Sequential Computed Tomography Morphometry–derived Muscle Loss Is Associated With Kidney Dysfunction and Mortality in Kidney Transplant Recipients Undergoing Serial Clinically Indicated Computed Tomography Imaging Authors: Dominik Kylies, Julian Kylies, Fabian Haas, Isabel Molwitz, Tobias B. Huber, Malte A. Kluger, Florian Grahammer Published: 2026 ABSTRACT Background.Kidney transplantation is the optimal treatment for end-stage kidney disease, yet long-term outcomes remain heterogeneous, underscoring the need for improved clinical risk stratification. Sarcopenia assessed by computed tomography (CT)–based morphometry has emerged as a promising biomarker across multiple conditions, including solid organ transplantation. Importantly, muscle metrics can be derived from CT scans obtained for clinical indications, enabling opportunistic screening without additional imaging burden. However, longitudinal muscle trajectories after kidney transplantation derived from indication-driven CT imaging and their associations with graft dysfunction and survival remain poorly characterized.Methods.In this retrospective single-center study, CT morphometry at the L3 level was analyzed in kidney transplant recipients who underwent 2 clinically indicated abdominal CT scans. Skeletal Muscle Index, Paraspinal Muscle Index, Psoas Muscle Index, and visceral adipose tissue were quantified. Longitudinal changes between both time points were related to kidney allograft dysfunction, defined as major adverse kidney events, and overall survival. Survival analyses were performed using Kaplan-Meier estimates and Cox proportional hazards models.Results.Seventy-three kidney transplant recipients were included. Patients with allograft dysfunction demonstrated markedly accelerated muscle loss compared with those maintaining stable graft function. Muscle decline was primarily driven by progression to end-stage kidney disease, whereas isolated ≥25% estimated glomerular filtration rate decline without graft failure was not associated with significant sarcopenia progression. Adverse CT morphometry was independently associated with reduced survival in Kaplan-Meier and multivariable Cox regression analyses adjusted for age, sex, and major adverse kidney events.Conclusions.Among kidney transplant recipients undergoing indication-driven imaging, posttransplant skeletal muscle loss is closely linked to allograft dysfunction and independently associated with mortality. Opportunistic CT-based morphometry may therefore represent a valuable tool for risk stratification in vulnerable transplant populations. Prospective studies are warranted to validate these findings and assess interventional implications. INTRODUCTION Kidney transplantation is the treatment of choice for patients with end-stage kidney disease (ESKD), a global health burden.1 Compared with dialysis, it is superior in cost-effectiveness, quality of life, and overall survival.2,3 Although advances in surgical techniques and medical management have markedly improved short-term outcomes,4,5 in the long-term, transplant recipients still face a wide spectrum of complications6 and the long-term course after kidney transplantation remains highly variable. Accordingly, there is a need for novel prognostic markers to enable individualized posttransplant care and informed clinical decision-making, particularly in long-term transplant recipients with complicated clinical courses. In this context, quantitative assessments of skeletal muscle mass and sarcopenia, indicated by reduced muscle strength, have emerged as promising candidate biomarkers.7 Their impact on adverse clinical outcomes, morbidity, and mortality has already been demonstrated across a broad spectrum of conditions,8-15 including chronic kidney disease,16-20 where the link between glomerular filtration rate (GFR) decline and muscle loss has been well established.21 Among the available methods for muscle quantification, computed tomography (CT)-based morphometry emerged as an increasingly used objective assessment tool8,22-24 as it allows quick, standardized, and reproducible quantification of muscle and adipose mass via indices such as Skeletal Muscle Index (SMI), Psoas Muscle Index (PMI), Paraspinal Muscle Index (PSMI), and visceral adipose tissue (VAT).25 Importantly, the increased use of CT imaging, frequently performed for clinical indications, may facilitate the applicability of this approach in real-world cohorts.25 This approach has also been increasingly used in solid organ transplantation, where sarcopenia emerged as a promising biomarker26: In liver transplant recipients, sequential CT-morphometric assessment of muscle parameters was linked to survival after liver transplantation.27 In heart transplant recipients, pectoral muscle parameters before heart transplantation were linked to clinical posttransplant outcomes.28 Similarly, in kidney transplantation, the presence of sarcopenia before or at the time of transplantation is known to be associated