**3. Results**

#### *3.1. Study Population*

A total of 121 patients were included in this study: 61 in the intervention group and 60 in the control group. An overview of patient characteristics, underlying diagnoses for LTx, comorbidities and immunosuppression is shown in Tables 1 and 2. There were only small di fferences in the demographic data between the study groups. The control group had a more extended warm ischemic time (*p* = 0.005). As coimmunosuppression, patients received mycophenolate mofetil (MMF) at a daily dose of 1000 (500–2000) mg (LCPT) and 1500 (500–2000) mg (control), everolimus at 2.0 (0.5–5.0) mg (LCPT) and 2.0 (2.0–4.0) mg (control), prednisolone at 5.0 (5.0–7.5) mg (LCPT and control) and sirolimus at 1.0 mg (control). In the intervention group, 45 patients su ffered from chronic kidney disease (CKD, categories 2–4). The control group showed a similar distribution in 39 LT recipients. No patients were in CKD category 5 or on dialysis. In the absence of kidney biopsies, the underlying renal disease remained unclear. The median interval between transplantation and study onset was 2.8 (0.1–20.8) years in the intervention group and 6.6 (0.2–16.5) years in the control group (*p* < 0.001). The reasons for a conversion from standard-release Tac to LCPT were CNIT (n = 7), neurotoxicity (n = 5) and prevention of side e ffects via better bioavailability of LCPT (n = 49).


**Table 1.** Patient characteristics.

Statistics: Values shown as mean ± standard deviation or number (percentage). a *t*-test, b Fisher's exact test. LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; LTx, liver transplantation; BMI, body mass index; CIT, cold ischemic time; WIT, warm ischemic time; CMV, cytomegalovirus.


**Table 2.** Underlying diagnoses for LTx, comorbidities and immunosuppression at study start.

Statistics: Values shown as number (percentage). All *p*-values from Fisher's exact tests. LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; LTx, liver transplantation; NASH, nonalcoholic steatohepatitis; CKD, chronic kidney disease (categories set with reference to [23]). MMF, mycophenolate mofetil; CNIT, calcineurin inhibitor nephrotoxicity.

#### *3.2. C*/*D Ratio*

At study start (baseline), the C/D ratio in the intervention group was comparable to that in the control group (1.68 (0.30–13.45) vs. 1.76 (0.38–7.40) ng/mL×1/mg, respectively; *p* = 0.362, Table 3). During the 12-month evaluation period, no significant changes in the C/D ratio were observed in the control group. After 12 months, the median C/D ratio was approximately at the baseline level (1.75 (0.49–6.40) ng/mL × <sup>1</sup>/mg; *p* = 0.847). In the control group, there was a slight decrease in both the daily Tac dose at study end compared with that at baseline (2.5 (0.5–10.0) vs. 2.8 (0.5–10.0) mg, respectively; *p* = 0.084), as well as in the median Tac trough level (4.7 (1.5–14.3) ng/mL at study onset to 4.1 (1.6–15.6) ng/mL after 12 months; *p* = 0.082). However, the di fferences in both cases were not significant.

In contrast, the C/D ratio in patients switched to LCPT was 50% higher 12 months after conversion than that at baseline (2.52 (0.58–6.40) vs. 1.68 (0.30–13.45) ng/mL × <sup>1</sup>/mg, respectively; *p* < 0.001). A significant increase in the C/D ratio was already observed in this group 3 months after study onset (2.03 (0.33–13.60) ng/mL × <sup>1</sup>/mg; *p* = 0.008). Regarding the daily Tac dose, a significant reduction of 33.3% was observed after 12 months compared with that at baseline (2.0 (0.4–7.8) vs. 3.0 (1.0–22.0) mg, respectively; *p* < 0.001)). Moreover, the Tac trough level was significantly reduced at study end (4.4 (2.2–11.8) vs. 6.0 (1.5–26.9) ng/mL at study onset; *p* < 0.001).

To confirm that conditions were stable before study onset, C/D ratios, Tac doses and trough level 3 months before enrolment were also obtained. There were no significant di fferences between the groups at t−<sup>3</sup> (Table 3) nor between study start and 3 months earlier within a group. Patients in the intervention group showed similar median C/D ratio compared with that at baseline (1.44 (0.24–6.20) vs. 1.68 (0.30–13.45) ng/mL × <sup>1</sup>/mg, respectively; *p* = 0.204). Daily Tac dose differed significantly due to single outlier values shortly after transplant (3.0 (0.5–12.0) (t−3) vs. 3.0 (1.0–22.0) (t0) mg; *p* = 0.049), while Tac trough level showed no considerable differences (5.0 (2.4–15.3) (t−3) vs. 6.0 (1.5–26.9) (t0) ng/mL; *p* = 0.722).

