**E**ffi**cacy and Safety of Belatacept Treatment in Renal Allograft Recipients at High Cardiovascular Risk—A Single Center Experience**

**Hannes Neuwirt 1,\*, Irmgard Leitner-Lechner 2, Julia Kerschbaum 1, Michael Ertl 1, Florian Pöggsteiner 1, Nicolas Pölt 1, Julius Mätzler 1, Hannelore Sprenger-Mähr 3, Michael Rudnicki 1, Peter Schratzberger 1, Iris E. Eder <sup>4</sup> and Gert Mayer <sup>1</sup>**


Received: 2 July 2019; Accepted: 1 August 2019; Published: 3 August 2019

**Abstract:** Belatacept is an attractive option for immunosuppression after renal transplantation. Renal allograft function is superior when compared to calcineurin inhibitor (CNI) based therapy in "de novo" treated patients and it has also been proposed that individuals at high cardiovascular (CV) risk may benefit most. In this retrospective cohort study, we assessed the efficacy and safety of treating patients at high cardiovascular risk with Belatacept (*n* = 34, for 1194 observation months) when compared to a matched control group of 150 individuals under CNI immunosuppression (for 7309 months of observation). The estimated glomerular filtration rate (eGFR) increased for patients taking Belatacept but decreased during CNI-based therapy (+2.60 vs. <sup>−</sup>0.89 mL/min/1.73 m2/year, *<sup>p</sup>* <sup>=</sup> 0.006). In a multivariate Cox regression model, Belatacept remained the only significant factor associated with the improvement of eGFR (HR 4.35, 95%CI 2.39–7.93). Belatacept treatment was not a significant risk factor for renal allograft rejection or graft loss. In terms of safety, the only significant risk factor for de novo cardiovascular events was a pre-existing cerebrovascular disease, but Belatacept was not associated with a significant risk reduction. Belatacept treatment was not associated with an increased risk of severe infections, cytomegalo virus (CMV) or BK-virus reactivation, malignancy or death in the multivariate Cox regression analysis. Belatacept is an efficient and safe option for patients after renal transplantation at high cardiovascular risk.

**Keywords:** kidney transplantation; Belatacept; cardiovascular high risk; outcome

#### **1. Introduction**

Calcineurin inhibitors (CNIs) are currently the standard immunosuppressive therapy after renal transplantation. Their introduction into clinical practice has improved short-term outcomes dramatically. Unfortunately, the rate of late allograft loss has not significantly improved [1] and it is generally accepted that CNI nephrotoxicity contributes to this problem. Thus, multiple studies have investigated the impact of CNI-free immunosuppression on renal allograft function and patient and graft survival. The use of mammalian target of rapamycin inhibitors (mTORi) is impeded by drop-out rates of up to 40% due to side effects [2] and furthermore is associated with higher allograft rejection rates [3].

Belatacept inhibits T-cell activation by blocking a co-stimulatory signal by binding to CD80/CD86 on antigen presenting cells [4]. It is currently approved for the de novo immunosuppression of renal allograft recipients in combination with mycophenolic acid and steroids. Studies have demonstrated an improved allograft survival over several years when compared to cyclosporine A-based immunosuppression [5,6]. Furthermore, a protocol for switching patients from CNIs to Belatacept has been published [7,8]. This conversion improved kidney function relative to the baseline and was safe concerning risk of death or transplant loss. Finally, it has been proposed that Belatacept-based regimens might have beneficial effects, especially in patients at high cardiovascular (CV) risk (reviewed by [9]). One mechanism might be a reduction of pulse wave velocity in patients treated with Belatacept compared to CNI-treated patients [10,11]. However, "real world" data on renal outcomes and especially safety in the latter individuals are sparse.

Thus, we conducted a retrospective cohort study in renal allograft recipients at high CV risk and compared the efficacy and safety of Belatacept treatment in 34 patients to the outcomes of 150 patients treated with CNI (mainly tacrolimus) based immunosuppression.