with poor posttransplant outcomes.29,30 Despite these promising data, evidence on posttransplant muscle parameters in kidney transplant recipients remains limited, particularly with respect to longitudinal assessments using CT-based morphometry at multiple time points. This is largely attributable to the fact that, after kidney transplantation, imaging is typically performed for clinical indications rather than according to standardized follow-up protocols. As a consequence, longitudinal CT-based analyses after kidney transplantation have so far been scarcely performed, as they would be largely restricted to patients with multiple indication-driven CT scans, a subgroup that is likely to represent a highly heterogenous cohort of transplant recipients with more complicated clinical courses. Nevertheless, this vulnerable patient population may particularly benefit from improved prognostic stratification. In this context, CT-based morphometric assessments derived from clinically acquired imaging may provide valuable insights into muscle mass trajectories after kidney transplantation and their implications for survival, thereby supporting more individualized posttransplant management strategies. In this study, we therefore investigated longitudinal muscle mass trajectories in relation to kidney allograft function and mortality using sequential CT-based morphometry at multiple time points after kidney transplantation, focusing on a cohort of transplant recipients with indication-driven imaging. MATERIALS AND METHODS Study Design and Patient Selection A waiver of written informed consent was obtained from the local ethics committee (ID: 2025-300649-WF) due to the retrospective analysis of anonymized data in this study. All analyses were conducted in accordance with institutional and ethical guidelines, including the Declaration of Helsinki. Kidney transplant patients were eligible for inclusion when the following criteria were met: 18 y or older, availability of artifact-free CT imaging at the L3 vertebral level at ≥2 time points after transplantation (time point (t)1 and t2), and complete clinical and laboratory data at both time points. Exclusion criteria were age younger than 18 y, insufficient imaging data availability, or quality. Between 2013 and 2023, of 689 kidney transplant recipients screened, 73 patients who met all inclusion and no exclusion criteria were included in our study. The presence of an additional prior solid organ transplant was not considered a formal exclusion criterion in our study. Although the majority of patients had received only a kidney allograft, 12 patients also had a history of another solid organ transplant (pancreas: n = 8, liver: n = 3, lung: n = 1). For all patients included, demographic data, laboratory data (including serum creatinine and estimated GFR [eGFR]), clinical data (including comorbidities and dialysis status), and imaging data were obtained at both time points of the study. eGFR was calculated using the 2021 Chronic Kidney Disease Epidemiology Collaboration creatinine-based equation. Major adverse kidney event (MAKE) was defined as ESKD of the renal allograft (sustained dialysis or eGFR <15 mL/min) or a decrease in eGFR ≥25% between time points in accordance with previous reports.31,32 The overall study design is illustrated in Figure 1A. Patient selection, study design, and demonstration of CT-based morphometric analysis. A, Overview of the study design (was partially created with BioRender.com). B, CT cross-section at the L3 level (image 1) with exemplary segmentation of SMI (image 2) and VAT (image 3). CT, computed tomography; SMI, Skeletal Muscle Index; VAT, visceral adipose tissue. Analysis of CT Morphometrics Axial CT slices were obtained from the local clinical picture archiving and communication systems at the level of the L3 level for further morphometric analysis using the Fiji imaging software (version 2.3.0/1.53q, Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany). The L3 region was chosen because it has been demonstrated to be well representative of the whole-body muscle volume.33 For the assessment of cross-sectionally measured skeletal muscle and VAT mass, if available, the portal-venous contrast scan phase was preferably used, although other scan phases were used if the portal-venous contrast scan phase was not available in accordance with previous reports.34 A semi-automated, observer-supervised threshold-based method (segmentation of muscle mass: –29 to 150 HU) was applied, following previously established protocols.22,27,35 To determine the SMI, a region of interest (ROI) was manually defined along the outer and inner parameters of the whole abdominal muscle, thereby excluding L3, VAT, and visceral organs as well as outer adipose tissue. Supervised thresholding was then performed to identify skeletal muscle tissue as described previously.36 The other muscle parameters, PSMI, PMI, were obtained using manual ROI annotation in accordance with previous reports.36,37 