Objectives: Tacrolimus is a cornerstone immunosup-pressant in kidney transplant; however, optimal trough levels during the late maintenance phase remain uncertain. This study investigated the association between tacrolimus trough levels measured 10 to 15 years after transplant and long-term clinical outcomes. Material and Methods: A retrospective review was conducted of 1747 recipients who received kidney transplants between 1990 and 2005. Among them, 290 patients maintained graft function beyond 15 years with continuous tacrolimus use from post-transplant year 10 to posttransplant year 15. Out-comes were compared between patients with mean tacrolimus trough levels <4 ng/mL and ≥4 ng/mL, with follow-up through December 2024. Propensity score matching was performed using 1:1 nearest-neighbor matching to balance baseline characteristics.
Results: In the matched cohort (n = 204), no significant differences were observed in kidney graft survival or rejection-free survival between groups. Secondary outcomes, including infection, calcineurin inhibitor toxicity, and malignancy, also showed no significant differences, except for a higher incidence of sepsis in the lower-level group (P = .049). In the unmatched cohort, the cumulative incidence of de novo donor-specific antibody formation beyond 15 years did not differ significantly, although a lower trend was observed in the higher-level group.
Conclusions: In long-term kidney transplant recipients with stable graft function, modest differences in tacrolimus trough levels during the late maintenance phase were not significantly associated with graft survival or rejection outcomes. Nonetheless, the obser-ved sepsis risk and the immunological implications of de novo donor-specific antibody formation underscore the importance of personalized tacrolimus exposure strategies and continued immunological surveillance in long-term transplant care.

Key words : Drug monitoring, Kidney transplantation

Introduction

Kidney transplant remains the optimal treatment for patients with end-stage renal disease, and long-term graft survival is a central goal in posttransplant management. Tacrolimus, a calcineurin inhibitor (CNI), is a cornerstone of immunosuppression the-rapy and effectively prevents allograft rejection by inhibiting T-cell activation.¹ Given its narrow therapeutic window, therapeutic drug monitoring is essential to balance efficacy and toxicity.² Overall tacrolimus exposure has been associated with the development of de novo donor-specific antibodies (dnDSA) during the first year after transplant, with lower levels demonstrating a uniform, graded increase in risk.³ Consistently, reduced tacrolimus exposure during this early period has also been linked to a higher risk of acute rejection and graft failure.^3,4 To mitigate these risks, several studies have recommended maintaining tacrolimus trough levels >5 ng/mL during the first 12 months after transplant.⁵

Although the early posttransplant period has been extensively studied with respect to optimal tacrolimus exposure, data on appropriate trough levels during the late maintenance phase, parti-cularly beyond 10 years after transplant, remain limited. Several studies have described the clinical and immunological characteristics of long-term kidney graft survivors; however, the relationship between immunosuppressant levels and clinical outcomes has not been clearly defined.^6-11 An analy-sis using data from the national Scientific Registry of Transplant Recipients has reported improved median graft survival in both deceased donor kidney transplant (from 8.2 years in 1995-1999 to an estimated 11.7 years in the most recent era) and living donor kidney transplant (from 12.1 years in 1995-1999 to an estimated 19.2 years in 2014-2017).¹² Consequently, further research is needed to optimize immunosuppressant use in long-term kidney transp-lant survivors. In this context, it remains unclear whether maintaining higher tacrolimus trough levels confers additional benefits in graft survival or rejection prevention among stable long-term recipients. Moreover, excessive CNI exposure during the late phases may increase the risk of CNI-related toxicity, infection, or malignancy.¹

Therefore, this study aimed to evaluate whether tacrolimus trough levels maintained between 10 and 15 years after kidney transplant are associated with long-term outcomes, including graft survival, rejection-free survival, infection, toxicity, and malig-nancy, in recipients with kidney allograft survival exceeding 15 years.

Materials and Methods

This single-center, retrospective observational study was conducted at Asan Medical Center, Seoul, Republic of Korea. Kidney transplant at this center began in 1990. All patients who underwent kidney transplant between January 1990 and December 2005 were reviewed. Patients were eligible for inclusion if their kidney allografts survived for >15 years and if tacrolimus therapy was continuously maintained between 10 and 15 years after kidney transplant. Tacrolimus trough levels and relevant clinical outcomes were analyzed. Patients aged <18 years at the time of transplant were excluded. Eligible patients were followed through December 2024. All living donors in this study were either a relative of the recipient or the spouse of the recipient. The study protocol was approved by the Institutional Review Board of Asan Medical Center (IRB No. 2023-0359), and the requirement for informed consent was waived due to the retrospective design of the study. The study protocols conformed to the ethical guidelines of the Declaration of Helsinki.