#### **2. Materials and Methods**

#### *2.1. Patient Population*

Belatacept has been used in 42 renal transplant recipients at our center since 2010 and in this retrospective cohort study, all patients at high cardiovascular risk (definition see below) have been included (*n* = 34). Eighteen patients were treated de novo and 16 were converted at a median of 1.6 months (interquartile range (IQR), 0.6–4.2 months) after transplantation, mainly due to biopsy confirmed or clinically suspected renal CNI toxicity. No patient in this group was returned to CNI therapy thereafter. As we were interested in studying the efficacy and safety in patients on either CNI or Belatacept therapy, the day of conversion was taken as the baseline in these individuals and all clinical endpoints were adjudicated to the Belatacept group. Due to the early conversion, we excluded the time on CNI from any calculation. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Medical University Innsbruck (Project-ID: 1137/2019).

The median observation duration on Belatacept was 35 months, no patient was lost for follow-up and the total period on therapy analysed was 1194 months. All renal allograft recipients (*n* = 309) on CNI-based immunosuppression that received their transplant between 1 January, 2010 and 31 December, 2012 formed the control cohort. Of these, 150 also fulfilled the criteria for high cardiovascular risk. No patients were lost for follow-up, the median follow-up was 48 months and the total analysed period of months on therapy was 7309. High cardiovascular risk was defined by the presence of any significant pre-transplant coronary artery disease (CAD) confirmed on angiography, a history of myocardial infarction, peripheral artery disease (PAD) (cardiovascular disease) or stroke (cerebrovascular disease) or the presence of diabetes mellitus in combination with arterial hypertension.

#### *2.2. Endpoints*

Efficacy endpoints were renal allograft function as assessed by a change of eGFR on therapy, number of rejection episodes (either confirmed by biopsy or clinically based on an improvement of allograft function after anti-rejection therapy) or graft loss. The estimated glomerular filtration rate (eGFR) was calculated using the abbreviated MDRD formula. ΔeGFR was calculated by dividing the difference between eGFR at last follow-up and the baseline by the number of follow-up years. The safety endpoints were de novo cardiovascular events (new myocardial infarction, newly diagnosed CAD of any stage, newly diagnosed peripheral artery disease), severe infections (defined as infection leading to the admission of the patient to hospital), cytomegalo virus (CMV) reactivation (diagnosed by PCR with or without a clinical CMV infection), BKV reactivation (as determined by PCR in serum

and/or urine), de novo malignancy and death. All efficacy and safety endpoints were identified using patients' records.

#### *2.3. Statistics*

The reported values represent either medians and interquartile ranges (IQR) or the number of patients and percentages of the respective cohort. Proportions were compared using the Chi2 or Fisher exact tests. Non-parametric tests were used to compare continuous variables. The factors potentially associated with the eGFR, the eGFR-slope (eGFR), and efficacy and safety parameters were assessed using a Cox regression analysis. In particular, those factors were: Belatacept treatment, recipient age, male gender of the recipient, recipient BMI, a CMV high risk mismatch (D+/R-), the presence of diabetes mellitus or arterial hypertension, the presence of cerebrovascular or cardiovascular disease, the time on renal replacement therapy (RRT) before renal transplantation (RTx), number of previous RTx, number of HLA mismatches, intraoperative urine production (initial diuresis), number of post-operative (PO)—meaning after renal transplantation—hemodialysis sessions (HDs), the absence or presence of steroids at discharge, the presence of serum-creatinine at discharge, the absence or presence of steroids at the last follow-up, the extended criteria donor (ECD) organ, the male sex of the donor, and donor age. A history of rejection was also included, with an exception for the endpoint rejection episodes. Variables with a *p*-value < 0.05 in univariate analysis were included in the multivariate analysis, where again a *p*-value < 0.05 was considered statistically significant. The analysis was performed using SPSS Version 24.

#### **3. Results**

#### *3.1. Baseline*

Baseline data are shown in Table 1. Belatacept patients spent a shorter period of time on renal replacement therapy (RRT) before renal transplantation, were more likely to suffer from CV disease or hypertension and less likely to have diabetes mellitus compared to CNI patients. The baseline data were not significantly different between de novo and converted Belatacept patients, except for the higher proportion of male recipients in de novo patients (17/18 vs. 8/16, *p* = 0.003).

