CYP3A5 genotype affects time to therapeutic tacrolimus level in pediatric kidney transplant recipients

Abstract

Background:

Optimal management of immunosuppression in kidney transplantation requires a delicate balance of efficacy and toxicity. Tacrolimus (TAC) dose requirements are significantly impacted by genetic variation in CYP3A5 polymorphisms, however the impact that genotype has on clinical outcomes in the pediatric kidney transplant population remains unclear.

Methods:

We evaluated a retrospective cohort of 98 pediatric kidney transplant recipients. The primary exposure was CYP3A5 genotype, which classified each recipient into the expresser (at least one CYP3A5*1 allele) or non-expresser group (only CYP3A5*3 alleles). The primary outcome was time to achieve a steady therapeutic TAC concentration. Secondary outcomes include incidence of early allograft rejection and calcineurin inhibitor (CNI) nephrotoxicity during the first year post-transplant.

Results:

The study cohort included 55 (56%) expressers and 43 (44%) non-expressers of the CYP3A5*1 allele. Expressers had a significantly longer time to achieve a steady therapeutic TAC concentration than non-expressers (log rank, P = 0.03). Expressers had a trend for higher incidence of early allograft rejection (29.1% vs 16.3%, log rank, P = 0.16). Early biopsy-proven CNI nephrotoxicity was seen in 60% of recipients, with no differences in the rate between expressers and non-expressers.

Conclusions:

Pediatric kidney transplant recipients with the CYP3A5*1 allele (expressers) take a longer time to achieve therapeutic TAC levels than those with the CYP3A5*3 allele (non-expressers). However, we observed no significant differences in acute rejection or CNI nephrotoxicity based on CYP3A5 genotype. Thus CYP3A5 genotype was not observed to have an immediate impact on early transplant outcomes.

1 |. INTRODUCTION

Kidney transplant is the best treatment for children with ESRD, providing better survival and quality of life in comparison with dialysis.¹ Optimal long-term management of kidney transplant patients requires a balance between prevention of allograft rejection and medication toxicities, such as opportunistic infections and nephrotoxicity. The first-line immunosuppression regimen for pediatric kidney recipients is often a combination of a CNI and antimetabolite, with or without corticosteroids.² CNIs, such as TAC, are started after kidney transplant at a standard weight-based dose and then titrated to a desired target concentration using therapeutic drug monitoring. TAC levels are highly variable in the initial post-transplant period, and some evidence suggests that higher early TAC levels are associated with decreased risk of acute rejection.^3,4 While target concentrations provide a standardized framework for dosing, they do not guarantee optimal immunosuppression efficacy or lack of toxicity. Furthermore, optimal immunosuppression may vary between patients and within a patient over time post-transplant.

Recent studies have sought to understand how to personalize kidney transplant immunosuppression with the aim of improving clinical outcomes. Genetic variants in enzyme activity are one method of personalization, since they can significantly affect drug metabolism and immunosuppressant levels.^5–12 TAC is metabolized by the CYP450 system and is largely metabolized by the CYP3A family of enzymes.¹¹ Studies have demonstrated higher TAC dose requirements to reach a target concentration in patients who carry the CYP3A5*1 allele (active enzyme, expressers) vs patients with the CYP3A5*3 allele (inactive enzyme, non-expressers).^6,11–14 In addition, the time to achieve a therapeutic TAC level was prolonged in adult kidney transplant recipients with the CYP3A5*1 allele.^5,13 A prospective study in adults demonstrated that patients required fewer dose adjustments and that therapeutic levels were achieved more rapidly in 75% of patients when the dosing regimen was based on CYP3A5 genotype.⁶ Similarly, one study in a Hispanic pediatric population found that CYP3A5 expressers were less likely to have a TAC trough level of 7 ng/mL or higher compared to non-expressers at 4 weeks post-transplant.¹⁵

Despite the influence of CYP3A5 genotype on TAC dose requirements, and guidelines for dosing TAC based on CYP3A5 genotype,¹⁶ studies have not conclusively shown a consistent association between CYP3A5 genotype and clinical outcomes such as rate of acute rejection^5,6,17–19 or CNI nephrotoxicity.^18–22 Previous studies have also had limitations including a low proportion of Black recipients,^5,17,18 short follow-up periods,^6,15 or lack of surveillance biopsy data to accurately assess acute rejection and CNI nephrotoxicity.²⁰ Thus, how genotype should be used to influence clinical practice remains unclear.