Given the extended study period, institutional immunosuppression protocols likely evolved in response to updates in clinical guidelines and drug availability, resulting in variability in the selection and dosing of immunosuppressive agents. All patients received standard immunosuppression therapy fol-lowing kidney transplant, generally consisting of a combination of CNI agents, antimetabolites, and corticosteroids. For the present study, only patients who remained on tacrolimus continuously between 10 and 15 years after transplant were included. Patients who switched to cyclosporine or other CNI agents, or patients who discontinued CNI therapy during this period, were excluded. Tacrolimus dosing was adjusted to achieve target trough levels according to standard clinical practice, and trough levels were regularly monitored throughout the follow-up period. The use of concurrent immunosup-pressive agents, including mycophenolate mofetil or azathioprine, corticosteroids, and sirolimus, were also recorded to account for differences in immuno-suppressive effects that could influence long-term graft outcomes.

Baseline characteristics of the study participants, including age, sex, body mass index, medical history, etiology of end-stage renal disease, dialysis history prior to transplant, maintenance immunosuppressive medications, and tacrolimus trough levels, were collected. Donor characteristics and transplant-related variables were also reviewed. When available, data on human leukocyte antigen (HLA) mismatches and the presence of DSAs were incorporated into the analysis.

The primary outcomes were kidney graft survival and rejection-free survival. Kidney graft survival was defined as the time from kidney transplant to graft failure, with graft failure defined as the resumption of dialysis or direct second transplant. Rejection-free survival was defined as the time from transplant to the first pathologically confirmed rejection episode. Rejection was diagnosed by allograft biopsy accor-ding to the Banff criteria. Protocol biopsies were not performed; biopsies were conducted when renal pathology was clinically suspected based on elevated serum creatinine levels or proteinuria. Biopsy spe-cimens were evaluated using light, electron, and immunofluorescence microscopy, with C4d staining performed in all cases.

Secondary outcomes included the incidence of clinically significant infection, CNI toxicity, and malignancy occurring beyond 15 years after transplant. Cytomegalovirus disease was defined as tissue-invasive disease, including cytomegalovirus esophagitis, colitis, and retinitis. BK virus infection was diagnosed irrespective of nephropathy if plasma BK virus DNA levels were >10 000 copies/mL. Sepsis was defined as the isolation of clinically significant pathogens from blood cultures. Additional infectious complications, such as acute pyelonephritis and pneumonia, were identified based on clinical diagno-ses and supported by microbiology or radiology findings when available. The CNI toxicity was defined as histologically confirmed CNI-induced injury on allograft biopsy. Malignancy was defined as any cancer diagnosis confirmed by tissue biopsy.

Testing for DSA was performed at the discretion of the treating physician, primarily in cases of unexplained graft dysfunction raising clinical suspicion of rejection. In patients with positive DSA results, follow-up testing was conducted more frequently based on clinical judgment. However, no standardized protocol for routine DSA surveillance was in place during the study period.

We used SAS software version 9.4 (SAS Institute) to perform propensity score matching and R software version 4.4.2 (R Foundation for Statistical Computing) to conduct all other analyses. We presented continuous data as mean ± SD and categorical data as frequencies (with percentages). We compared categorical variab-les with the chi-square test or the Fisher exact test and continuous variables with the 2-sample t test or Wilcoxon rank-sum test as appropriate.

To reduce potential confounding, propensity score matching was performed for comparisons of kidney graft survival and rejection-free survival. Matching variables included recipient age, sex, dialysis history prior to transplant, comorbidities (diabetes mellitus and hypertension), HLA mismatches, donor-related factors, the use of concomitant immunosup-pressive agents other than tacrolimus, and the presence of rejection episodes or CNI toxicity prior to 15 years after transplant. One-to-one greedy nearest neighbor matching without replacement was performed, applying a caliper width of 0.2 SD of the logit of the propensity score to minimize baseline differences between groups and achieve stringent matching. After 1:1 propensity score matching, survival analy-ses were conducted using the matched cohort. Cox proportional hazards regression models with robust variance estimators were used to assess the association between tacrolimus trough levels and graft survival. Hazard ratios and 95% confidence intervals were reported, and statistical significance was evaluated using the Wald test. We generated Kaplan-Meier curves to illustrate survival differences between groups.