Similarly, to measure VAT a ROI capturing visceral organs and adipose tissue was drawn along the inner border of the skeletal muscle and the L3 was then followed by supervised thresholding as previously described.38,39 All skeletal muscle parameters as well as VAT were quantified and normalized to patient height (cm2/m2).27,36,37 CT-morphometric analyses were performed in the entire cohort (N = 73) as well as in a predefined subcohort restricted to recipients undergoing first-time kidney transplantation without a history of other solid organ transplantation (n = 44). Within this subcohort, where applicable, additional stratified analyses were conducted for living donor transplant recipients (n = 14) and deceased donor transplant recipients (n = 30). A representative annotated CT cross-sectional image is shown in Figure 1B. Statistical Analysis Statistical analyses and data visualization were performed using GraphPad Prism (version 10.4.1) and RStudio (R version 4.4.2). The Wilcoxon matched-pairs signed-rank test was used to assess longitudinal changes in CT-morphometric metrics and, where applicable, in clinical and laboratory parameters. Differences in CT-morphometric and laboratory parameters between 2 groups were assessed using the Mann-Whitney U test, and comparisons between 3 groups were assessed using the Kruskal-Wallis test. Survival Analysis and Cox Proportional Hazards Regression Analysis Kaplan-Meier survival analyses were used to compare overall survival between patients with pronounced PSMI loss (≥14%) and those with stable or less pronounced PSMI changes (<14%). Survival curves were compared using the log-rank (Mantel-Cox) test. For all survival analyses, follow-up time was defined from tCT2, which served as the index time point to harmonize follow-up across patients. To identify the most informative morphometric parameter for survival stratification, receiver operating characteristic curve analyses were performed for the different morphometry trajectory measures. The parameter demonstrating the highest discriminatory performance was selected for the primary survival analysis. PSMI change yielded the best area under the curve (0.73, P = 0.0018) and was therefore used for Kaplan-Meier analyses. The optimal cutoff value for PSMI change was determined using Youden’s Index (J). This approach identified a 14% decrease in PSMI as the most predictive threshold, with patients exceeding this level of muscle loss exhibiting significantly worse survival than those with smaller declines. To evaluate whether longitudinal changes in PSMI between tCT1 and tCT2 were independently associated with survival, Cox proportional hazards regression analyses were performed. PSMI change was entered as the primary explanatory variable. Univariable Cox models were first fitted to assess crude associations with survival, followed by multivariable Cox proportional hazards models adjusted for age, sex, and the presence of MAKE. Model comparisons were conducted using likelihood ratio tests. Effect estimates are reported as hazard ratios (HRs) with corresponding 95% confidence intervals (CIs). Associations between categorical variables were assessed using the Fisher exact test, as appropriate. All statistical tests were 2-sided, with a P value of <0.05 considered statistically significant. This was an exploratory study without confirmatory sample size testing. All P values are thus considered descriptive. RESULTS Baseline Clinical Characteristics The median time between kidney transplantation and first CT (t1) was 4 (interquartile range [IQR], 21) mo. No statistically significant differences in the time between kidney transplantation and first CT scan (t1) were observed between kidney transplant patients with and without MAKE (3 (IQR, 20) mo versus 4.5 (IQR, 19.25) mo, P = 0.4266). The median time between t1 and t2 for all patients was 17 (IQR, 44) mo, with no significant difference between the subgroups of kidney transplant patients with and without MAKE (24 [IQR, 37] mo versus 12 [IQR, 47] mo, P = 0.2729). The median age at baseline was 58 y (IQR, 25), with no significant age difference between patients with MAKE (62 y; IQR, 25) and those with stable transplant function (55 y; IQR, 22; P = 0.33). The proportion of allocations within the Eurotransplant Senior Program was numerically higher in the MAKE group compared with the non-MAKE group (44.0% versus 27.6%), although this difference did not reach statistical significance (P = 0.26). Overall, 38.4% of the cohort were women, with comparable sex distribution between groups (39.4% versus 37.5%, P = 0.87). Median baseline eGFR was 29 mL/min (IQR, 27) without significant differences between groups (29 versus 30 mL/min, P = 0.95). The median Liu Comorbidity Index40,41 at baseline was 3 (IQR, 3), and there were no significant differences between MAKE and non-MAKE patients (P = 0.48). The most common underlying kidney disease was IgA nephropathy (16.4%), followed by autosomal dominant polycystic kidney