FormalPara Key Points

AKI often occurs in the early phase after heart transplantation, is caused by variable factors and is associated with morbidity and mortality

Supratherapeutic whole-blood tacrolimus trough concentrations often occur early after heart transplantation

A supratherapeutic whole-blood tacrolimus trough concentration is a cofactor of AKI in the first week after heart transplantation

1 Introduction

Over 104,000 heart transplantations have been performed worldwide since 1967 [1]. The immunosuppressive regimen has improved considerably since then [1,2,3]. The introduction of tacrolimus, a very effective immunosuppressive drug, has substantially contributed to increased survival and decreased rejection rates [1,2,3,4]. Despite its success, tacrolimus often has serious side effects, such as nephrotoxicity [3]. Tacrolimus-induced nephrotoxicity frequently evolves into chronic kidney disease (CKD) [5]. The occurrence of CKD in heart-transplanted patients is reported to be 26% after 1 year, 52% after 5 years and in 68% by 10 years. Of these patients, 81% has been treated with the immunosuppressant tacrolimus as the preferred calcineurin inhibitor [6]. It has been acknowledged that CKD after heart transplantation contributes considerably to increasing mortality rates over time [7,8,9]. Logically, prevention of acute kidney injury (AKI) might prevent subsequent CKD in heart transplants [10].

The etiology of AKI in the perioperative phase is often multifactorial. Various factors collectively contribute to the development of AKI, e.g., a high baseline creatinine, a long surgery time, the use of cardiopulmonary bypass, shock, inflammation, the administration of blood products, and nephrotoxic drugs [10,11,12,13,14]. At present, the evidence on tacrolimus trough concentrations being related to AKI after heart transplantation is circumstantial and largely derived from other solid organ transplantations. Therefore, the association between AKI after heart transplantation and tacrolimus is still not fully elucidated.

Our research hypothesis was that supratherapeutic whole-blood tacrolimus trough concentrations are an independent factor in the development of AKI in adult heart transplant recipients. In addition, we analyzed whether AKI after heart transplantation is associated with subsequent development to CKD.

2 Patients and Methods

2.1 Inclusion and Exclusion Criteria

Data of all heart transplantation patients at the University Medical Center Utrecht between April 2005 and December 2012 were retrospectively examined. No multi-organ transplantations were performed. Patients who died within 24 h were excluded as well as patients with preoperative glomerular filtration rate (GFR) < 40 ml/min defined by the Modification of Diet in Renal Disease formula (MDRD) [15]. In these patients, tacrolimus was postponed for several days and basiliximab was used as immunosuppression. Patients who have died on the first day could not be analyzed for kidney injury, because of low exposure to tacrolimus and the delayed increase in plasma creatinine.

2.2 Immunosuppressive Regimen and Dosing

The protocol of the transplantation center demanded that tacrolimus was started at an oral dose of 2 mg twice daily. Further, dosing was based on tacrolimus whole-blood trough concentrations at 6 a.m. (12 h post-dose). A whole-blood tacrolimus trough concentration between 9 and 15 ng/ml was considered therapeutic in the first 2 months and thereafter tapered towards 5–8 ng/ml providing no rejection was encountered [16]. Blood samples for tacrolimus concentration were immediately drawn before the administration and, therefore, represent trough levels. Steady state was not needed for dose adjustments. Corrections on basis of trough concentrations, kidney and liver function, gut motility and interactions with other drugs were left to the discretion of the transplantation cardiologist. Accompanying immunosuppression comprised corticosteroids [prednisolone 50 mg intravenously directly postoperative followed by 25 mg twice daily and tapered off to 20 mg twice daily orally after 6 days] and mycophenolate mofetil [1000 mg orally twice daily]. Basiliximab was not administered in combination with tacrolimus.

2.3 Tacrolimus Assay

The measurements from days 1 to 14, and at 1, 3, 6 and 12 months after transplantation were used for analysis using a micro-particle enzyme immunoassay in accordance with the required quality standards. The lower limit of quantification was 2 ng/ml and intraday imprecision was ± 15% (Abbott IMx™ assay II, Abbott laboratories, Malvern, USA) [17].

