Higher number of tacrolimus dose adjustments in kidney transplant recipients who are extensive and intermediate CYP3A5 metabolizers

Abstract

Kidney transplant recipients carrying the CYP3A5*1 allele have lower tacrolimus troughs, and higher dose requirements compared to those with the CYP3A5*3/*3 genotype. However, data on the effect of CYP3A5 alleles on post-transplant tacrolimus management are lacking.

The effect of CYP3A5 metabolism phenotypes on the number of tacrolimus dose adjustments and troughs in the first six months post-transplant was evaluated in 78 recipients (64% Caucasians). Time to first therapeutic concentration, percentage of time in therapeutic range (TTR), and estimated glomerular filtration rate were also evaluated.

Fifty-five kidney transplant recipients were CYP3A5 poor metabolizers (PM), 17 were intermediate metabolizers (IM), and 6 were extensive metabolizers (EM). Compared to PMs, EMs/IMs had significantly more dose adjustments (6.1 vs 8.1, p=0.015). Overall, 33.82% of trough measurements resulted in a dose change. There was no difference in the number of tacrolimus trough measurements between PMs and EM/IMs. The total daily tacrolimus dose requirements were higher in EMs and IMs compared to PMs (<0.001). TTR was ~50% in the PMs and EMs/IMs groups.

CYP3A5 EM/IM metabolizers have more tacrolimus dose changes and higher dose requirements which increases clinical management complexity. Larger studies are needed to assess the cost and benefits of including genotyping data to improve clinical management.

1. Introduction

Tacrolimus-based regimens are standard of care for maintenance immunosuppression for nearly every organ including the kidney. Use of tacrolimus is complex due to its narrow therapeutic index and high inter- and intra-patient variability as well as time varying exposure levels that necessitates therapeutic drug monitoring to ensure appropriate trough blood concentrations. Low concentrations increase the risk of rejection and high concentrations increase the risk of toxicity. In recent years, high intra-patient variability has been associated in some studies with a greater risk of adverse graft outcomes^1–5.

Variability in tacrolimus trough concentrations among patients is associated with non-adherence; interactions with food, other drugs, disease; and genetic variation⁶. Tacrolimus is extensively metabolized by the cytochrome P450 (CYP) 3A5 enzyme which is a highly polymorphic gene. Common CYP3A5 variant alleles such as CYP3A5*3, *6, and *7 are loss of function (LoF) alleles encoding for a nonfunctional protein. Most Caucasians carry one or two CYP3A5*3 alleles. African Americans are much less likely to carry CYP3A5 LoF alleles and as a result have significantly greater clearance of tacrolimus, lower trough concentrations, and higher dose requirements than Caucasians^7–10. While CYP3A5 genotype has reliably demonstrated effect on metabolism and dose requirements of tacrolimus, it has inconsistently been associated with estimated glomerular filtration rate (eGFR), acute rejection, and graft survival^11–13. Given the strong association between genetic variation and dose the Clinical Pharmacogenetics Implementation Consortium (CPIC) has published guidelines for tacrolimus dosing based on CYP3A5 phenotypes¹⁴.

The impact of CYP3A5 phenotype on clinical management of tacrolimus (such as utilization of healthcare resources) has been poorly studied. The primary objectives of this study were to evaluate if CYP3A5 metabolism phenotypes were associated with the number of tacrolimus trough concentrations measured and dose adjustments made to achieve trough concentrations in the therapeutic range in the first 6 months posttransplant.

2. Material and Methods

2.1. Study Population

Study participants were kidney transplant patients enrolled from Hennepin Healthcare in the National Institutes of Health funded Genomics of Transplantation (GEN03) multi-center, prospective, observational, genome wide association study (GWAS) between 2013 and 2016^8,15–17. Participants for the current study were selected from the GEN03 participants and were included if they received de novo immediate release tacrolimus-based maintenance immunosuppression and had CYP3A5 genetic information available from the GWAS. Patients were excluded if they were age < 18 years at time of enrollment into GEN03, were switched from tacrolimus to alternative therapy within 6 months of transplant, had incomplete genotype data available for the CYP3A5 alleles *3, *6, and *7, or had incomplete 6-month follow-up data within the electronic medical record (EMR). This study was approved by the local institutional review board (IRB #12–3506).