disease (13.7%) and diabetic nephropathy (13.7%). Interstitial nephritis accounted for 6.8%, whereas the largest fraction was classified as other or unclear etiologies (49.3%). Living donor transplantation was performed in 26% of patients, with no significant difference between groups (24.2% versus 27.5%, P = 0.79). Patient demographics and comprehensive clinical characteristics are detailed in Table 1. The CT indications are summarized in Table S1 (SDC, https://links.lww.com/TXD/A871). TABLE 1.Baseline patient characteristicsOverallMAKEStable transplant P Patients, n733340Age, y58 (25)62 (25)55 (22)0.33Female sex, %38.439.437.50.87eGFR, mL/min29 (27)29 (28)30 (29)0.95 ≥60 (n patients)954NA 45–59 (n patients)1037NA 30–44 (n patients)1789NA 15–29 (n patients)20911NA <15 (n patients)1789NALiu Comorbidity Index3 (3)3 (4)3 (3)0.48Cause of kidney disease, % IgAN16.421.212.5NA ADPKD13.79.117.5NA Diabetes13.79.117.5NA Interstitial nephritis6.89.15.0NA Other/unclear cause49.351.547.5NALiving donor transplant, %26.024.227.50.79Maintenance immunosuppression, % Corticosteroids93.293.995.0NA mTOR inhibitor39.745.535.0NA MMF/MPA58.951.557.5NA Tacrolimus71.260.675.0NA Cyclosporine A19.227.317.5NABaseline characteristics of the study cohort. Continuous variables if not stated otherwise are presented as median with IQR. P values, where provided, indicate differences between patients with MAKE and those with stable transplant function. “NA” indicates that statistical testing was not performed because of small subgroup sizes. Minor rounding discrepancies in individual values may occur.ADPKD, autosomal dominant polycystic kidney disease; eGFR, estimated glomerular filtration rate; IgAN, immunoglobulin A nephropathy; IQR, interquartile range; MAKE, major adverse kidney event; MMF, mycophenolate mofetil; mTOR, mammalian target of rapamycin. Baseline patient characteristics Baseline characteristics of the study cohort. Continuous variables if not stated otherwise are presented as median with IQR. P values, where provided, indicate differences between patients with MAKE and those with stable transplant function. “NA” indicates that statistical testing was not performed because of small subgroup sizes. Minor rounding discrepancies in individual values may occur. ADPKD, autosomal dominant polycystic kidney disease; eGFR, estimated glomerular filtration rate; IgAN, immunoglobulin A nephropathy; IQR, interquartile range; MAKE, major adverse kidney event; MMF, mycophenolate mofetil; mTOR, mammalian target of rapamycin. Overall Trajectory of Muscle Mass and Adipose Tissue in Kidney Transplant Recipients We first analyzed the overall muscle mass trajectory of the entire cohort. Here, sequential CT-based morphometry revealed a modest yet statistically significant decline in skeletal muscle mass between t1 and t2. The median SMI significantly decreased from 44.01 (IQR, 9.35) cm2/m2 at t1 to 39.68 (IQR, 9.29) cm2/m2 at t2 (relative change: −8.54% [IQR, 16.32], P < 0.0001; Figure 2A). PSMI significantly declined from 14.61 (IQR, 3.54) cm2/m2 to 13.88 (IQR, 4.02) cm2/m2, with a relative change of –3.55% (IQR, 19.09; P = 0.0049; Figure 2B). Similarly, the PMI showed a significant reduction from 6.32 (IQR, 2.17) cm2/m2 to 5.68 (IQR, 2.74) cm2/m2 (relative change: −10.2% [IQR, 19.66], P < 0.0001; Figure 2C). VAT displayed a trend toward a more pronounced decline from 55.55 (IQR, 41.23] cm2/m2 at t1 to 48.53 [IQR, 37.20] cm2/m2 at t2 (relative change: –2.72% [IQR, 56.41]) that did not reach statistical significance (P = 0.11; Figure 2D). Similar trajectories were observed in predefined subgroups of kidney transplant recipients undergoing their first transplantation without a history of other solid organ transplantation. Analyses were performed separately for the overall subgroup combining living and deceased donor transplants as well as stratified by donor type (living and deceased donor, respectively). Across these subgroup analyses, longitudinal patterns of muscle mass and adipose tissue changes were comparable with those observed in the overall cohort (Table S2, SDC, https://links.lww.com/TXD/A871). Longitudinal changes in CT morphometry of muscle mass and VAT in kidney transplant recipients. A, SMI showed a significant decline between t1 and t2 in all kidney transplant recipients. B, Similarly, PSMI also declined significantly during the observation period between time points. C, Kidney transplant recipients demonstrated a significant reduction in PMI between t1 and t2. D, VAT did not differ significantly between t1 and t2 in kidney transplant recipients. CT, computed tomography; PMI, Psoas Muscle Index; PSMI, paraspinal skeletal muscle index; SMI, Skeletal Muscle Index; VAT, visceral adipose tissue. Association Between MAKE and Longitudinal Changes in Muscle and Adipose Tissue After establishing the overall trajectory of muscle and visceral fat mass, revealing a modest decline in skeletal muscle parameters across the entire cohort, we next analyzed whether