#### *3.2. Renal Transplantation*

Concerning renal transplantation (RTx, Table 2), the donor type was significantly different between CNI- and Belatacept-treated patients. This was primarily driven by a higher proportion of living donors and deceased donors that died due to circulatory reasons in the Belatacept group. The Belatacept patients received more organs from female donors and donors were older (61 vs 49.5 years) and had a higher BMI compared to CNI patients. The proportion of patients with intraoperative urine production (initial diuresis) was lower (76% vs. 95%) and the number of hemodialysis sessions (HDs) was significantly higher in Belatacept patients. Hence, renal allograft function at discharge, as assessed by serum creatinine (1.52 vs. 2.20 mg/ml, *p* = 0.001) and eGFR (MDRD) (44.5 vs. 28 mL/min/1.73 m2, *p* = 0.001), was significantly worse in Belatacept patients.

#### *3.3. E*ffi*cacy*

The median follow-up (Table 3) was 1462 and 1054 days in CNI and Belatacept patients (*p* = 0.084), respectively. Belatacept was continued in all patients with a maintained graft function during follow-up (31/34). The number of patients on steroids at follow-up and the proportion of hypertensive patients were higher in the Belatacept group.



The median and interquartile range (IQR) are depicted except for nominal variables, where the number of patients (*n*) and percentages are shown. *p*-values < 0.05 are marked with an asterisk \*. BMI: body mass index, RTx: renal transplantation, RRT: renal replacement therapy, ADPKD: autosomal dominant polycystic kidney disease, ANCA: antineutrophil cytoplasmic antibody, NOS: not otherwise specified, n.c.: not calculated.

**Table 2.** Data at the time of Renal Transplantation (RTx).


Median and IQR are depicted except for nominal variables, where the number of patients (*n*) and percentages are shown. *p*-values < 0.05 are marked with an asterisk \*. DD: deceased donor, CVA: cerebrovascular event, SAB: subarachnoidal bleeding, SDH: subdural hematoma, ECD: extended criteria donor, CMV mismatch: "−" means sero-negative, "+" means sero-positive, PO HDs: postoperative hemodialysis sessions, RTx: renal transplantation, S-Creatinine: serum creatinine.


**Table 3.** Data at last follow-up.

Median and IQR are depicted except for nominal variables, where the number of patients (*n*) and percentages are shown. *p*-values < 0.05 are marked with an asterisk \*. Data on cerebro- and cardiovascular diseases show cumulative numbers of events at follow-up. De novo events are depicted in Table 5. IS-CNI/mTOR/Bela: immunosuppression concerning tacrolimus (Tac), cyclosporine A (CsA), Belatacept (Bela), mTOR inhibitors (mTORi). IS-Antimetabolites: immunosuppression concerning mycophenolate mofetil (MMF), mycophenolic acid (MPA), azathioprine (Aza).

Concerning efficacy, renal allograft function as assessed by serum creatinine/eGFR improved in Belatacept-treated patients and slightly worsened in CNI patients, yielding a non-significant difference between the groups at follow-up (in contrast to the time of discharge after RTx). ΔeGFR of the patients on Belatacept was positive compared to CNI patients (+2.60 vs. <sup>−</sup>0.89 mL/min/1.73 m2/year, *p* = 0.006). The median ΔeGFR in the whole cohort (Belatacept plus CNI patients, *n* = 184) was + 0.35 mL/min/1.73 m2 (Table 4). The only factor significantly associated with a ΔeGFR above the median in the multivariate model (adjusted for significant factors in the univariate analysis including Belatacept treatment, recipient BMI, number of postoperative HDs, the presence of serum-creatinine at discharge and donors' age) was Belatacept treatment (HR 4.35, 95%CI 2.387–7.926, *p* < 0.001, Table 4). Rejection episodes and graft loss were not significantly different between Belatacept and CNI patients (Table 5). Neither Belatacept nor any other parameter was a significant risk factor for rejection in the univariate Cox regression. Univariate correlated risk factors for graft loss were a higher recipient BMI (HR 1.126, 95%CI 1.038–1.222, *p* = 0.004), number of postoperative HDs (HR 1.178, 95%CI 1.041–1.333, *p* = 0.009) and higher serum-creatinine at discharge (HR 1.598, 95%CI 1.179–2.166, *p* =0.03). BMI (HR 1.116, 95%CI 1.003–1.242, *p* = 0.043) and number of postoperative HDs (HR 1.253, 95%CI 1.027–1.530, *p* = 0.027) remained significant after multivariate adjustments. Belatacept was not a risk factor for graft loss (HR 0.987, 95%CI 0.283–3.417, *p* = 0.980).