2.2. Data Collection

Data were obtained from the GEN03 study database with supplemental data collected from the EMR for the first 6 months post-transplant. Data obtained prospectively through the GEN03 GWAS study database included demographics, CYP3A5 genotypes, acute rejection, and serum creatinine at 1, 3, and 6 months post-transplant. Supplemental data collected to evaluate effect on clinical management included all measured tacrolimus trough blood concentrations, dose at time of trough measurement, if dose adjustments were made, timing of trough concentration relative to dose, and goal trough range. Due to the difficulty in consistently identifying longitudinal drug-drug interactions through review of medical records, concomitant drug-drug interactions were not recorded.

2.3. Genotype Data

Three common CYP3A5 LoF genetic variants [CYP3A5*3 (rs776746), CYP3A5*6 (rs10264272), and CYP3A5*7 (rs41303343)] were taken from the GWAS chip for each participant and used in this analysis¹⁵. Genotype data were then used to create CYP3A5 metabolism phenotypes which were assigned based on the CPIC guidelines for CYP3A5 genotype and tacrolimus dosing⁷. Patients with any two LoF alleles were designated poor metabolizers (PM), with any one LoF allele were designated intermediate metabolizers (IM), and with no LoF alleles were designated extensive metabolizers (EM).

2.4. Immunosuppression Management

Patients received maintenance immunosuppression of tacrolimus, mycophenolate mofetil, and corticosteroids tapered to a maintenance dose of prednisone 5 mg daily over 14 to 90 days. Patients at high risk for rejection were initiated at a dose of tacrolimus 0.15 mg/kg/day based on ideal body weight with a tacrolimus trough concentration target of 8–12 ng/mL for the first three months followed by 5–10 ng/mL lifelong thereafter. High risk determination was based on at least two of the following risk factors: complete 6 antigen HLA mismatch, cytotoxic panel reactive antibody >20%, African American race, age < 40 years, donor specific antibody (DSA) deemed clinically relevant, and re-transplantation. Patients with standard or low risk of rejection were initiated at a dose of tacrolimus 0.1 mg/kg/day based on ideal body weight with a lifelong tacrolimus trough concentration target of 4–8 ng/mL. Antibody induction was antithymocyte globulin or basiliximab, depending on rejection risk.

2.5. Outcomes and Statistical Analysis

Participant baseline characteristics and clinical outcomes were summarized descriptively in Table 1 using means for continuous variables and frequency counts with percentage for categorical variables. The primary analysis was the association of CYP3A5 metabolism phenotypes with number of tacrolimus trough concentrations measured and number of tacrolimus dose adjustments for any reason made in the first 6 months posttransplant. The number of tacrolimus trough measurements was defined as the total number of trough concentrations reported in the EMR and determined to be appropriately obtained (within 11–13 hours of previous dose with a documented collection time). The number of tacrolimus dose adjustments was defined as all dose increases or decreases, and was determined through review of pharmacy records in the EMR. Secondary endpoints were time to first therapeutic trough concentration, percentage of time in therapeutic range (TTR), and eGFR at 6 months. Time to first therapeutic trough concentration was defined as first trough measurement within goal range from the time of kidney transplant. TTR was calculated using the Rosendaal method¹⁸. Trough measurements reported as < 2 ng/mL were imputed as 1.99 ng/mL. eGFR was estimated using Modification of Diet in Renal Disease (MDRD) equation¹⁹. The association between our primary endpoints (number of tacrolimus trough concentrations and number of tacrolimus dose adjustments) and CYP3A5 metabolism phenotypes was evaluated using quasi-likelihood based Poisson regression adjusted for age at transplant, gender, body mass index (BMI) at baseline, history of diabetes at baseline and initial tacrolimus dose. The EM and IM phenotype groups were combined and compared to the PM group.

Table 1.