We conducted analyses of cumulative incidence of dnDSA using unmatched data, constructed Kaplan-Meier survival curves, and evaluated dif-ferences between groups using the log-rank test.

Results

At our center, 1747 patients underwent kidney transplant between January 1990 and December 2005. Of this group, 1024 patients had functioning grafts for >15 years, among whom 294 kidney transplant recipients continuously received tacrolimus from 10 to 15 years after kidney transplant. Patients aged <18 years at the time of transplant (n = 4) were excluded. Accordingly, tacrolimus drug levels and clinical outcomes were analyzed in the remaining 290 patients (Figure 1).

The mean tacrolimus trough drug level between 10 and 15 years after kidney transplant was 4.07 ± 1.47 ng/mL (range, 1.35-9.08 ng/mL) with a median of 3.77 ng/mL (Figure 2). Based on this distribution, a cutoff value of 4 ng/mL was selected, and patients were categorized into 2 groups for subsequent analyses: patients with tacrolimus trough levels <4 ng/mL (lower-level group) and patients with levels <4 ng/mL (higher-level group).

Baseline characteristics of the study population prior to propensity score matching are presented in Table 1. These data provide a comprehensive overview of the entire cohort before adjustment for potential confounders. Although propensity score matching was applied to balance covariates between comparison groups in subsequent analyses, unmatched baseline characteristics are reported here to illustrate the initial distribution and clinical profile of the overall study cohort. Some baseline characteristics, such as sirolimus use, had a low prevalence in the study population and were therefore excluded from the matching process to preserve statistical stability and ensure the robustness of the analysis.

After 1:1 propensity score matching, 102 patients were included in each group. Table 2 presents the balance assessment of matching variables following propensity score matching. Covariate balance between the lower-level group and the higher-level group was evaluated using the absolute standardized difference, for which a value <0.10 was generally considered indicative of adequate balance. Although a small number of variables slightly exceeded this threshold, the matching was considered acceptable for subse-quent analyses, given the marginal nature of these deviations and the overall covariate balance achieved.

Table 3 summarizes the event frequencies of the primary outcomes observed during the 15-year follow-up period in the matched sample. In the matched cohort, Cox proportional hazards regression analysis demonstrated no significant difference in rejection-free survival between the lower-level group and higher-level group (hazard ratio, 0.72; 95% CI, 0.40-1.33; P = .297) (Table 4). Similarly, no significant difference in kidney graft survival was observed between the 2 groups (hazard ratio, 0.97; 95% CI, 0.54-1.74; P = .925). Kaplan-Meier survival curves for rejection-free survival and kidney graft survival in the matched cohort are presented in Figure 3. Consistent with the Cox regression findings, the survival curves demonstrated no significant dif-ferences between the lower-level group and the higher-level group for either outcome, as assessed by the Wald test.

Among the secondary outcomes, no significant differences were observed between the lower-level group and higher-level group in the overall incidence of infections, CNI toxicity, or malignancy. However, the incidence of sepsis was significantly higher in the lower-level group versus the higher-level group (P = .049) (Table 5).

As an additional analysis, the cumulative inci-dence of dnDSA formation beyond 15 years after transplant was evaluated in the unmatched cohort using Kaplan-Meier analysis. Patients who showed DSA-positive test results before 15 years after transplant (n = 11) were excluded, resulting in 279 patients included in the analysis. Table 6 presents event frequencies of dnDSA incidence in the unmatched sample over the follow-up period. Although the cumulative incidence tended to be lower in the higher-level group, the difference was not statistically significant (log-rank P = .169) (Figure 4).

Evidence on immunosuppressant drug levels in long-term kidney graft survivors is limited. Consequently, immunosuppression management in this population is often guided by physician experience, informed by patients’ medical history and current medical status. This study examined the association between tacrolimus trough levels during the late maintenance phase, specifically between 10 and 15 years after kidney transplant, and long-term outcomes in recipients with sustained graft function. Among patients matched for key clinical variables, no marked differences were observed in kidney graft survival or rejection-free survival between the lower-level group and higher-level group. In addition, the incidence of major adverse events, including infection, CNI toxicity, and malignancy, did not differ signi-ficantly between groups, except for sepsis, which occurred more frequently in the lower-level group (P = .049). Furthermore, in the unmatched cohort, the cumulative incidence of dnDSA formation beyond 15 years after transplant did not differ significantly between the 2 groups, although a lower incidence was observed in the higher-level group.