**Table 4.** Cox regression for eGFR better than the median (+0.35 mL/min/1.73 m2/year) at last follow-up. We calculated univariate hazard ratios (HR) for risk factors. All significant univariate risk factors were used in the multivariate model. BMI: body mass index, RTx: renal transplantation, PO HD: postoperative hemodialysis session, ECD: extended criteria donor.


<sup>1</sup> adjusted for Belatacept treatment, BMI, number of postoperative (PO) hemodialysis sessions (HDs), eGFR at discharge and donor age. *p*-values < 0.05 are marked with an asterisk \*.



#### *3.4. Safety*

We found no difference between CNI- and Belatacept-treated patients concerning all safety endpoints, except for severe infection (Figure 1, Table 5).

**Figure 1.** KM-Plot for severe infection (Log-Rank *p* = 0.013), defined as infection leading to the admission of the patient to hospital.

Concerning severe infections, defined as infections leading to the admission of the patient to hospital, we found more infections in the Belatacept group (38.2 vs. 23.3%, Log-Rank *p* = 0.013). The type of severe infection also differed between the groups, with a higher proportion of diarrhoea, urinary tract infections and sepsis, but fewer instances of pneumonia, in Belatacept-treated patients. In the univariate Cox regression analysis, Belatacept treatment, number of postoperative HDs, the presence of creatinine at discharge, ECD and donor age were significant risk factors for severe infections (Table 6), while the male sex of the recipient was a protective factor. In the multivariate analysis, no risk factor remained significant (including Belatacept), whereas the male sex of the recipient remained a significant protective factor for severe infection in our cohort. It is noteworthy that Belatacept was not a significant risk factor in any Cox regression analysis for all other safety endpoints. Risk factors for CMV reactivation were number of postoperative HDs (HR 1.123, 95%CI 1.050–1.201, *p* = 0.001) and the presence of serum-creatinine at discharge (HR 1.335, 95%CI 1.087–1.638, *p* = 0.006) of which none remained significant in a multivariate model. A risk factor for BKV reactivation in patients' plasma was treatment with steroids at follow-up (HR 3.358, 95%CI 1.246–9.051, *p* = 0.017) whereas the male sex of the donor was protective (HR 0.362, 95%CI 0.142–0.917, *p* = 0.032). The treatment with steroids at follow-up remained significant in the multivariate model (HR 2.850, 95%CI 1.042–7.796, *p* = 0.041). BKV reactivation in patients' urine was significantly correlated with recipient (HR 1.041, 95%CI 1.013–1.069, *p* = 0.004) and donor age (HR 1.020, 95%CI 1.001–1.038, *p* = 0.036), of which recipient age remained multivariately significant (HR 1.033, 95%CI 1.003–1.064, *p* = 0.029).


**Table 6.** Cox regression for severe infection. All the significant risk factors from the univariate Cox-Regression are shown and were included in the multivariate analysis. *p*-values < 0.05 are marked with an asterisk \*.

<sup>1</sup> adjusted for belatacept treatment, male recipient sex, number of post-operative (PO) hemodialysis sessions (HDs), creatinine ad discharge, extended criteria donor (ECD), donor age.