2.4 Definition of Kidney Injury

Acute kidney injury (AKI) was classified according to the Kidney Disease Improving Global Outcomes Network (KDIGO) criteria, which distinguishes three classes (See also online resource Table S.1) [18]. Urine data were unavailable; therefore, AKI classification was solely based on serum creatinine concentration: AKI stage 1; Increase in serum creatinine ≥ 26 μmol/L or 150–200% from baseline, AKI stage 2; Increase in serum creatinine > 200% and ≤ 300% from baseline, AKI stage 3; Increase in serum creatinine > 300% or ≥ 354 μmol/L with an acute increase of minimally 44 μmol/L or initiation of renal replacement therapy. Baseline creatinine was the last creatinine prior to surgery. Indications for renal replacement therapy were stage 3 combined with one of the following characteristics: hyperkalemia, severe hypervolemia, uncorrectable metabolic acidosis and serious uremia.

Recovery of AKI was defined as a reduction in peak AKI stage within 3 days, which is consistent for 48 h between day 1 to day 14 [19]. Persistence of AKI was determined at 1 month by comparing the creatinine at 1 month to the baseline creatinine among patients with AKI between day 1 and day 14. Chronic kidney disease (CKD) was determined as the estimated GFR using the “CKD Epidemiology Collaboration equation” [20]. CKD was defined as having a stage 3, 4 or 5 and was assessed after 3 and 6 months and 1 year.

2.5 Collection and Definitions of other Covariates Considered in the Analyses

Drugs interacting with tacrolimus were also documented. Co-medication increasing tacrolimus blood concentrations encompassed macrolides, azoles, calcium antagonists, haloperidol and amiodarone (Additional information is given in the online resource Table S.1, showing the definitions of the covariates). Corticosteroids are known to decrease tacrolimus blood concentrations [21]. Liver dysfunction increases tacrolimus blood concentrations. Liver injury was defined as bilirubin > 34 μmol/L or alanine aminotransferase (ALT) > 90 U/L for men and > 70 U/L for women [22].

Systemic inflammatory response syndrome (SIRS) was determined according to the definition of the American College of Chest Physicians (ACCP) and the Society of Critical Care Medicine Consensus Conference (SCCM) [23]. The only modification to the SIRS criteria was that the heart frequency, providing one SIRS point, was established as 100/min instead of 90/min, because in many heart transplant patients the standard pacemaker configuration is set at 100/min. Shock was specified as mean arterial pressure < 60 mmHg or use of at least one of the following vasopressors/inotropes: norepinephrine, epinephrine, phenylephrine, vasopressin, dopamine, dobutamine or milrinone. We also collected information on potentially nephrotoxic drugs other than tacrolimus that were used in this cohort of heart transplantation patients, particularly (val)acyclovir or (val)ganciclovir (CMV prophylaxis), tobramycin, gentamicin, trimethoprim/sulfamethoxazole (Pneumocystis jiroveci prophylaxis), penicillins, furosemide, vancomycin and amphotericin B. The covariates SIRS, shock and nephrotoxic drugs were not available between days 7 and 14.

2.6 Statistics

Statistical analyses are outlined in the online resource Table S.2. Variables are presented as mean (with standard deviation [SD]), median (with interquartile range [IQR]) or number (proportion) where appropriate. A Generalized Estimating Equations (GEE) procedure was used to test whether AKI was significantly associated with prior supratherapeutic whole-blood tacrolimus trough concentrations, after correction for possible confounders like shock and baseline characteristics (preoperative ventricular assist device (VAD), ischemic cardiomyopathy (ICM), diabetes mellitus (DM), surgery time > 400 min and extracorporeal membrane oxygenation postoperative (ECMO)). GEE analysis was chosen, because it accounts for the correlation among the repeated observations for a given patient [24]. The outcome variable “AKI” was “0” when not meeting the KDIGO criteria and “1” when meeting one of the KDIGO classes (1, 2, or 3).

Linear mixed models were used to study the relationships between different variables (liver injury, other administered drugs and time) and supratherapeutic whole-blood tacrolimus trough concentration. Effects were considered significant when p values were lower than or equal to 0.005 (instead of 0.05 as usual, to compensate for multiple comparisons). Kaplan–Meier analyses were used to compare two groups of patients: group 1 included patients without AKI between day 1 and day 14; group 2 included patients having AKI between day 1 and day 14. Statistical analyses were carried out using SPSS version 15.0 for Windows and SAS version 9.2 for Windows (SAS Institute Inc., USA).