Most of the previous studies of pharmacogenomics in transplant immunosuppression have focused on adults. Findings regarding the impact of pharmacogenomic variants on clinical outcomes in adults may not extrapolate well to pediatric transplant recipients. Factors which are unique to the pediatric transplant recipient and may impact the optimal level of immunosuppression include a more naïve and robust immune system, seronegativity for viral infections (ie, EBV, CMV, BKV), variable drug metabolism with development and maturation, and liquid formulations of medications.^23,24 Improving renal allograft longevity is particularly important in pediatric kidney transplant recipients who are expected to outlive their primary renal allograft and require excellent allograft function to promote healthy growth and development. Herein, we sought to understand the effect of CYP3A5 genotype on clinical outcomes in pediatric kidney transplant recipients. We hypothesized that CYP3A5 genotype significantly impacts time to achieve therapeutic TAC level, allograft rejection rates, and CNI nephrotoxicity in pediatric kidney recipients during the first post-transplant year.

2 |. METHODS

2.1 |. Study population

We evaluated pediatric patients who received a first kidney transplant at COA/UAB (COA/UAB) from 2004 to 2014. Eligible patients were identified from the list of kidney transplant recipients maintained by the program administration. Patients were excluded if they received a dual organ transplant, had a previous kidney transplant, or received maintenance immunosuppressants other than TAC, antimetabolites, or steroids immediately after transplant. The study was approved by the Institutional Review Board at UAB (X150803005, 9/28/2015) and was conducted with adherence to ethical principles for research consistent with the Declaration of Helsinki.

We obtained whole-blood samples for DNA extraction during follow-up visits in the COA and UAB outpatient kidney transplant clinics from 2015 to 2016. Informed consent/assent was obtained for the collection of blood samples at follow-up visits. For patients who were no longer actively followed in the COA and UAB transplant clinics, remnant DNA samples were obtained from our histocompatibility laboratory if available. A waiver of consent was obtained for the use of remnant samples. Demographic information and clinical information for the first post-transplant year were collected retrospectively from the medical record.

We reviewed factors known to impact TAC dosing, including presence of delayed graft function (need for dialysis in the first 7 days post-transplant) and post-op day of TAC initiation. Medication lists at the time of hospital discharge were reviewed for potential interactions. The following medications which may have clinically relevant interactions with TAC were included²⁵: antibiotics (azithromycin, chloramphenicol, ertapenem, erythromycin, fusidic acid, levofloxacin), antifungals (caspofungin, clotrimazole, fluconazole, isavuconazonium sulfate, itraconazole, ketoconazole, posaconazole, voriconazole), antivirals (asunaprevir, efavirenz, foscarnet, glecaprevir and pibrentasvir, grazoprevir, ledipasvir, letermovir, nelfinavir, ombitasvir, paritaprevir, ritonavir, and dasabuvir, ritonavir, saquinavir, telaprevir, other protease inhibitors), antiepileptics (fosphenytoin, phenytoin, stiripentol), antineoplastic agents (ceritinib, crizotinib, dabrafenib, dasatinib, enzalutamide, idelalisib, mitotane, palbociclib, siltuximab, temsirolimus), non-dihydropyridine calcium channel blockers (diltiazem, verapamil), proton pump inhibitors (esomeprazole, lansoprazole, omeprazole), and others (bosentan, cinacalcet, conivaptan, danazol, deferasirox, dronedarone, fosaprepitant, mifeptristone, netupitant, ranolazine, sarilumab, sevelamer, tocilizumab).

The standard immunosuppression protocol during the study period included non-depleting induction with daclizumab or basiliximab, and triple maintenance therapy with TAC, mycophenolate, and prednisone. Pneumocystis and CMV prophylaxis were provided with trimethoprim-sulfamethoxazole and valganciclovir, respectively. Surveillance allograft biopsies were performed at 3 and/or 6 months post-transplant as standard of care. Allograft biopsies were also performed when indicated for allograft dysfunction and suspicion of rejection. TAC levels were drawn daily while inpatient immediately post-transplant and at all outpatient visits. TAC target levels were 12–14 ng/mL during the first week post-transplant, 10–12 ng/mL during weeks 2–7 post-transplant, 7–10 ng/mL during weeks 8–11 post-transplant, and 5–7 ng/mL at 12 weeks post-transplant and beyond. TAC dose adjustments were made at the discretion of the pediatric nephrologist to achieve these targets. CYP3A5 genotype was not known at the time of transplant or at subsequent clinic visits during the study period.