Baseline Recipient Characteristics by CYP3A5 Metabolism Phenotype

	Poor Metabolizer (n = 55)	Intermediate Metabolizer (n = 17)	Extensive Metabolizer (n = 6)
N (%) Male	30 (54.5)	11 (64.7)	5 (83.3)
Age at Transplant, y (mean + SD)	50.9 (14.2)	46 (14.5)	44.5 (10.7)
BMI at Transplant (mean + SD)	28 (5.9)	28.8 (5.5)	31.1 (6.2)
N (%) Caucasian	44 (80)	6 (35.3)	0 (0)
N (%) African American	4 (7.2)	6 (35.3)	3 (50)
N (%) Asian	2 (3.6)	4 (23.6)	2 (33.3)
N (%) American Indian or Alaska Native	2 (3.6)	1 (5.8)	0 (0)
N (%) race not reported	3 (5.4)	0 (0)	1 (16.7)
Cause of Kidney Disease, n (%)
Congenital	11 (20)	1 (5.9)	0
Diabetes	12 (21.8)	2 (11.8)	1 (16.7)
Glomerular Disease	13 (23.6)	9 (52.9)	3 (50)
Hypertension	4 (7.3)	2 (11.8)	1 (16.7)
Polycystic Kidneys	8 (14.5)	1 (5.9)	0
Other/Unknown	7 (12.3)	2 (11.8)	1 (16.7)
Deceased Donor, n (%)	24 (43.6)	6 (35.3)	5 (83.3)
Re-Transplant, n (%)	9 (16.4)	1 (5.9)	1 (16.7)
Biopsy-proven Acute Rejection at 12 Months, n (%)	2 (3.6)	0	0
Graft Failure at 12 months, n (%)	2 (3.6)	0	0

For our secondary endpoints, the association between TTR and eGFR with CYP3A5 metabolism phenotypes was evaluated using multivariable linear regression analysis with TTR and eGFR modeled as continuous variables. The association between time to first therapeutic trough and CYP3A5 metabolism phenotype was evaluated using a Cox regression analysis. Co-variates for the linear and Cox regression models were the same as those used in the primary analysis. The association between dose normalized trough concentrations [tacrolimus trough (ng/mL) / total daily dose (mg)] over time and CYP3A5 metabolism phenotypes was evaluated using mixed effect regression model. To account for the change in trough concentrations over time in our mixed effect regression model, we included time in an interaction term with the binary phenotype (EM/IM compared to PMs) and as a random coefficient. All analyses were performed using STATA 14.

3. Results

Ninety-three patients were enrolled into the GEN03 GWAS study at Hennepin Healthcare and evaluated for inclusion. Five patients were excluded because they received a de novo cyclosporine-based regimen, eight were excluded for incomplete genotype information, and two were excluded for incomplete follow-up. A total of 78 patients were included in this analysis.

Baseline characteristics are described in Table 1. Fifty-five patients (70.5%) were PMs, seventeen patients (21.8%) were IMs, and six patients (7.7%) were EMs. Multiple tacrolimus trough measurements were collected per recipient and a total of 1533 were identified in the 6 months after kidney transplant. After excluding trough measurements without documented collection time after previous dose or where the previous dose was not taken 11–13 hours prior to sampling, 1425 troughs (93%) were available for evaluation (mean of 18.2 troughs per individual).

The number of trough measurements and dose adjustments by metabolism phenotype and association analyses are shown in Table 2. Overall 33.82% (482/1425) of trough measurements resulted in a dose change. Patients with the CYP3A5 EM/IM phenotypes had 1.28 times (95% CI: 1.05; 1.56, p=0.015) more tacrolimus dose adjustments when compared to PMs. EM/IM and PM phenotypes had on average 8.1 and 6.1 dose adjustments, respectively. Patients with CYP3A5 EM/IM phenotypes did not have more tacrolimus trough measurements compared to PMs (p=0.66).

Table 2.