No significant differences were detectable in the control group between baseline and 3 months before: C/D ratio (1.69 (0.40–9.20) vs. 1.76 (0.38–7.40) ng/mL × <sup>1</sup>/mg, respectively; *p* = 0.626), Tac daily dose (2.5 (0.5–9.0) vs. 2.8 (0.5–10.0) mg, respectively; *p* = 0.362) and Tac trough level (4.4 (1.5–14.7) vs. 4.7 (1.5–14.3) ng/mL, respectively; *p* = 0.742).


**Table 3.** Tacrolimus concentration/dose (C/D) ratio, daily dose and blood trough concentration.

To confirm that conditions were stable before enrolment, values 3 months prior to study onset are given for all patients who had already undergone liver transplantation (n = 54 vs. 58). In the intervention group (LCPT), values 3 months before and the day before the first LCPT intake (study onset) were determined when s-r-Tac was administered. LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; s-r-Tac, standard-release tacrolimus. *p*-values from Mann–Whitney U-test.

As shown in Figure 2, the C/D ratio at study end was significantly higher in patients on LCPT than in the control group (2.52 (0.58–6.40) vs. 1.75 (0.49–6.40) ng/mL × <sup>1</sup>/mg, respectively; *p* = 0.009). The median Tac trough level and the daily dose were significantly higher in the intervention group at study onset (Table 3). After 12-month follow-up, the Tac dose in the LCPT group was significantly reduced compared with that in the control group (2.0 (0.4–7.8) vs. 2.5 (0.5–10.0) mg, respectively; *p* = 0.047). However, the Tac trough level was comparable in the two groups at study end (4.4 (2.2–11.8) vs. 4.1 (1.6–15.6) ng/mL, respectively; *p* = 0.283).

**Figure 2.** Boxplots of C/D ratio among patients receiving LCPT (dark grey) or standard-release Tac (light brown) at baseline and 3, 6, 9 and 12 months later. There were significant differences between the two study groups at 6 and 12 months after conversion. *p*-values reflect differences between the groups at each time point.

#### *3.3. Renal Function*

At baseline (study onset, t0), patients in the control group had a higher mean eGFR than patients switched to LCPT (Figure 3), although the difference (ΔeGFR) was not significant (*p* = 0.157). However, mean ΔeGFR in patients on LCPT had significantly improved at 6 months after conversion (*p* = 0.029). In contrast, patients on standard-release Tac showed a significant decline of mean ΔeGFR 9 months after study initiation (*p* = 0.006). Over the 12-month evaluation period, mean ΔeGFR continued to improve significantly in patients receiving LCPT (*p* = 0.001), whereas mean ΔeGFR continued to deteriorate in the control group (*p* < 0.001). In a pairwise comparison between the groups, eGFR values did not differ significantly (Supplementary Table S1).

**Figure 3.** Glomerular filtration rate (eGFR; mL/min/1.73 m2) over time and the difference from baseline at each time point (ΔeGFR ± SEM) in each study group. Improved renal function with a significantly increased mean ΔeGFR was already observed 3 months after conversion to LCPT (dark grey). *p*-values reflect comparison of ΔeGFR between the study groups.

While absolute eGFR values are meaningful to only a limited extent, eGFR slope (ΔeGFR) relative to the baseline can be used as additional empirical support (Table 4). Three months before study onset, there were no significant differences within the study groups relative to baseline. In the intervention group, renal function increased 6 months after conversion (*p* = 0.029). In contrast, LT recipients in the control group showed a significant decline of eGFR 9 months after study initiation (*p* = 0.006). Over the 12-month evaluation period, renal function continued to significantly improve in patients receiving LCPT (*p* = 0.001), whereas eGFR continued to deteriorate in the control group (*p* < 0.001).


**Table 4.** Slope analysis (ΔeGFR) of glomerular filtration rate (eGFR; mL/min/1.73 m2).

The "estimate" value describes the difference between the respective time point and the baseline (ΔeGFR). A negative value shows a decline and a positive value an improvement of eGFR. LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; *p*-values within a group are relative to the baseline.

In further analysis, the eGFR values of the patients suffering from diabetes mellitus and arterial hypertension were compared between the groups.

At every time point, patients with diabetes mellitus had significantly lower eGFR than patients without it, regardless of the study group (Table 5). However, eGFR among diabetic patients recovered in a manner similar to that of nondiabetics upon switching to LCPT. In contrast, renal function deteriorated in patients maintained on standard-release Tac in a similar fashion, regardless of diabetes.