3 Results

3.1 Demographics

Between April 2005 and December 2012, 114 patients underwent heart transplantation in our hospital. In 4 patients, tacrolimus was postponed and basiliximab was part of the initial immunosuppressive strategy. No patient died within 24 h. We analyzed 110 patients of whom five died within the first 30 days and additionally three within 1 year. Causes of death were right ventricular failure (2), pulmonary bleeding (1), primary graft failure (2), acute rejection (1), and infection (2). Two patients were transferred to other hospitals and three were lost to follow-up. Table 1 shows the patient characteristics. Within 1 year, three patients stopped tacrolimus due to severe side effects as neuro- and nephrotoxicity and switched to sirolimus (N = 1) or cyclosporine (N = 2). (The frequencies of AKI and CKD are also shown in the online resource Figure S.1).

Table 1 Patients’ characteristics
Full size table

3.2 Renal Function

Figure 1 shows the distribution of serum creatinine over time. The median baseline creatinine was 102 µmol/L (IQR 84-126 µmol/L). The highest median creatinine of 117 µmol/L was observed at day 3 and slowly decreased to 91 µmol/L at 1 year. The frequency of patients presenting at least one episode of AKI between day 1 and day 6 was 57% (63 out of 110). AKI was most prevalent on day 3 (44%). There was a decrease in frequency of AKI to 17% at day 14, with an increase at 1 month to 25%. The most serious AKI stage 3 [based on “KDIGO criteria”] was most often observed on day 4 (6%). Renal replacement therapy was needed in 6% of patients within the first 14 days (7 out of 110). Recovery from AKI was observed in 15 out of 64 patients (24%). Five patients among the 110 (4.5%) had a recurrent AKI. Persistent kidney injury at one month was not associated with the occurrence of CKD (p = 0.7). The frequency of patients with CKD was 19% at one year. CKD or death was not significantly related to the occurrence of AKI (p = 0.14).

Fig. 1
figure 1

Tukey boxplot of creatinine between day 1 and year 1. Day 0 is baseline creatinine. Box: 25th, median and 75th percentiles, whiskers: maximum value excluding outliers, dots: outliers more or less than 3/2 upper or lower quartile

Full size image

3.3 Variables Influencing AKI

GEE analyses showed that AKI was associated with prior supratherapeutic whole-blood tacrolimus trough concentrations both in the first week (OR 1.66; 95% CI 1.20–2.31) and in the first 2 weeks (OR 2.10; 95% CI 1.35–3.25) after transplantation, even after correction for the variables ICM, VAD, DM, age, contrast administration, surgery time > 400 min and ECMO post-operation. Among the patients who presented a supratherapeutic level, the median duration was 2 days (minimum 1 day, maximum 5 days). Nephrotoxic drugs as described in the method section were frequently co-administered during the first week after transplantation (95%, 104 out of 110) and most often on day one and two (72 and 75%, respectively). At day 6, 33% of the patients used nephrotoxic drugs. Shock was observed in 96% of patients (105 out of 110) during the first 6 days. The occurrence of shock was most often seen on day one (84%) and decreased to 50% on day 6. The frequency of SIRS was highest on day 2 (86%) with a decrease to 31% on day 6 (Table 2).