Outcomes by CYP3A5 Metabolism Phenotype

Outcomes	Mean (95% CI)	Relative Risk (95% CI)	P-value ‡
Number of tacrolimus trough measurements †
Poor Metabolizers	19.5 (18.3 – 20.7)	Reference	Reference
Intermediate & Extensive Metabolizers	20.2 (17.7 – 22.7)	0.97 (0.86 – 1.09)	0.66
Number of tacrolimus dose adjustments †
Poor Metabolizers	6.1 (5.4 – 6.8)	Reference	Reference
Intermediate & Extensive Metabolizers	8.1 (6.8 – 9.1)	1.28 (1.05 – 1.56)	0.015
Time to first therapeutic trough in days †
	Median (Range)	Hazard Ratio (95% CI)	P-value
Poor Metabolizers	6 (2–61)	Reference	Reference
Extensive & Intermediate metabolizers	5 (2–31)	0.86 (0.49 – 1.49)	0.59
Time in Therapeutic Range, % †
	Mean (95% CI)	Percent Difference (95% CI)	P-value
Poor Metabolizers	49.7 (12.4 – 94)	Reference	Reference
Extensive & Intermediate metabolizers	51.02 (44.8 – 57.2)	0.02 (−0.08 – 0.11)	0.72
eGFR at 6 months, mL/min/1.73m 2
	Mean (SD)	Difference in eGFR level (95% CI )	P-value
Poor Metabolizers	60 (20.3)	Reference	Reference
Extensive & Intermediate metabolizers	58.8 (18.1)	1.18 (−8.71 – 11.06)	0.81
Dose-normalized tacrolimus troughs [ng/ml per mg tacrolimus daily dose]
	Mean (SD)	Change in dose-normalized tacrolimus trough (95% CI)	P-value
Poor Metabolizers	2.05 (1.67)	Reference	Reference
Extensive & Intermediate metabolizers	0.84 (0.53)	−0.57 (−1.01 ; −0.13)	0.012

^†adjusted for age, gender, BMI, diabetes at time of transplant, and initial tacrolimus dose,

^‡p-values are a comparison to the reference group (poor metabolizers)

For our secondary outcomes, we found that CYP3A5 metabolism phenotypes were not associated with time to first therapeutic trough, TTR, and eGFR at 6 months (Table 2). The time to first therapeutic trough was not different between CYP3A5 EM/IM and PM phenotypes [hazard ratio (HR) = 0.86 (0.49; 1.49, p=0.59)]. Both CYP3A5 metabolism phenotype groups had ~50% TTR and there was no difference between the groups (51.4% vs 49.7%, p=0.72). The eGFR at 6 months was not different between the EM/IM and PMs groups (p=0.81). Trough concentrations were not different across the phenotypes (Supplemental Table 1). EM and IMs had significantly higher dose requirements and lower dose-normalized troughs relative to the PMs (Supplemental Table 1 and 2).

4. Discussion

This study identified that the number of tacrolimus dose adjustments were associated with the CYP3A5 metabolism phenotype in adult kidney transplant recipients. There were significantly more dose adjustments in the EM/IM phenotype group compared to the PM group in the first 6 months post transplant (8.1 vs 6.1 respectively, p=0.015). This is consistent with expectations since most transplant protocols use a standardized tacrolimus dose which has been designed and refined for the PM phenotype (i.e. CYP3A5 non-expressers) ²⁰, and several studies have previously reported an association between the inclusion of CYP3A5 phenotype information and improved clinical outcomes and management of kidney transplant recipients ^21,22. A randomized controlled trial of 280 kidney transplant recipients found that participants who received a tacrolimus dose based on their CYP3A5 phenotype required fewer dose adjustments²³. Another randomized controlled trial of 63 kidney transplant recipients found that the genotype-guided group had an increased proportion of participant reaching therapeutic tacrolimus concentrations within the first few days post transplant²¹. In a retrospective cohort study of 125 adult kidney transplant recipients, those who had a CYP3A5 EM/IM phenotype had higher total hospitalization costs compared to PM²². A recent study in 85 pediatric kidney and heart transplant recipients also found that CYP3A5 phenotype was associated with greater healthcare resources utilization (such as number of dose adjustment and tacrolimus troughs) among heart transplant recipient, but not kidney transplant recipients²⁴. However, other studies reported that genotype based tacrolimus dosage did not decrease the incidence of long term clinical outcomes such as the incidence of biopsy-proven acute rejection and graft survival in kidney transplant recipients^21,23,25.

While it may seem that increased number of dose adjustments is a minor issue, each dose adjustment represents an opportunity for a medication error. Dose adjustments are generally communicated electronically or by phone which may be forgotten, missed, or misunderstood by patients. Frequent medication changes after transplant have been shown to affect medication adherence^26,27. A review of lung transplant recipients found that 74% of medication errors stemmed from improperly updated pill containers or medication lists²⁸. As the number of troughs measured and dose changes made increases, resource utilization from laboratory, phlebotomy, pharmacy, and nursing may be increased in an already resource limited setting. We did not evaluate the cost of these resources, but in a recent study, total first year Medicare reimbursement differed by CYP3A5 phenotype where EM had greater spending (p = 0.02)²⁹. A recent study showed that de novo DSA development was more likely to occur in CYP3A5 expressors (EM or IMs) than nonexpressors (PM) (19% vs 10%, p=0.02) and that expressors experienced more antibody mediated rejection and lower survival rates than non-expressors³⁰. Management of antibody mediated rejection may be associated with greater costs.