The only univariate risk factor for the safety endpoint de novo cardiovascular events was pre-existing cerebrovascular disease (HR 6.144, 95%CI 1.026–36.798, *p* = 0.047). All other parameters, and especially Belatacept (HR 1.938, 95%CI 0.214–17.591, *p* = 0.557), were not significant. Concerning malignancy, we found no significant factor in the univariate Cox regression. Univariate risk factors for death were recipient age (HR 1.044, 95%CI 1.006–1.083, *p* = 0.023), HLA mismatch (HR 2.27, 95%CI 1.011–5.099, *p* = 0.047) and the presence of creatinine at discharge (HR 1.350, 95%CI 1.018–1.791, *p* = 0.037). Recipient age was the only significant risk factor for death in multivariate Cox regression (HR 1.036, 95%CI 1.001–1.074, *p* = 0.046).

#### **4. Discussion**

We conducted a retrospective cohort study of renal allograft recipients at high cardiovascular risk either treated with a Belatacept- or CNI-based immunosuppressive regimen. eGFR improved with Belatacept treatment, but slightly decreased during CNI therapy and, in the multivariate analysis, Belatacept treatment was the only significant factor for the improvement of ΔeGFR. This is in line with the BENEFIT and BENEFIT-EXT studies, which also demonstrated an increase of GFR over a follow-up of seven years [12–14]. However, the CNI comparator cohort consisted of cyclosporine A-treated patients only, whereas our group was mainly treated with tacrolimus (68% of patients). Our recipients were of a similar age (56 years) to recipients in BENEFIT-EXT, but older compared to BENEFIT (43 years). Additionally, our CV high risk cohorts consisted of more diabetic patients (CNI group: 57%, Belatacept group: 32% vs. BENEFIT: 15%, BENEFIT-EXT: 16%) and had worse donor characteristics (living donors: CNI group 4%, Belatacept group 15% vs. BENEFIT 58%, BENEFIT-EXT not reported). Furthermore, the BENEFIT studies did not report the number of patients suffering from established cardiovascular disease, which was substantial in our Belatacept (94%) and CNI patients (75%). Nevertheless, and although Belatacept patients had inferior renal allograft function at the time of discharge after transplantation, serum creatinine levels and eGFR were similar at follow-up in this high CV risk cohort compared to CNI-treated patients. Bertrand et al. [15] and Le Meur et al. [16] reported similar results in 17 and 25 patients treated with Belatacept, because of vascular damage and CNI intolerance.

Belatacept was not associated with an increased risk of rejection in our patients. BENEFIT-EXT [17] reported a higher risk in Belatacept-treated patients, whereas BENEFIT [18] found no significant difference. However, our CNI cohort was mainly treated with tacrolimus, which is generally accepted to have a slightly higher immunosuppressive potential, rather than cyclosporine A as in the BENEFIT(-EXT) studies. Belatacept was not a risk factor for graft loss in our cohort (HR 0.987, 95%CI 0.283–3.417, *p* = 0.980), which is in line with the literature [17,18].

Concerning safety and in contrast to Florman et al. [13], we found that Belatacept treatment was associated with an increased incidence of severe infections in the univariate Cox regression. The most obvious differences were higher event rates of diarrhoea, urinary tract infection and sepsis. Sepsis as an endpoint has not been reported in any previous studies of Belatacept patients. In line with our data, gastroenteritis and urinary tract infections were numerically higher in the switch studies [7,8,19]. However, in the multivariate regression analysis, Belatacept treatment was not associated with an increase of this endpoint. One explanation for the higher incidence of gastroenteritis (diarrhoea) might be the more frequent use of mycophenlate in Belatacept patients (76.47 vs. 60.67%; 1.26–fold more), which is known to have such gastrointestinal side-effects. The reason for this is that physicians tend to prescribe the triple combination of mycophenolic acid, steroids and Belatacept, as only this regimen was approved for renal allograft recipients. However, it cannot be ruled out that some examples of mycophenolate-associated diarrhoea have been misdiagnosed as infectious diarrhoea, therefore increasing the proportion of diarrhoea in Belatacept patients compared to CNI patients, although the fold-change of diarrhoea (8.8 vs. 0.7%, 12.57-fold) substantially exceeds the use of mycophenlate in the Belatacept compared to the CNI group (1.26-fold, see above). Additionally, the proportion of mycophenolate-treated patients in the switch studies was almost identical between the groups (and higher compared to our data (approx. 94%)) [7,8,19].