Table 2 GEE analyses to test the variables influencing AKI
Full size table

3.4 Tacrolimus Blood Concentrations

We analyzed a total of 473 tacrolimus whole-blood trough concentrations within the first week after transplantation. In 34% of patients (37 out of 110), a supratherapeutic concentration was measured at least once. Whole-blood tacrolimus trough concentrations ranged from 1.5 to 35 ng/mL with the highest median at day 4, 5, 6 and 7 of 10 ng/mL. At month 3, the median tacrolimus concentration was 9 ng/mL, indicating a supratherapeutic concentration. Median whole-blood tacrolimus trough concentrations showed to be below therapeutic range during the first 3 days, at day 10 and 12, at month 6 and at 1 year (see Fig. 2). At day 4, 5 and 6 the highest frequencies of supratherapeutic whole-blood tacrolimus trough concentrations were measured (16% on all days). At day, 14 3% had a supratherapeutic level and at 1 month this frequency increased to 8%. (See Table S.3 in the online resource, demonstrating the variables influencing whole-blood tacrolimus concentrations). Liver injury was often observed within the first week after transplantation (56%, 61 out of 110). The highest occurrence of liver injury was observed on day 2 post-transplantation (28%) with a decrease to 16% on day 6. Supratherapeutic whole-blood tacrolimus trough concentrations were not significantly associated with liver injury (p = 0.03) (See Table S.3 in the online resource, demonstrating the variables influencing whole-blood tacrolimus concentrations). Almost all patients used drugs increasing tacrolimus concentrations within the first week after transplantation (95%, 104 out of 110). Drugs increasing tacrolimus concentrations were most frequently used on day 1 (82%). From day 2 on drugs decreasing tacrolimus concentrations were applied. The prevalence of the administration of these drugs was highest on day 2 (71%). One patient was treated with high dose methylprednisolone for 6 days because of suspected rejection. Drugs, which could affect tacrolimus concentrations, increasing as well as decreasing, were not significantly associated with whole-blood tacrolimus trough concentrations (p values 0.86 and 0.07, respectively, See also in the online resource Table S.3).

Fig. 2
figure 2

Tukey boxplot of whole-blood tacrolimus trough concentrations between day 1 and year 1. The therapeutic range of 9–15 ng/ml and 5–8 ng/ml is indicated as gray bars in the figure. Box: 25th, median and 75th percentiles, whiskers: maximum value excluding outliers, dots: outliers more or less than 3/2 upper or lower quartile

Full size image

4 Discussion

We found a high incidence rate of AKI after heart transplantation. AKI was independently associated with supratherapeutic tacrolimus concentrations, which were most often observed in the first week.

The high incidence rate of AKI (57%) in our cohort of heart transplant patients was in accordance with Tjahjono et al., in which AKI was seen in 59% of the patients within the first week [12]. These high incidence rates showed to be higher than in a cohort of lung transplants and may be based on the high frequency of shock and on the high median baseline creatinine, reflecting a decreased renal function pre-operatively [12, 25]. The high baseline creatinine is in accordance with previous observations of heart transplants [6]. A diminished renal function may further deteriorate in case of shock and inflammation, the use of a pulmonary bypass circuit, increased administration of blood products as well as with the administration of nephrotoxic drugs, such as tacrolimus [12]. It was suggested that a supratherapeutic tacrolimus concentration at day 3 is a risk factor for the development of AKI. This was observed in a pediatric heart transplantation cohort, though it did not reach significance [11]. We showed that a prior supratherapeutic whole-blood tacrolimus trough concentration was an independent predictor of AKI in an adult heart transplantation cohort.

The occurrence of AKI is a predictor of CKD. In a group of 300 heart transplantation patients, AKI was related to CKD [10]. The risk of progression to CKD is related to the stages of AKI. Even mild AKI, stage 1, increases the risk of CKD [26]. We could not relate AKI to CKD. This lack of association may have been caused by a too small sample size or too short follow-up period. Nonetheless, we observed a low recovery rate of AKI in our cohort [27, 28]. Recovery of AKI in heart transplants is important. AKI persistent at 1 month after transplantation has been shown to significantly decrease survival rates compared to patients with complete renal recovery [1, 10, 13, 28]. Moreover, earlier reports showed a steady increase in the percentage and severity of CKD after heart transplantation in the first 10 years with a median time to progression to CKD stage 4 of 3 years [1, 9]. Hence, the importance of AKI in the development of CKD shows to be pivotal and tacrolimus may have an important part in it.

Tacrolimus is an ongoing assault to the kidneys, because it is administered continuously. Nephrotoxicity induced by tacrolimus is assumed to be caused by acute vascular and tubular damage, and chronic irreversible tubule-interstitial fibrosis [29,30,31,32,33,34,35,36,37]. Renal biopsies show a gradual increase of arteriolar hyalinosis, glomerulosclerosis as well as interstitial fibrosis in patients treated with calcineurin inhibitors [38]. Paradoxically, anatomical abnormalities may go unnoticed and tacrolimus nephrotoxicity might be present without clinical loss of renal function. Nankivell et al. observed a mean GFR of 60 ml/min in the presence of grade I nephropathy and of 50 ml/min in the presence of chronic nephropathy of grade II or higher [38].