In our study, there was no difference in the number of trough measurements by CYP3A5 phenotype. This is not surprising since most transplant centers have set therapeutic drug monitoring protocols in the early posttransplant period and additional measurements in between protocol directed troughs may not be necessary after a dose change given the regularity of protocol troughs. Time in therapeutic range in the first 6 months post-transplant was similar between the CYP3A5 phenotype groups (49.7%–51.0%) and was lower than optimal TTR. Time in therapeutic range has been shown to be important towards clinical outcomes in numerous studies^31–33.

We observed a low frequency of biopsy-proven acute rejection (n = 2) in our study population. A larger number of immunologic events will be needed to determine the role of CYP3A5 phenotype in the risk of rejection, as 46 DSA events were needed to detect an association between de novo DSA development and CYP3A5 expression status ^30,34. A previous meta-analysis comparing CYP3A5*3 and CYP3A5*1 alleles showed an increased risk of acute rejection for patients that carry a CYP3A5*1 allele³⁵.

The differences in tacrolimus dose requirements and dose-normalized troughs among the CYP3A5 metabolism phenotypes are well known²⁰. We also found significantly higher tacrolimus dose requirements in EM (~4-fold) and IM (~1.6-fold) relative to patients with a PM phenotype (p=<0.001). Dose-normalized troughs were 2-fold higher in the PM phenotype group compared to the EM/IM group (p=0.02).

We were not able to exclude the role of race as a confounder in our study, as the CYP3A5*1 genotype is highly correlated with African ancestry. Further, we did not evaluate patient non-adherence or dietary information which may impact tacrolimus absorption. The presence of CYP3A4*22 and drug-drug interactions which may have resulted in a small subset of patients (3–5%) with a misclassified metabolism phenotype. These factors may influence the number of trough measurements and dose adjustments and an understanding all the factors which increase resource utilization may provide an opportunity to provide genotype directed therapeutic drug monitoring, early intervention in at risk individuals thereby improving tacrolimus post-transplant management. Future studies should evaluate additional endpoints such as length of stay, time to stable dose and cost expenditures in a larger population.

Pharmacogenomic testing is becoming more common in clinical practice, and patients are now presenting for transplantation with testing already conducted. Patients without testing could be genotyped while on the waitlist and available at time of transplant for guiding the starting dose or decisions about the regularity of trough monitoring. CYP3A5 genotypes can be measured alongside additional genetic variants as part of a multigene panel which would also provide information relevant for drugs that patients may receive over the lifetime of transplant care. Higher tacrolimus doses are often used in African American patients due to lower HLA matching, and pharmacogenomic testing provides a method to move away from race-based medication decisions and instead use pharmacogenomic testing to select treatments, doses and direct post-transplant monitoring.

Abbreviations Page

BMI

body mass index

CI

confidence interval

CPIC

Clinical Pharmacogenomics Implementation Consortium

CYP3A4

Cytochrome P450 (CYP) 3A4 enzyme

CYP3A5

Cytochrome P450 (CYP) 3A5 enzyme

DeKAF

the Deterioration of Kidney Allograft Function study

dnDSA

de novo Donor Specific Antibody

DSA

Donor Specific Antibody

eGFR

estimated glomerular filtration rate

EM

extensive metabolizer

EMR

electronic medical record

GEN03

Genomics of Kidney Transplantation study

GWAS

genome-wide association study

HR

hazard ratio

IM

intermediate metabolizer

IPV

intra-patient variability

kg

kilogram

LoF

loss of function

MDRD

Modification of Diet in Renal Disease

mg

milligram

mL

milliliter

ng

nanograms

NIAID

National Institute of Allergy

NIH

National Institutes of Health

PM

poor metabolizer

TTR

time in therapeutic range

Data Availability Statement

Aggregated data that support the findings of this study are available from the corresponding author upon reasonable request.