In our cohort, Belatacept treatment was not a risk factor for CMV reactivation, malignancy or death. This is in line with the published data cited above. Additionally, Belatacept treatment was not a risk factor for BKV reactivations either in patients' serum or urine. Unfortunately, data on these endpoints were not reported in the BENEFIT and BENEFIT-EXT studies. Nevertheless, our data is in line with the phase II studies, which showed only slightly increased cumulative incidence rates (0.85 vs. 0 [19]) and events (4 vs. 0% [8] and 2 vs. 0% [7]) in Belatacept patients. Unfortunately, no statistics were calculated in these studies.

Published data suggest a beneficial impact of Belatacept on arterial stiffness and metabolic parameters (e.g., arterial hypertension and lipid profile) or post-transplant diabetes mellitus. The authors concluded that this could improve kidney transplant recipients' survival by reducing events related to those factors [9,17,18,20]. However, available data from the long-term outcomes of these studies do not show a significant difference in severely adverse events (including cardiac or vascular disorders) [12,14]. Concerning patients with high cardiovascular risk, the only study that has been published so far was a post-hoc analysis of patients with pre-existing diabetes of the BENEFIT and BENEFIT-EXT cohorts. Patient survival and renal function were numerically but not significantly higher in Belatacept patients at 12 months' follow-up and, unfortunately, cardiovascular events were not reported [8]. Hence, to the best of our knowledge, this is the first report on cardiovascular events in Belatacept compared to CNI-treated patients. We found no difference between Belatacept- and CNI-treated patients concerning de novo cardiovascular events with a cumulative follow-up of 1194 months in Belatacept (*n* = 35) and 7309 months in CNI patients (*n* = 150).

Our study has limitations. Firstly, this is a retrospective study, which by nature does not provide the same data quality as a prospective design. Secondly, the size of the study population is relatively small, as we included only 34 Belatacept patients and 150 CNI patients as a comparator. Thirdly, the baseline characteristics of time on RRT, primary renal disease, diabetes, cardiovascular disease and arterial hypertension were different between our two populations (Table 1) and it is possible that statistical methods were not able to correct appropriately for this issue. Fourthly, the median time of follow-up was longer in CNI patients (Belatacept: 1054 vs. CNI: 1462 days) but not statistically significant (*p* = 0.084). From our point of view, the duration of follow-up is still significant, although one might argue that a longer follow-up would have been beneficial, especially for the end point "cardiovascular event". However, the number of studies that have published data of renal allograft recipients on Belatacept-based immunosuppression is generally limited. In total, until the end of 2014, the data of 521 Belatacept patients, which were compared to CNI-treated controls, were published [21]. Recently, one study of 17 Belatacept patients matched to 18 control patients, and two studies of 25 and 6 cases that were converted from CNI to Belatacept without a control population, were published [15,16,22]. The randomized controlled trials BENEFIT [18], BENEFIT-EXT [17] and the switch study [8] originally

reported one-year results of 181, 129 and 81 belatacept patients compared to a CNI-treated cohort of similar size. Hence, we believe that our cohort size and follow-up period is considerable and contributes information in a real world setting.

In conclusion, we believe that Belatacept is an efficient, beneficial and safe option for renal allograft recipients at high cardiovascular risk. In our cohort, Belatacept treatment was associated with a superior graft function compared to a CNI-treated cohort and was not a risk factor for renal allograft rejection, -loss, severe infection, CMV- or BKV-reactivation, malignancy or death.