To avoid early tacrolimus nephrotoxicity, decreased doses and delayed introduction have been suggested. To preclude early rejection when tacrolimus is postponed, an interleukin-2 receptor inhibitor or mTOR inhibitor could be administered [39,40,41,42,43,44]. Unfortunately, mTOR inhibitors are poorly tolerated, with almost one-third discontinuing the drug within 1 year after transplantation [41, 44,45,46]. Rejection rates may be higher in tacrolimus free regimens; therefore, tacrolimus is still the preferred immunosuppressive drug after heart transplantation [41].

Tacrolimus was carefully dosed in our cohort, reflected by the low starting dose and the (sub-)therapeutic median whole-blood concentrations. Dosages were substantially lower than in a cohort of lung transplantation patients [25]. Yet, tapering of corticosteroids, increasing tacrolimus blood concentrations, was started in a later stage as in the lung transplantation cohort [25]. Nevertheless, we found a high frequency of supratherapeutic concentrations during the first 2 weeks. This cohort of heart transplants showed a higher frequency of liver injury and shock compared to the cohort of lung transplantation patients, reflecting the higher frequency of clinical instability [25]. Furthermore, we observed a high prevalence of drugs influencing the blood tacrolimus concentrations. Although none of the beforementioned variables individually influenced the development of subsequent supratherapeutic tacrolimus concentrations, all combined might still contribute to supratherapeutic tacrolimus concentrations. Tacrolimus dosing is complex and extremely difficult in the clinically unstable phase after heart transplantation.

To prevent from supratherapeutic whole-blood trough concentrations, an even more tentative dosing scheme could be used. Targeting at the narrow therapeutic range may have contributed to the supratherapeutic levels. Starting with a low dose and using only an upper level may prevent peaks in the concentrations. This qualitative dosing scheme may prevent erratic tacrolimus concentrations and thus supratherapeutic tacrolimus concentrations.

4.1 Limitations of this Study

There are some notable limitations to this study due to its retrospective character. Several variables influencing pharmacokinetics of tacrolimus, for instance, the effect of variations in HDL, alpha-1-acid glycoprotein, acidosis, changes in fluid balance, gut motility and variations in concentrations or activity of CYP 3A4/5 and P-glycoprotein, could not be studied.

In our study, only 18 patients were still followed at 1-year post-transplantation and developed a CKD. A larger sample size would increase the reliability of our results, e.g., the potential influence of tacrolimus on the occurrence of CKD is likely to remain undetected in our study (Kaplan–Meier analysis). Similarly, it is plausible that our study had inadequate power to detect the effect of other drugs and liver injury on the whole-blood tacrolimus trough concentration (mixed model analysis).

Trough concentrations were assumed, though with one sample per 24-h period the assumption may be insufficient. Especially, clinically unstable patients may lack steady state and large changes in half-life times may occur. This may hamper the interpretation of the blood concentrations.

We could not analyze all the known factors related to AKI after heart transplantation, e.g., the amount of blood products administered, inotrope and ventilation duration. Moreover, ultrasound, biopsy, and urine analyses have not been performed. Therefore, not all causes of kidney injury have been investigated in this study and they may have had an effect on kidney function.

Another limitation is the estimation of AKI itself. At present, the KDIGO criteria are considered to be the best option. However, plasma creatinine provides only an indication of renal function. Plasma creatinine in heart transplantation patients shortly after surgery may overestimate renal function [47]. Moreover, CKD-EPI tends to overestimate GFR [9]. Uniformity in defining kidney injury as well as recovery from AKI may be an important improvement for the comparison of renal outcome in heart transplant recipients.

5 Conclusions

This study shows that AKI after heart transplantation is  correlated to supratherapeutic whole-blood tacrolimus trough concentrations. Low recovery rates of AKI might even be a reflection of ongoing tacrolimus toxicity. Therefore, the prevention of tacrolimus nephrotoxicity early after transplantation may be crucial in preserving kidney function. The appropriate tailoring of tacrolimus dosing early after transplantation could be a key factor in improving transplantation outcomes.