The absence of a marked difference in kidney graft survival and rejection-free survival between the lower-level and higher-level groups suggests that tacrolimus trough levels within this range may not be a critical determinant of long-term graft outcomes in stable recipients. By 10 years after transplant, most patients included in this study had already demonstrated long-term graft stability, which may have diminished the influence of further adjustments in immunosuppressive intensity. The inclusion of a cohort with long-term stable graft function, poten-tially representing an immunologically low-risk population, may have contributed to the lack of dif-ference observed between groups. These findings are consistent with previous reports that have indicated an attenuation of the association between tacrolimus exposure and graft outcomes over time, particularly in recipients with favorable baseline characteristics and prolonged immunological stability.¹³

In addition to the primary outcomes, secondary outcomes, including the incidence of infection, CNI toxicity, and malignancy, did not differ significantly between the lower-level group and the higher-level group. These findings suggest that modest differences in tacrolimus trough levels during the late maintenance phase may not substantially influence the develop-ment of long-term complications in recipients with stable graft function.¹³ It is possible that the relatively low event rates in this cohort limited the ability to detect subtle differences between groups.

Notably, the incidence of sepsis was significantly higher in the lower-level group, despite the general assumption that higher CNI exposure increases susceptibility to infection. One possible explanation is confounding by indication, whereby patients in the lower-level group may have undergone tacrolimus dose reduction due to prior infections or comorbidities that also predisposed them to subsequent sepsis events. This finding may also reflect immunological and pharmacokinetic dysregulation during infection, because evidence suggests that acute infections, particularly gastrointestinal infections, can alter tacrolimus metabolism and bioavailability through cytokine-mediated suppression of cytochrome P450 enzymes and P-glycoprotein activity.¹⁴ Notably, previous studies have identified acute rejection as an independent risk factor for sepsis in kidney transplant recipients, suggesting that episodes of heightened alloimmune activation may predispose patients to subsequent immune dysregulation and infection.¹⁵ Although the present study did not assess alloreactive T-cell activity or dynamic changes in immunology profiles, the observed association bet-ween lower tacrolimus trough levels and increased sepsis incidence may reflect an indirect pathway involving insufficient immunosuppression, subclinical alloactivation, and secondary vulnerability to infection. Although this finding warrants cautious interpretation, it raises important questions regarding the balance between adequate immunosuppression and infection risk in long-term transplant recipients and under-scores the need for individualized monitoring stra-tegies in this population.

Although the frequency of dnDSA development did not differ significantly between groups, the higher-level group demonstrated a trend toward a lower cumulative incidence. This tendency may reflect a continued immunological benefit of main-taining higher tacrolimus exposure, even among long-term transplant survivors. Previous studies have consistently shown that low tacrolimus levels during the early posttransplant period are associated with an increased risk of dnDSA formation; however, limited data are available regarding whether this association persists beyond the first decade after transplant. In a recent long-term study by Unagami and colleagues, no significant association was shown between tacrolimus levels maintained in the range of 4 to 6 ng/mL and dnDSA development over a 7-year follow-up, although few patients in that study had trough levels <4 ng/mL.¹⁶ Although the present fin-dings are not conclusive, the observed trend supports the hypothesis that tacrolimus exposure may continue to mitigate the risk of alloantibody formation even during the late posttransplant phase. Notably, Maggiore and colleagues recently reported that, even among ultra-long-term transplant survivors, the emergence of dnDSA remains strongly associated with graft failure, reinforcing the clinical relevance of dnDSA monitoring regardless of time since transplant.¹⁷

In this retrospective study of long-term kidney transplant recipients with functioning grafts beyond 15 years, tacrolimus trough levels maintained between 10 and 15 years after transplant were not significantly associated with kidney graft survival, rejection-free survival, or the incidence of most major complications. However, a numerically lower cumulative incidence of dnDSA formation was observed in the higher-level group, suggesting that tacrolimus exposure may continue to play a role in long-term alloimmune control, even among clinically stable recipients. These findings underscore the potential relevance of dnDSA monitoring and tacrolimus exposure opti-mization in long-term transplant care. Further prospective studies are warranted to clarify the immunological implications of tacrolimus dosing during the late maintenance phase and to guide individualized immunosuppression strategies beyond the first decade after transplant.