patients experiencing kidney dysfunction defined by MAKE would also display a more pronounced loss of muscle and adipose tissue compared with those patients who maintain a stable kidney function. For that, we divided our cohort into 2 groups: patients with MAKE and those without, thereby maintaining stable kidney function during the study. Here, we found that overall, kidney transplant patients experiencing MAKE displayed a significantly higher loss of muscle mass compared with those with a stable kidney function: during the observation period, kidney transplant patients experiencing MAKE showed a significantly greater decline in SMI, decreasing by –13.58% (IQR, 20.48) compared with –3.56% (IQR, 15.20) in the stable kidney function group (P = 0.002; Figure 3A). Similarly, PSMI declined by –12.52% (IQR, 23.56) in the MAKE group, whereas remaining essentially unchanged in the stable group (−0.24% [IQR, 11.74], P = 0.0114; Figure 3B). A comparable and statistically significant pattern was also observed for PMI, which decreased by –17.32% (IQR, 24.17) in MAKE patients, compared with a –6.39% (IQR, 14.35) change in the stable group (P = 0.0041; Figure 3C). VAT area declined by –12.31 (IQR, 61.53) in MAKE patients and by –1.11% (IQR, 50.63) in the stable group, showing a nonsignificant trend toward greater VAT loss in MAKE (P = 0.1865; Figure 3D). Differences in CT morphometry changes between kidney transplant recipients with MAKE and stable kidney function. A, SMI declined significantly more in patients with MAKE compared with those with stable kidney function. B, Like SMI, PSMI also showed a significantly greater reduction over time in the MAKE group compared with stable kidney function during the observation period. C, PMI loss as well was significantly more pronounced in patients with MAKE than in those with stable kidney transplant function. D, VAT showed a trend toward a more pronounced loss in the MAKE group but did not reach statistical significance. CT, computed tomography; MAKE, major adverse kidney event; PMI, Psoas Muscle Index; PSMI, paraspinal skeletal muscle index; SMI, Skeletal Muscle Index; VAT, visceral adipose tissue. Importantly, at t1, we did not observe any statistically significant differences in baseline body composition between patients with and without subsequent MAKE for all morphometry parameters: (SMI: 43.58 [IQR, 11.88] cm2/m2, versus 44.75 [IQR, 7.18] cm2/m2, P = 0.79; Figure S1A, SDC, https://links.lww.com/TXD/A871; PSMI: 14.39 [IQR, 4.01] cm2/m2 versus 14.83 [IQR, 3.21] cm2/m2, P = 0.50; Figure S1B, SDC, https://links.lww.com/TXD/A871; PMI: 6.15 [IQR, 2.15] cm2/m2 versus 6.80 [IQR, 2.25] cm2/m2, P = 0.26; Figure S1C, SDC, https://links.lww.com/TXD/A871; VAT: 55.41 [IQR, 33.02] cm2/m2 versus 57.93 [IQR, 42.94] cm2/m2, P = 0.79; Figure S1D, SDC, https://links.lww.com/TXD/A871). The analysis was repeated in predefined subgroups restricted to kidney transplant recipients undergoing their first transplantation without a history of other solid organ transplantation. Subgroup analyses were performed for the combined cohort of living and deceased donor transplants as well as stratified by donor type (living and deceased donor transplantation). Across these subgroup analyses, patients with MAKE consistently exhibited more accelerated loss of muscle and VAT tissue than those with stable kidney function, mirroring the patterns observed in the overall cohort (Table S3, SDC, https://links.lww.com/TXD/A871). In summary, although baseline (t1) body composition did not differ between groups, patients with MAKE showed a significantly greater loss of skeletal muscle mass over time than those with stable kidney function. Comparative Analysis of CT Morphometry in MAKE Subcomponents We also performed further subgroup analysis of patients developing the combined outcome MAKE to investigate whether both suboutcomes (ESKD and ≥25% decrease in eGFR but not reaching ESKD) are individually associated with a more accelerated loss of muscle and adipose tissue compared with patients with a stable kidney function or whether the loss of muscle tissue is primarily driven by 1 suboutcome. To achieve this, we thus divided patients experiencing MAKE further into 2 groups: ESKD (n = 22 patients) and a decrease in eGFR ≥25% without experiencing ESKD (n = 11 patients), and compared their muscle and adipose tissue trajectories to those who maintain a stable kidney function during the study. For SMI, patients progressing to ESKD showed a significantly greater decline of −16.28% (IQR, 18.27) compared with those with stable transplant function (–3.56%; IQR, 15.20; P = 0.0014; Figure 4A). Conversely, patients with a ≥25% loss in eGFR who did not progress to ESKD showed no significantly increased loss in SMI (–8.03%; IQR, 8.42) compared with those with stable transplant function (P = 0.6113; Figure 4A). Relative changes in CT morphometry in kidney transplant recipients with stable kidney function and the MAKE suboutcomes ESKD and ≥25% eGFR loss