2.2 |. TAC levels

During the study period, our laboratory used the Abbott IMx (2004–2010) and Abbott ARCHITECT i1000 (2010–2014) TAC assays. Internal validation studies showed a tight correlation between TAC levels obtained with the two methods (R = 0.9885).

2.3 |. DNA samples and genotyping

DNA extraction from whole-blood samples was performed by the UAB Sample Processing and Analytic Nexus laboratory in the Center for Clinical and Translational Science using the Gentra PureGene system. Residual DNA samples from UAB’s histocompatibility laboratory were obtained from long-term storage at −80°C. PCR amplification and DNA sequencing for the CYP3A5 single nucleotide polymorphism (rs776746) were performed at the UAB Heflin Center for Genomics using standard approaches.

2.4 |. Variable definitions

The primary exposure was CYP3A5 genotype, with “expressers” defined as patients with at least one CYP3A5*1 allele (*1/*1 or *1/*3) and “non-expressers” as patients with only the CYP3A5*3 allele. The primary outcome was time to steady therapeutic TAC level, defined as the number of days between initiation of TAC to the second consecutive TAC level within the target range. Secondary outcomes included biopsy-proven acute rejection and CNI nephrotoxicity. Acute rejection was diagnosed by allograft biopsy according to the most updated Banff criteria at the time of each biopsy. CNI nephrotoxicity was diagnosed by allograft biopsy by the presence of striped fibrosis, arteriolar hyalinosis, or tubular vacuolization.

2.5 |. Statistical methods

Demographic and clinical covariates were compared between the expresser and non-expresser groups, using t tests for normally distributed continuous variables, Wilcoxon rank-sum tests for non- normally distributed variables, and chi-square or Fisher’s exact tests for categorical variables. Time-to-event analyses were performed for the primary and secondary outcomes using Kaplan-Meier methods and log-rank tests at a significance level of α = 0.05. Patients were censored at the last available follow-up visit. Additional exploratory outcomes included development of allograft inflammation (borderline changes or acute rejection), BK nephropathy, and eGFR by the CKD-Schwartz Equation²⁶ at 1 year post-transplant or last follow-up. Comparisons for exploratory outcomes were made using Wilcoxon rank-sum tests for non-normally distributed continuous variables and chi-square or Fisher’s exact tests for categorical variables.

3 |. RESULTS

We identified 179 eligible children who received a primary kidney transplant during the study period from 2004 to 2014. Of these, we were able to obtain DNA samples from 98 subjects. Residual DNA samples were obtained from UAB’s histocompatibility laboratory in 60 subjects; DNA was extracted from fresh whole-blood samples during routine visits in 38 patients. Patients not included in the study were slightly older at the time of transplant compared to included patients (median age 14 vs 12), but there were no significant differences in donor type, gender, or race. The study population included 55 (56%) expressers and 43 (44%) non-expressers of the CYP3A5*1 allele. Demographic information and clinical information for the expresser and non-expresser groups are displayed in Table 1. Compared with non-expressers, expressers were older at the time of transplant, had a higher proportion of Black race, and a higher proportion of deceased donor type. There were no statistically significant differences in etiology of ESRD, insurance type, or proportion of patients on dialysis prior to transplant (see Table 1). There were 4 patients in the expresser group that experienced delayed graft function, defined as need for dialysis in the first 7 days post-transplant, but this difference was not statistically significant (Table 1).

TABLE 1.