Despite advances in transplant immunology, substantial uncertainty remains regarding how best to balance long-term immunosuppression and immune surveillance in stable recipients. Existing studies have demonstrated that high intrapatient variability (IPV) in tacrolimus trough levels is associated with an increased risk of dnDSA formation and graft loss.^18-20 However, the effect of cumulative tacrolimus exposure and individualized immunolo-gical risk profiles on long-term outcomes remains unclear and warrants further investigation. Future research should focus on refining risk stratification methods and explore integrative monitoring tools, such as pharmacodynamic biomarkers and longi-tudinal immune profiling, to support more precise and durable immunosuppression strategies in long-term transplant recipients. In particular, the influence of tacrolimus variability, cumulative exposure, and individual immune risk profiles on late alloimmune events, including dnDSA formation, warrants further investigation.

This study had several strengths. First, we included a well-defined cohort of kidney transplant recipients who maintained graft function for >15 years, allo-wing evaluation of long-term outcomes in a uniquely stable population. Second, the use of propensity score matching helped reduce confounding by balancing key baseline characteristics between groups. Additionally, detailed clinical data on immunosup-pression protocols, graft function, rejection history, and infection status were available, supporting robust analyses.

However, the study also had several limitations. Its retrospective, single-center design may limit the generalizability of the findings. Although propensity score matching was performed, residual confoun-ding from unmeasured variables cannot be excluded. For example, infection history prior to 15 years after transplant was not included as a matching factor, which may have influenced both tacrolimus dosing and subsequent infection risk, including sepsis. The sample size, particularly after matching, was rela-tively small, and may have limited the statistical power to detect subtle differences in rare outcomes. Furthermore, because the study period included patients who received transplants before the imple-mentation of electronic medical records, some variables, such as detailed HLA mismatch data, were missing or inconsistently documented. This limita-tion may have affected the completeness of risk stratification and the robustness of propensity score adjustment. Finally, tacrolimus exposure was assessed based on trough levels, which may not fully capture individual pharmacodynamic variability over time. Given that high IPV has been associated with increased risk of acute rejection, dnDSA formation, and poor graft outcomes in previous studies, the absence of IPV data may have limited the ability to fully evaluate the effect of tacrolimus exposure consistency on long-term clinical outcomes.

Conclusions

In this retrospective study of long-term kidney transplant recipients with functioning grafts of >15 years, tacrolimus trough levels maintained between 10 and 15 years after transplant were not significantly associated with kidney graft survival, rejection-free survival, or the incidence of most major complications. These findings suggest that modest differences in tacrolimus exposure within the lower therapeutic range may not substantially influence long-term outcomes in clinically stable recipients. However, a trend toward lower dnDSA incidence in the higher-level tacrolimus group highlights a potential role for ongoing immunological modulation in this population. Overall, these findings support the importance of individualized immunosuppression strategies and continued dnDSA monitoring in long-term transplant care. Further prospective studies are warranted to define the optimal level of tacrolimus exposure that balances efficacy and safety during the late posttransplant period.

References:

1. Bentata Y. Tacrolimus: 20 years of use in adult kidney transplantation. What we should know about its nephrotoxicity. Artif Organs. 2020;44(2):140-152. doi:10.1111/aor.13551
   CrossRef - PubMed
2. Kershner RP, Fitzsimmons WE. Relationship of FK506 whole blood concentrations and efficacy and toxicity after liver and kidney transplantation. Transplantation. 1996;62(7):920-926. doi:10.1097/00007890-199610150-00009
   CrossRef - PubMed
3. Davis S, Gralla J, Klem P, et al. Lower tacrolimus exposure and time in therapeutic range increase the risk of de novo donor-specific antibodies in the first year of kidney transplantation. Am J Transplant. 2018;18(4):907-915. doi:10.1111/ajt.14504
   CrossRef - PubMed
4. Gaynor JJ, Ciancio G, Guerra G, et al. Lower tacrolimus trough levels are associated with subsequently higher acute rejection risk during the first 12 months after kidney transplantation. Transpl Int. 2016;29(2):216-226. doi:10.1111/tri.12699
   CrossRef - PubMed
5. Han A, Jo AJ, Kwon H, et al. Optimum tacrolimus trough levels for enhanced graft survival and safety in kidney transplantation: a retrospective multicenter real-world evidence study. Int J Surg. 2024;110(10):6711-6722. doi:10.1097/JS9.0000000000001800
   CrossRef - PubMed
6. Bererhi L, Pallet N, Zuber J, et al. Clinical and immunological features of very long-term survivors with a single renal transplant. Transpl Int. 2012;25(5):545-554. doi:10.1111/j.1432-2277.2012.01451.x
   CrossRef - PubMed
7. Diaz LS, Fraile MG, Lopez ACA, Fernandez MLS, Cano NR, Diaz-Corte C. More than 25 years together: basis for a long-lifespan kidney transplant. Transplant Proc. 2019;51(2):334-336. doi:10.1016/j.transproceed.2018.10.007
   CrossRef - PubMed
8. Dujardin A, Chesneau M, Dubois F, et al. Clinical and immunological follow-up of very long-term kidney transplant recipients treated with calcineurin inhibitors indicates dual phenotypes. Kidney Int. 2021;99(6):1418-1429. doi:10.1016/j.kint.2020.09.036
   CrossRef - PubMed
9. McCaughan JA, Courtney AE. The clinical course of kidney transplant recipients after 20 years of graft function. Am J Transplant. 2015;15(3):734-740. doi:10.1111/ajt.13041
   CrossRef - PubMed
10. Nascimento E, Lucas-Junior FM, Fabreti-Oliveira RA, et al. Kidney transplantation with ultralong-term (42 years) survival of a 100-year-old graft. Transplant Proc. 2016;48(9):3079-3084. doi:10.1016/j.transproceed.2016.02.064
    CrossRef - PubMed
    
    
11. Reimann AV, Nilsson J, Wuethrich RP, Mueller TF, Schachtner T. Entering the third decade after kidney transplantation: excellent graft function refers to superior graft but not patient survival. Transpl Int. 2022;35:10675. doi:10.3389/ti.2022.10675
    CrossRef - PubMed
12. Poggio ED, Augustine JJ, Arrigain S, Brennan DC, Schold JD. Long-term kidney transplant graft survival: making progress when most needed. Am J Transplant. 2021;21(8):2824-2832. doi:10.1111/ajt.16463
    CrossRef - PubMed
13. Jung HY, Seo MY, Jeon Y, et al. Tacrolimus trough levels higher than 6 ng/mL might not be required after a year in stable kidney transplant recipients. PLoS One. 2020;15(7):e0235418. doi:10.1371/journal.pone.0235418
    CrossRef - PubMed
14. Percy C, Hassoun Z, Mourad M, et al. Impact of acute infection requiring hospitalization on tacrolimus blood levels in kidney transplant recipients. Transplant Proc. 2017;49(9):2065-2069. doi:10.1016/j.transproceed.2017.09.019
    CrossRef - PubMed
15. Schachtner T, Stein M, Reinke P. Sepsis after renal transplantation: clinical, immunological, and microbiological risk factors. Transpl Infect Dis. 2017;19(3). doi:10.1111/tid.12695
    CrossRef - PubMed
    
16. Unagami K, Ishida H, Furusawa M, et al. Influence of a low-dose tacrolimus protocol on the appearance of de novo donor-specific antibodies during 7 years of follow-up after renal transplantation. Nephrol Dial Transplant. 2021;36(6):1120-1129. doi:10.1093/ndt/gfaa258
    CrossRef - PubMed
17. Maggiore U. Twenty years, and more to come: learning what makes some transplants ultra-long survivors. Transpl Int. 2022;35:11036. doi:10.3389/ti.2022.11036
    CrossRef - PubMed
18. Rodrigo E, Segundo DS, Fernandez-Fresnedo G, et al. Within-patient variability in tacrolimus blood levels predicts kidney graft loss and donor-specific antibody development. Transplantation. 2016;100(11):2479-2485. doi:10.1097/TP.0000000000001040
    CrossRef - PubMed
19. Shuker N, Shuker L, van Rosmalen J, et al. A high intrapatient variability in tacrolimus exposure is associated with poor long-term outcome of kidney transplantation. Transpl Int. 2016;29(11):1158-1167. doi:10.1111/tri.12798
    CrossRef - PubMed
20. Whalen HR, Glen JA, Harkins V, et al. High intrapatient tacrolimus variability is associated with worse outcomes in renal transplantation using a low-dose tacrolimus immunosuppressive regime. Transplantation. 2017;101(2):430-436. doi:10.1097/TP.0000000000001129
    CrossRef - PubMed