without progression to ESKD. A, SMI declined significantly more in the ESKD group compared with the stable kidney function group, whereas the ≥25% eGFR loss group did not differ significantly from stable patients. B, PSMI showed a significantly greater reduction in the ESKD group compared with stable kidney function, whereas no significant difference was observed between the ≥25% eGFR loss group and stable patients. C, PMI loss was significantly more pronounced in ESKD compared with stable kidney function, with no significant difference between the ≥25% eGFR loss group and stable patients. D, VAT showed a trend toward greater loss in the ESKD group that did not reach statistical significance, and the ≥25% eGFR loss group did not differ significantly from the stable group. CT, computed tomography; eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; PMI, Psoas Muscle Index; SMI, Skeletal Muscle Index; VAT, visceral adipose tissue. Similarly, transplant patients with ESKD experienced a significant decline in PSMI of −15.9% (IQR, 21.57) compared with –3.55% (IQR, 19.09) in the stable group (P = 0.0032; Figure 4B). In contrast, the change in the ≥25% GFR loss group (−0.67%; IQR, 24.80) was not significantly different from the stable group (P > 0.9999; Figure 4B). For PMI, the percentage decrease were −20.03% (IQR, 25.46) in ESKD group and –10.20% (IQR, 18.01) in the ≥25% GFR loss group compared with −6.39% (IQR, 14.35) in the stable group, representing a statistically significant difference between the ESKD and stable groups (P = 0.0023) but no significant difference between the ≥25% GFR loss and the stable group (P = 0.8190; Figure 4C). For VAT, patients with ESKD showed a reduction of −26.67% (IQR, 61.11) compared with –1.11% (IQR, 50.63) in stable patients, indicating a trend toward greater loss that did not meet statistical significance (P = 0.1625; Figure 4D). The change in the ≥25% GFR loss group (−0.89%; IQR, 46.03) was not significantly different compared with the stable group (P > 0.9999; Figure 4D). In addition, the analyses were repeated in subgroups of first kidney transplant recipients without a history of other solid organ transplantation. Analyses were conducted separately within the combined cohort of living and deceased donor transplants as well as separately for living and deceased donor transplantation. Similar to the results observed in the entire cohort, longitudinal morphometric changes were primarily driven by progression to ESKD, whereas trajectories in patients with isolated eGFR decline without ESKD were comparable with those observed in patients with stable kidney function (Table S4, SDC, https://links.lww.com/TXD/A871). Overall, we found that although patients with ESKD showed statistically significant declines in muscle mass over time compared with patients with a stable kidney function, no statistically significant differences were found between those showing only a ≥25% loss in eGFR but not progressing to ESKD and those with a stable kidney function. Prognostic Relevance of Longitudinal CT-derived Muscle Loss for Patient Survival in Kidney Transplant Recipients Given the observed associations between graft dysfunction and muscle loss, we next assessed whether longitudinal changes in CT-based morphometry provide prognostic information for overall survival. First, Kaplan-Meier survival analysis demonstrated a significantly reduced survival in kidney transplant recipients exhibiting pronounced paraspinal SMI (PSMI) loss compared with those without relevant PSMI loss (log-rank P < 0.0001; Figure 5). The mean follow-up time was 27.0 ± 28.2 mo for the entire cohort. The median survival was 5 mo in the PSMI loss group. In contrast, median survival was not reached in the no-PSMI loss group due to <50% of observed events during follow-up. Association of longitudinal paraspinal muscle loss with survival in kidney transplant recipients. A, Kaplan-Meier survival curve comparing overall survival between kidney transplant recipients with pronounced PSMI loss and those without relevant PSMI loss. B, Forest plot showing HRs for mortality associated with PSMI loss derived from univariable (crude) and multivariable Cox proportional hazards regression analyses. Multivariable models were adjusted for age, sex, and MAKEs. HR, hazard ratio; MAKE, major adverse kidney event; PSMI, Paraspinal Skeletal Muscle Index. To further explore the association between muscle loss and mortality, uni- and multivariable Cox proportional hazards regression models were generated. In univariable Cox regression analysis, PSMI loss was significantly associated with an increased risk of death (HR, 8.598; 95% CI, 3.606-20.502; P < 0.0001), alongside age and the occurrence of MAKEs, whereas sex did not display a significant association with mortality (Table 2). In a multivariable Cox regression model adjusting for age, sex, and MAKE, PSMI loss remained independently and significantly associated with mortality (HR, 4.358; 95% CI, 1.604-11.837; P = 0.0039), whereas age also retained an independent association. TABLE 2.Univariable and multivariable Cox proportional hazards regression analyses for overall survivalVariableHR95% CI P Univariable analysis PSMI loss8.5983.606-20.502<0.0001 Age1.0681.031-1.1050.0002 Sex0.8260.380-1.7930.6290 MAKE2.5641.192-5.5140.0159Multivariable analysis PSMI loss4.3581.604-11.8370.0039 Age1.0501.014-1.0880.0069 Sex0.8880.392-2.0120.7770 MAKE1.4940.628-3.5530.3640Univariable and multivariable Cox proportional hazards regression analyses evaluating the association of PSMI loss, age, sex, and MAKE with overall survival after kidney transplantation. The multivariable model adjusts for age, sex, and graft dysfunction (MAKE).CI, confidence interval; HR, hazard ratio; MAKE, major adverse kidney event; PSMI, Paraspinal Skeletal Muscle Index. Univariable and multivariable Cox proportional hazards regression analyses for overall survival Univariable and multivariable Cox proportional hazards regression analyses evaluating the association of PSMI loss, age, sex, and MAKE with overall survival after kidney transplantation. The multivariable model adjusts for age, sex, and graft dysfunction (MAKE). CI, confidence interval; HR, hazard ratio; MAKE, major adverse kidney event; PSMI, Paraspinal Skeletal Muscle Index. Importantly, this association remained significant after adjustment for MAKE, indicating that longitudinal muscle loss provides prognostic information beyond graft function alone. In contrast, MAKE was no longer independently associated with mortality after multivariable adjustment (Table 2). To assess potential confounding by malignancies, additional analyses were performed on the basis of the presence of a malignant disease. No significant difference in survival was observed between patients with and without malignancy (Figure S2, SDC, https://links.lww.com/TXD/A871). Furthermore, the occurrence of MAKE did not differ significantly between patients with and without malignancy (52.9% versus 42.9%, P = 0.58). DISCUSSION In recent years, the assessment of sarcopenia and particularly CT-based morphometry has gained increasing attention as a potential prognostic assessment tool across multiple cohorts, including solid organ transplantation.27-30,42 In liver transplantation recipients, adverse CT morphometry was linked to increased mortality.27 In kidney transplantation, pretransplant sacropenia was associated with adverse outcomes.29 However, data on muscle trajectories measured by CT morphometry and their implications after kidney transplantation remain sparse. However, in kidney transplant recipients, data on muscle trajectories measured by CT morphometry after kidney transplantation and their associations with renal function and clinical outcomes remain sparse. The reasons for the paucity of longitudinal CT-based morphometry data in kidney transplant recipients include the lack of imaging data at standardized time points as well as an often highly heterogeneous clinical course, with diverse and frequently unpredictable complications that are difficult to standardize within a clinical cohort. Under this background, we conducted a retrospective pilot study to investigate the dynamics of muscle and adipose tissue changes in a selected subgroup of kidney transplant recipients with abdominal CT imaging available at 2 time points during transplantation. Using CT-based morphometry, we analyzed trajectories of muscle and VAT in relation to allograft function and, importantly, explored their potential use as prognostic imaging-based biomarkers. The key finding of our study is that longitudinal skeletal muscle loss, as assessed by CT-based morphometry, demonstrates independent prognostic relevance for patient survival in kidney transplant recipients. Specifically, patients with pronounced muscle loss exhibited significantly reduced survival compared with those with stable muscle mass. Importantly, in the context of the observed associations between allograft dysfunction and muscle loss, both allograft dysfunction and muscle loss were significantly associated with mortality in univariable analyses. However, after multivariable adjustment, only longitudinal muscle loss remained independently associated with death, whereas the association for allograft dysfunction was attenuated. These findings suggest that CT-based morphometry captures clinically relevant risk beyond graft function alone and may therefore serve as a promising imaging-derived biomarker for risk stratification in a subgroup of high-risk kidney transplant recipients undergoing serial imaging. However, these findings require confirmation in larger, ideally prospective studies. Despite these encouraging observations, several important limitations inherent to this approach must be critically discussed. The most prominent limitation is the highly selected patient cohort. Only recipients with 2 available abdominal CT scans were included, resulting in a substantial drop-out rate. Consequently, the