Demographic and clinical information by CYP3A5 genotype (n = 98)

Variables	Non-expressers (CYP3A5*3/*3) n = 43	Expressers (CYP3A5*1/*3, *1/*1) n = 55	P-Value*
Age at transplant, years, mean (SD)	7.8	11.5	<0.01
Race, n (%)
Black (Hispanic 0%)	1 (2.3%)	46 (83.6%)	<0.01
Non-black (Hispanic 6%)	42 (97.7%)	9 (16.4%)
Gender, n (%)
Male	27 (62.8%)	34 (61.8%)	0.92
Female	16 (37.2%)	21 (38.2%)
Insurance type, n (%)
Private	17 (39.5%)	15 (27.3%)	0.40
Government (Medicaid, Tricare)	19 (44.2%)	31 (56.4%)
Unknown	7 (16.3%)	9 (16.4%)
ESRD category, n (%)
Glomerular	13 (30.2%)	19 (34.6%)	0.88
CAKUT	20 (46.5%)	23 (41.8%)
Other/Unknown	10 (23.3%)	13 (23.6%)
Pretransplant dialysis, n (%)
Yes	28 (65.1%)	43 (78.2%)	0.15
No	15 (34.9%)	12 (21.8%)
Donor type, n (%)
Living	28 (65.1%)	13 (23.6%)	<0.01
Deceased	15 (34.9%)	42 (76.4%)
Pretransplant EBV status, n (%)
Positive	12 (27.9%)	37 (68.5%)	<0.01
Negative	31 (72.1%)	17 (31.5%)
Pretransplant CMV status, n (%)
Positive	12 (27.9%)	19 (34.5%)	0.48
Negative	31 (72.1%)	36 (65.5%)
Delayed graft function, n (%)
Yes	0 (0%)	4 (7.27%)	0.13
No	43 (100%)	51 (92.7%)
Length of follow-up, days, mean (SD)	354 (76)	366 (39)	0.37

^*Obtained using t tests for normally distributed continuous variables, and chi-square or Fisher’s exact tests for categorical variables.

All patients received non-depleting induction with an IL-2 inhibitor except one patient in the non-expresser group who received thymoglobulin for induction (Table 2). All patients received maintenance immunosuppression with TAC and mycophenolate immediately after transplant, and >90% of the patients received maintenance steroids (Table 2). There were no significant differences in the length of follow-up or number of TAC levels between the groups (Table 2). As expected, the non-expressers achieved a higher serum TAC concentration per weight-adjusted dose than expressers (Table 2). There were 5 patients in the expresser group who had delayed initiation of TAC at post-op day 2 or later due to poor initial allograft function (Table 2). The only potential medication interaction at the time of hospital discharge was concurrent use of a proton pump inhibitor, and there was no difference in the rate of proton pump inhibitor use between groups (Table 2).

TABLE 2.

Medications and TAC level monitoring by CYP3A5 genotype (n = 98)

Variables	Non-expressers (CYP3A5*3/*3) n = 43	Expressers (CYP3A5*1/*3,*1/*1) n = 55	P-Value*
Induction medication
Daclizumab	20 (46.5%)	28 (50.9%)	0.75**
Basiliximab	22 (51.2%)	27 (49.1%)
Thymoglobulin	1 (2.3%)	0 (0%)
Maintenance steroids
Yes	39 (90.7%)	52 (94.6%)	0.70
No	4 (9.3%)	3 (5.4%)
Post-op TAC start day
0–1	43 (100%)	50 (90.9%)	0.07
≥2	0 (0%)	5 (9.1%)
TAC levels, levels/month of follow-up, median (IQR)	2.8 (2.4–4.0)	2.8 (2.2–3.8)	0.47
TAC concentration per dose, (ng/mL)/(mg/kg/d), median (IQR)	56.8 (43.0–81.0)	32.5 (23.0–44.1)	<0.01
Medication interaction at hospital discharge (PPI)
Yes	15 (34.8%)	12 (21.8%)	0.15
No	28 (65.1%)	43 (78.2%)

^*Obtained using Wilcoxon rank-sum tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables.

^**Excludes 1 patient who received thymoglobulin to allow validity of chi-square test.

Time to reach therapeutic TAC level was significantly longer in the expressers compared to the non-expressers (P = 0.03, Figure 1). The median time to reach steady therapeutic TAC level was 47 days (IQR 23–77) days in the expresser group compared to 26.5 days (IQR 10.5–49.5) in the non-expresser group (Table 3). A total of 9 patients in the overall cohort did not reach a steady TAC level within the first post-transplant year or last available follow-up, six in the expresser group and three in the non-expresser group (Table 3). Of these 9 patients, four had a follow-up time of <300 days (two expressers and two non-expressers).

FIGURE 1.

Time to steady TAC level by CYP3A5 genotype, Kaplan-Meier (log rank, P = 0.0303)

TABLE 3.