**Author Contributions:** Conceptualization: H.N.; Data curation: H.N., I.L.-L., M.E., F.P., N.P., J.M., H.S.-M., M.R., P.S.; Formal analysis: H.N., M.E., F.P., N.P., J.M., M.R.; Methodology: H.N., J.K., G.M.; Project administration: H.N., G.M.; Software: H.N.; Supervision: H.N.; Validation: H.N., I.L.-L., M.R.; Writing-original draft: H.N., M.R., P.S., I.E.E., G.M.; Writing-review & editing: H.N., I.E.E., G.M.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Analysis of the E**ff**ects of Day-Time vs. Night-Time Surgery on Renal Transplant Patient Outcomes**

### **Nesrin Sugünes 1, Anna Bichmann 2, Nadine Biernath 1, Robert Peters 1, Klemens Budde 3, Lutz Liefeldt 3, Thorsten Schlomm <sup>1</sup> and Frank Friedersdor**ff **1,\***


Received: 30 June 2019; Accepted: 12 July 2019; Published: 18 July 2019

**Abstract:** Sleep deprivation and disruption of the circadian rhythms could impair individual surgical performance and decision making. For this purpose, this study identified potential confounding factors on surgical renal transplant patient outcomes during day and night. Our retrospective cohort study of 215 adult renal cadaver transplant recipients, of which 132 recipients were allocated in the "day-time" group and 83 recipients in the "night-time" group, primarily stratified the patients into two cohorts, depending on the start time. Within a 24 h operational system, "day-time" was considered as being from 8 a.m. to 8 p.m. and "night-time" from 8 p.m. to 8 a.m.. Primary outcomes examined patient and graft survival after three months and one year. Secondary outcomes included the presence of acute rejection (AR) and delayed graft function (DGF), as well as the rate of postoperative complications. In log-rank testing, "day-time" surgery was associated with a significantly higher risk of patient death (*p* = 0.003), whereas long-term graft survival was unaffected by the operative time of day. The mean cold ischemia time (CIT), which was 12.4 ± 5.3 h in the "night-time" group, was significantly longer compared to 10.7 ± 3.6 for those during the day (*p* = 0.01). We observed that "night-time" kidney recipients experienced more wound complications. From our single-centre data, we conclude that night-time kidney transplantation does not increase the risk of adverse events or predispose the patient to a worse outcome. Nevertheless, further research is required to explore the effect of fatigue on nocturnal surgical performance.

**Keywords:** night-time renal transplantation; graft survival; patient survival/outcome; surgical complications

#### **1. Introduction**

Kidney transplant outcomes have improved in recent years through novel technical approaches and immunosuppressive therapy [1–4]. There is still a deleterious impact of surgical complications on graft and patient survival [5,6]. Several risk factors of surgical complications have been identified, including donor and recipient characteristics, organ recovery and surgical implantation techniques [7]. Recipients with a prolonged cold ischemia time (CIT) have a greater risk for delayed graft function (DGF) and diminished long-term allograft survival [8]. To reduce CIT, surgery is initiated at any time of the day to preserve the organ quality. Further risk factors which are detrimental for patient outcome are human factors, including physical and mental fatigue and sleep deprivation, which are known to affect communication, attention and situational awareness, as well as psychomotor function [9,10]. It has been hypothesized that sleep deprivation reduces the performance of surgeons by affecting cognitive and fine motor skills [11,12]. In a technically demanding field, such as renal transplantation, meticulous preparation and excellent suturing techniques are required to prevent vascular and urologic complications [7]. The impact of physician fatigue on the medical error rate and clinical outcomes has been actively researched [13–15]. A number of studies demonstrated that operative outcomes were not related to sleep deprivation [16–18], whereas others link mental fatigue to surgical complication rates after general procedures [19]; and mortality after liver transplants [20]. To our knowledge, the literature regarding the impact of night-time surgery on outcomes after kidney transplantation is underrepresented and recent studies have reported conflicting results [21–24]. For this purpose, we conducted a retrospective cohort study to examine the association between the time of day of transplantation surgery (night-time vs. day-time) on surgical renal transplant patient outcomes. The primary outcomes examined were patient and graft survival after three months and one year. Secondary outcomes included the presence of acute rejection (AR) and DGF and other postoperative complications. We hypothesize, that renal transplantation surgery performed during the night-time would have inferior outcomes compared to those performed during the day.

#### **2. Methods**

We performed a retrospective cohort study of all adult patients undergoing cadaver renal transplantation at Charité University Hospital Campus Mitte, between 01.01.2011 and 31.12.14. Data on kidney transplantation and operative variables, as well as follow-up data, were obtained retrospectively from internal SAP (System, Anwendung, Produkte) and national TBase (Kidney Transplant Information System) electronic databases. The entire analysis was in adherence with correct scientific research work terms of the Charité Medical University of Berlin including full anonymization of patient data ('Good Scientific Practice', version 29/03/18).