**Table 5.** Glomerular filtration rate (eGFR; mL/min/1.73 m2) in diabetic and nondiabetic patients.

LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; t0 to t12, time points (months). eGFR values shown as mean ± standard deviation. *p*-values from *t*-test.

Patients with arterial hypertension in both study groups had a lower mean eGFR than patients with normal blood pressure at each time point (Table 6). However, renal function recovered in patients treated with LCPT and deteriorated in those maintained on standard-release Tac over the course of the study, regardless of the presence of arterial hypertension.

**Table 6.** Glomerular filtration rate (eGFR; mL/min/1.73 m2) in patients with and without arterial hypertension.


LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; t0 to t12, time points (months). eGFR values shown as mean ± standard deviation. *p*-values from *t*-test.

Multivariable analysis was performed to identify independent predictors of alterations in renal function expressed as ΔeGFR (Supplementary Table S2). Conversion to LCPT was the only identified independent predictor of significant changes in eGFR.

#### *3.4. Liver Function*

During the entire follow-up, we monitored the graft function (Table 7). LT recipients in the LCPT group showed significantly lower serum bilirubin concentrations than the control group at all time points. However, the median values in both study groups remained within the lower part of the normal range throughout the course of the study. Regarding the parameters ALT and INR, no differences were observed between the groups.


**Table 7.** Assessment of liver function over time in each study group.

LCPT, once-daily MeltDose® tacrolimus; Tac, tacrolimus; ALT, alanine transaminase; INR, international normalized ratio; *p*-values from Mann–Whitney U-test.

## **4. Discussion**

The present study shows that the conversion of LT recipients from standard-release Tac to LCPT was beneficial in regard to renal function. This may be due to the improved bioavailability of LCPT which led to a significant increase in C/D ratio.

Notably, the median daily Tac dose declined by 33.3% among LT recipients after conversion. A dose reduction of approximately 30% with a comparable area under the curve (AUC) was reported in recent studies of KT and LT recipients [20,24,25]. In those studies, this finding was also attributed to the greater bioavailability of LCPT.

In our cohort, the median C/D ratio among LT recipients who switched to LCPT had increased by 50% at 12 months after conversion. The C/D ratio among patients maintained on standard-release Tac remained unchanged over the 12-month period. In accordance with these data, Franco et al. described a 35% increase in the C/D ratio among KT recipients after conversion from IR-Tac and a 83.3% increase among those who were switched from ER-Tac to LCPT [26]. In the study by Rostaing et al., KT recipients had a 20% higher C/D ratio 12 months after conversion to LCPT and a 24.4% higher C/D ratio 24 months after conversion [27]. In contrast, Kami ´nska et al. showed that the C/D ratio of KT recipients converted from IR-Tac to ER-Tac did not change significantly [28]. To our knowledge, the present study is the first to describe a significant increase in the C/D ratio after a switch to LCPT among LT recipients.

In a previous study, we explored the impact of the C/D ratio on renal function after kidney transplantation (KTx) [14]. Fast metabolizers, defined as patients with a C/D ratio < 1.05 ng/mL × <sup>1</sup>/mg, showed a strong association with decreased renal function compared with slow metabolizers in a 24-month follow-up. Similar results were confirmed among LT recipients in a 36-month follow-up study [13]. In that cohort, the cut-o ff value for fast metabolizers was defined as a C/D ratio < 1.09 ng/mL × <sup>1</sup>/mg. In a 5-year follow-up, KT recipients with a lower Tac C/D ratio showed a higher risk of renal impairment as well as higher mortality rates [17]. Recently, several studies confirmed these findings [19,29,30] and a further negative impact of fast Tac metabolism on increased kidney allograft rejection rates and BK virus infections was demonstrated [17,18,31].

Given these results, we postulated that a higher C/D ratio after conversion to LCPT is associated with nephroprotection. Surprisingly, we already observed significant improvement of renal function 6 months after conversion. Twelve months after conversion, the mean ΔeGFR was 4.7 mL/min/1.73 m<sup>2</sup> higher than at baseline. In contrast, eGFR had deteriorated significantly in patients maintained on standard-release Tac 9 months after study onset and ΔeGFR had decreased by 4.3 mL/min/1.73 m<sup>2</sup> at 12 months.

After conversion to LCPT, the median trough level declined from 6.0 ng/mL at study onset to 4.6 ng/mL (month 3) without a subsequent decrease until month 12. A lower Tac trough level in the LCPT group has already been reported in a prospective study, although the same target trough level was given [27]. Alongside better bioavailability of LCPT, trough level reduction might be another reason for the increase in renal function. However, median trough levels did not vary considerably between subsequent time points (t3–t12) while renal function showed further recovery. Notably, median Tac trough levels were also slightly reduced in the control group (t0–t12), although eGFR showed further decline over the 12-month follow-up. Therefore, we postulate that improvement of bioavailability and a reduced peak Tac level after conversion to LCPT are factors more relevant to the increase in eGFR than the reduction in Tac trough levels alone.