Clinical outcomes during first year post-transplant by CYP3A5 genotype (n = 98)

Variables	Non-expressers (CYP3A5*3/*3) n = 43	Expressers (CYP3A5*1/*3, *1/*1) n = 55	P-Value*
Reached steady TAC level during follow-up, n (%)
Yes	40 (93.0%)	49 (89.1%)	0.73
No	3 (7.0%)	6 (10.1%)
Time to steady TAC levela, days, median (IQR)	26 (10–49)	47 (21–76)	0.02
Acute rejection, n (%)
Yes	7 (16.3%)	16 (29.1%)	0.14
No	36 (83.7%)	39 (70.9%)
Allograft inflammationb, n (%)
Yes	14 (32.6%)	28 (50.9%)	0.07
No	29 (67.4%)	27 (49.1%)
CNI nephrotoxicityc, n (%)
Yes	23 (57.5%)	35 (63.6%)	0.54
No	17 (42.5%)	20 (36.4%)
BK nephropathyc, n (%)
Yes	3 (7.5%)	6 (10.9%)	0.73
No	37 (92.5%)	49 (89.1%)
Allograft loss, n (%)
Yes	0 (0%)	2 (3.6%)	0.50
No	43 (100%)	53 (96.4%)
eGFRd, median (IQR)	72.0 (58–91)	60.0 (47–80)	0.04

^aIncludes 89 patients who reached steady TAC level during follow-up.

^bAllorgraft inflammation includes borderline changes and acute rejection.

^cExcludes 3 non-expressers who did not have an allograft biopsy during the follow-up period.

^dExcludes 2 patients with allograft loss, 1 expresser with missing data, and 1 non-expresser with missing data.

^*Obtained using Wilcoxon rank-sum tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables.

We observed a trend for higher incidence of early acute rejection in expressers vs non-expressers (Table 3 and Figure 2) that did not reach statistical significance. We observed a similar trend for increased early allograft inflammation when biopsies with borderline changes were considered in addition to those with acute rejection (Table 3). The incidence of early CNI nephrotoxicity was high in the overall cohort (60%), but was similar between the expresser and non-expresser groups (Table 3 and Figure 3).

FIGURE 2.

Time to acute rejection by CYP3A5 genotype, Kaplan-Meier (log rank, P = 0.1634)

FIGURE 3.

Time to CNI nephrotoxicity by CYP3A5 genotype, Kaplan-Meier (log rank, P = 0.3930)

Additional analysis revealed that the expresser group had a significantly lower eGFR at last follow-up compared to non-expressers, but this finding could have been confounded by a higher proportion of deceased donors in the expresser group among other factors. No differences were seen between groups with respect to the other outcomes, including BK nephropathy and allograft loss (Table 3). There were no patient deaths during the first post-transplant year.

4 |. DISCUSSION

Genetic variants in CYP3A5 are known to significantly impact TAC dose requirements.^6,11–14 In this cohort study of pediatric kidney transplant recipients, we demonstrated that CYP3A5*1 allele expressers took significantly longer time to reach steady therapeutic TAC levels than non-expressers, with a median difference of 21 days. There were 5 patients in the expresser group who had delayed initiation of TAC on post-op day 2 or later (latest on post-op day 8); however, this was accounted for in our analysis, as time to steady therapeutic TAC level was measured from the day of TAC initiation. There were no differences between the groups in the proportion of patients using potentially interacting medications (eg, proton pump inhibitors). Our findings are similar to data from the Long-term Deterioration of Kidney Allograft Function (DeKAF) study which showed the median TAC trough in African Americans remained below 8 ng/mL and significantly lower than the median trough in White recipients over the first 6 months post-transplant.¹² The findings are also consistent with a study in Hispanic pediatric kidney transplant recipients that found CYP3A5 expressers to be less likely to have a TAC trough at or above 7 ng/mL at 4 weeks post-transplant.¹⁵

Despite the difference in time to reach steady target TAC levels, we did not detect a substantial impact on 1-year clinical outcomes, although we observed non-significant trends toward increased incidence of borderline changes and acute rejection. While expressers had significantly lower eGFR at 1 year post-transplant, this effect could have been confounded by a higher proportion of deceased donors in this group among other factors.