analyzed cohort likely represents a subgroup of patients with a complicated posttransplant course. The observed trajectories and their associations with clinical outcomes may therefore not be generalizable to the broader kidney transplant population. Indeed, a single-center study from Japan using an alternative method of muscle mass assessment reported an increase in muscle mass after transplantation in 66% of patients.43 Although direct comparison is challenging, this discrepancy likely reflects differences in cohort composition, with potentially fewer posttransplant complications in that population. Another limitation relates to the use of creatinine-based eGFR estimation. As serum creatinine is influenced by muscle mass, reduced skeletal muscle mass may lead to an overestimation of kidney function, which should be considered when interpreting associations between muscle loss and eGFR-based outcomes. The allocation program may represent a potential confounder, given its association with recipient age and clinical outcomes; however, no significant imbalance between groups was observed in our cohort. Further limitations include the single-center, retrospective design and the modest sample size, which may again limit generalizability and statistical power, particularly for subgroup analyses. Therefore, negative findings, such as the absence of a detectable association between ≥25% eGFR decline and sarcopenia, should be interpreted with caution. Another consideration concerns the survival analysis. Despite no significant difference in the time intervals between groups, the ranges of these intervals were broad. This heterogeneity complicated the definition of a uniform starting point for time-to-event analyses. To address this issue, we defined t2 as the index time point and analyzed survival thereafter, thereby harmonizing follow-up across patients despite variability in preceding intervals. Adding further heterogeneity, the cohort included patients undergoing first and subsequent kidney transplantations across different allocation modes, including Eurotransplant Kidney Allocation System, Eurotransplant Senior Program, and living donation. To mitigate potential bias and improve cohort homogeneity, key muscle trajectory analyses were repeated in a subcohort restricted to first kidney transplant recipients without a history of other solid organ transplantation and were additionally stratified by living versus deceased donor transplantation. Importantly, the overall morphometric signals were comparable with those observed in the full cohort. Furthermore, skeletal muscle density was not assessed because of heterogeneity in available CT contrast phases. Finally, sex-specific analyses were not performed because of limited overall sample size and subgroup numbers. Despite its limitations, to our knowledge, our study is among the first to investigate muscle and adipose tissue mass trajectories after kidney transplantation using sequential CT-based morphometry at multiple time points. The underlying biological mechanisms are not yet fully understood. Kidney dysfunction may accelerate muscle loss via systemic inflammation and catabolic metabolism. In this context, molecular muscle-kidney crosstalk has been described, directly linking kidney disease to muscle wasting.44,45 Conversely, patients with a complicated posttransplant course may be exposed to important confounding factors, such as recurrent infections and prolonged hospitalizations, which can adversely affect muscle mass and allograft function in parallel. Given our study design, no conclusions regarding causality can be drawn and our findings must thereby remain strictly descriptive. Despite the limitations as discussed earlier, the demonstration of associations among kidney dysfunction, sarcopenia, and mortality provides novel and potentially clinically meaningful insights in a real-world scenario that have not previously been reported in the transplant literature. These findings may open avenues for risk stratification and targeted interventions in vulnerable patients. Future prospective and ideally multicenter trials are warranted to validate our findings and further delineate the interplay between graft dysfunction, sarcopenia, and clinical outcomes after kidney transplantation. Furthermore, translational molecular studies would aid in deciphering the pathomechanisms linking muscle wasting to kidney function in kidney transplant patients. Uncovering the molecular mechanisms could aid in paving the way for the discovery of novel treatment inventions. Prospectively, the used and related techniques may gain further importance as nowadays automated 3-dimensional body composition analyses are feasible using artificial intelligence tools.46 In conclusion, kidney dysfunction may be associated with skeletal muscle loss, and longitudinal declines in skeletal muscle mass after kidney transplantation may serve as an independent predictor of posttransplant mortality. Supplementary Material