As an explanation for the nephroprotective potential of LCPT, Schütte-Nütgen et al. hypothesized that a lower daily Tac dose results in a lower peak serum concentration (Cmax), which in turn reduces the side e ffects of Tac overdosing within the first hours after drug intake [17]. In a review article on LT recipients, Baraldo reported that LCPT had a similar AUC after 24 h and a similar minimal blood concentration (Cmin), but had a significantly lower Cmax and a smaller Cmax/Cmin fluctuation ratio when compared with IR-Tac [32]. In addition, Bunnadaprist et al. postulated that there is a reduced cumulative Tac dose in KT recipients receiving LCPT [33]. In a recent study, we also showed that fast metabolizers with a C/D ratio < 1.05 ng/mL × <sup>1</sup>/mg had significantly higher Tac blood concentrations than slow metabolizers 2 h after Tac intake [16]. In the same study, we showed that a low C/D ratio was significantly associated with acute CNIT. Although renal biopsy is not routinely performed in LT recipients, we can assume that patients converted to LCPT su ffered less frequently from CNIT. In contrast, Kamar et al. reported similar renal function in de novo KTx recipients who were randomized to LCPT or ER-Tac in a 4-week follow-up [34]. Notably, Cmin and AUC0–24 were slightly higher in the LCPT group (at days 3, 7 and 14), a fact that might have influenced the results.

In the current study, the control group had an increased warm ischemic time (WIT) compared with the intervention group (~5 min). Prolonged cold and warm ischemic times can be associated with long-term allograft dysfunction [32]. Nevertheless, at the beginning of our study, the liver function parameters ALT and INR did not di ffer between the groups and median bilirubin was within the normal range. In a study by Laskey et al., increasing WIT during LTx was associated with a lack of renal recovery in the presence of pretransplant subacute kidney injury [35]. It was concluded that minimization of WIT could potentially avoid renal replacement therapy or the need for subsequent kidney transplantation. At the study start in our cohort, the control group showed even higher eGFR values despite increased WIT compared with the intervention group. Notably, the control group had a more extended interval between LTx and study onset than patients switched to LCPT (6.6 (0.2–16.5) vs.2.8 (0.1–20.8) years, respectively).

In regard to the Tac formulations used before study onset, IR-Tac was administered more frequently than ER-Tac in the intervention group and vice versa in the control group. A recent study on pharmacokinetics in a large transplant cohort showed similar Tac trough levels and bioavailability between these two formulations [36]. Notably, C/D ratio as well as C/D intrapatient variability was reported not to change considerably during conversion from IR-Tac to ER-Tac in KT recipients [28]. These findings justify our and others' approach of including patients taking either one of these formulations [19,29].

In the current study, patients su ffering from diabetes mellitus or arterial hypertension had reduced renal function. Interestingly, patients who were switched to LCPT (median C/D ratio increased from 1.68 to 2.52 ng/mL × 1/mg) showed considerable recovery of eGFR independent of the presence of both conditions. In accordance with these findings, Bardou et al. showed that slow Tac metabolizers (C/D ratio > 1.8 ng/mL × 1/mg) were less likely to su ffer from diabetes and hypertension after LTx [37].

Finally, we recognize that our study has limitations due to its retrospective design and the limited sample size from a single-centre. In addition, in this study, we cannot provide Tac Cmax, C2 (2 h after Tac intake) nor AUC, although higher Cmax or C2 could potentially induce higher CNIT. Therefore, we can only hypothesize that, after conversion to LCPT, lower C2 was a more relevant factor to the improvement of renal function than trough level reduction. Further investigations should also include data on the concentrations of di fferent Tac metabolites, which could be responsible for adverse e ffects, such as CNIT, infections and myelotoxicity [38,39]. Furthermore, given the retrospective design of this study, the study beginning in the control group had a wide range from March 2017 until August 2018 and the time period from LTx to the beginning of the study was significantly increased compared with that in the intervention group. The longer Tac exposure in the control group might have had a negative influence on renal function in this cohort. However, at t0, the control group showed even higher eGFR values than patients converted to LCPT (70.6 ± 19.3 vs. 65.3 ± 21.1, respectively).

Another limitation of the study is that the reasons for conversion to LCPT in our study were taken only from the clinical reports from our Outpatient Transplant Clinic. In addition, in contrast to the case for KTx recipients, renal biopsy is not routinely performed in LT recipients which limits our ability to analyse CNIT before study onset.