Most previous studies have not demonstrated a link between CYP3A5 polymorphisms and acute allograft rejection. However, due to limitations in study population and design, a degree of uncertainty remained with previous findings. Macphee et al,⁵ Hesselink et al,¹⁷ and Glowacki et al¹⁸ all reported no difference in acute rejection rates between expressers and non-expressers; however, these studies included a low proportion of Black recipients (5%−14%) and therefore a low proportion of expressers (18%−30%). Thervet et al⁶ demonstrated through a prospective, randomized trial that therapeutic TAC concentrations are achieved more quickly when dosing is based on CYP3A5 genotype, although no differences in acute rejection, delayed graft function, or allograft survival were seen. However, the study only included a short follow-up period of 3 months, and 90% of the cohort was Caucasian. Only one of the above studies included surveillance biopsies to evaluate for early subclinical inflammation. Quteineh et al¹⁹ did find a higher proportion of acute rejection in CYP3A5*1/*1 recipients, but this finding was based on a small number of *1/*1 homozygotes (13 out of 136 participants) in a predominately Caucasian European population. Our study adds to the literature in that it evaluated a pediatric population with a high proportion of Black recipients (48%), included surveillance biopsy data, and followed patients through the first post-transplant year.

Our cohort had a 60% incidence of early CNI nephrotoxicity, but the incidence did not vary significantly between expressers and non-expressers. Previous studies evaluating the relationship between CYP3A5 polymorphisms and CNI nephrotoxicity have provided conflicting results.^18–21,27 The largest of these, which evaluated 945 kidney transplant recipients from the (DeKAF) study, found no association between genetic polymorphisms and TAC-induced nephrotoxicity. However, this study relied on a clinical definition of CNI nephrotoxicity rather than histologic findings. Since surveillance biopsies were performed as the standard of care at our center, we were able to accurately assess the development of early CNI nephrotoxicity by 6 months post-transplant. However, as late surveillance biopsies were not done, it is not known whether these findings of CNI nephrotoxicity were sustained. Incidence of CNI nephrotoxicity may also vary with TAC trough targets levels, which are higher in the early post-transplant period.

Although our study had several strengths, we also recognize several limitations common to retrospective cohort studies. Although we were able to obtain DNA samples from more than half of eligible recipients, not all recipients during the time period were included. We confirmed that excluded recipients without DNA sampling had similar clinical and demographic characteristics to our study cohort. We were not able to reliably assess some clinically relevant confounders, such as non-adherence to immunosuppression therapy. There were 9 patients in the cohort that did not reach a steady TAC level because they did not have two consecutive TAC levels within the target range. These patients may have had poor medication compliance or other variations in medication administration (such as dosing with food) leading to level variability. CYP3A5 genotype varies with race, and we recognize other unmeasured factors associated with race can still influence results, including socioeconomic factors that influence access to health care and medications. Adjustments in TAC dose in response to levels were not standardized, but rather left to the provider, reflecting variability seen in clinical practice. This likely lead to increased variation in levels, although this source of variation should be randomly distributed across the CYP3A5 groups. Patients in this study received non-depleting induction and triple maintenance therapy with TAC, mycophenolate, and prednisone; thus our findings may not be generalizable to patients on alternative immunosuppression regimens. Although patients were started on the same, weight-based doses of mycophenolate and prednisone, we did not evaluate how these other immunosuppressant doses may have changed during the 1 year post-transplant follow-up period. We were not able to assess the incidence of de novo donor-specific antibody formation in this cohort, as this was not measured routinely through the study period.

CYP3A5 genotyping prior to kidney transplantation in pediatric recipients would likely aid in reducing time to achieve therapeutic TAC targets. However, given the unclear effect this may have on early clinical outcomes, and the wide availability of therapeutic drug monitoring, we are not yet able to recommend CYP3A5 genotyping for general pediatric clinical practice. Further study of how early attainment and maintenance of TAC targets influence allograft outcomes is warranted.

Abbreviations:

CKD

chronic kidney disease

CMV

cytomegalovirus

CNI

calcineurin inhibitor

COA

Children’s of Alabama

DNA

deoxyribonucleic acid

EBV

Ebstein-Barr virus

eGFR

estimated glomerular filtration rate

ESRD

end-stage renal disease

IQR

interquartile range

PCR

polymerase chain reaction

SD

standard deviation

TAC

tacrolimus

UAB

University of Alabama at Birmingham