#### *2.1. Study Population*

Transplants were stratified by the operative time of day. "Day-time" surgery was defined as surgery that started between 8:00 a.m. and 8:00 p.m. and "night-time" surgery was defined as surgery that started between 8:00 p.m. and 8:00 a.m. Twelve surgeons performed all the transplantations using standard surgical techniques. Kidneys were placed either in the right or left iliac fossa via an extraperitoneal approach. The renal graft vessels were anastomosed end-to-side to the recipient external or common iliac vessels. In all cases, except for one patient with urinary diversions (ileal conduit), a standard Lich–Gregoir ureteroneocystostomy was performed. A double-J ureteral stent was systematically inserted and removed six weeks later, followed by a urethral catheter for ten days postoperatively. All recipients received intravenous prophylactic antibiotics at the time of transplant. Graft function was monitored by Doppler ultrasound scanning, serum creatinine level and urine output measurements. The routine immunosuppression protocol that was initiated consisted of a triple regimen, including calcineurin inhibitors or a mammalian target of rapamycin inhibitors, mycophenolate mofetil (MMF), and steroids.

#### *2.2. Data Collection*

Patient and donor demographics and clinical data were collected by chart review. The parameters evaluated in this study were, recipient characteristics of age, gender, body mass index (BMI), comorbidities (hypertension, diabetes mellitus, cardiovascular disease, stroke and peripheral vascular disease), previous abdominal surgery, causation of end-stage renal disease, previous transplantation, duration of pre-transplant dialysis, and human leukocyte antigen (HLA) mismatches. The donor features were age, gender, BMI, site of donor kidney, number of graft arteries and the presence of graft vessels atherosclerosis. Perioperative factors included the surgeon's experience (consultant, resident), cold and warm ischemia time (WIT), and incidence of intraoperative complications. CIT was defined as the time between the start of cold perfusion and removal of the renal allograft from ice. Warm

ischemia time was defined as the time between the placement of the renal allograft into the iliac fossa of the recipient until revascularization of the kidney occurred.

#### *2.3. Outcome Measures*

The primary outcomes examined were patient and graft survival after three months and one year, respectively. Secondary outcomes included the presence of AR and DGF, as well as the rate of postoperative complications. Postoperative complications were examined for the first three months after surgery and defined according to the Clavien–Dindo classification system [20].

#### *2.4. Statistical Analysis*

Univariate comparisons were performed using the Chi-Square test or Fisher exact test for categorical variables. Continuous variables were tested with the non-paired Student t-test and the Mann–Whitney-U test for data with non-normal distribution. Categorical variables were displayed as n (%) and continuous variables mean ± standard deviation (SD); and nonparametric distribution as median (minimum-maximum). Patient and allograft survival rates were estimated using the Kaplan–Meier method and comparisons of survival rates were performed using the log-rank test. For all statistical measures, a *p*-value below 0.05 was considered statistically significant. Statistical analyses were performed using SPSS software (SPSS Inc., version 25, Armonk, NY, USA).

#### **3. Results**

The baseline characteristics and operative parameters in the two groups stratified according to the time surgery was performed are presented in Table 1. The two groups were similar with respect to most of the baseline characteristics, except for the higher distribution of male donors for the "day-time" group (*p* = 0.05). The mean CIT was 12.4 ± 5.3 h in the "night-time" group compared with 10.7 ± 3.6 for the "day-time" cohort (*p* = 0.01). The total operative time from skin incision to wound closure was similar in kidney transplants performed at all times. Considering the surgical expertise, 76.5% of "day-time" procedures were performed by a consultant compared to 72.3% during the "night-time" (*p* = 0.49). A total of six intraoperative surgical complications occurred in the overall cohort of 215 recipients (2.8%): renal artery stenosis (*n* = 2), renal vein injury (*n* = 2), renal vein thrombosis (*n* = 1) and iatrogenic bladder perforation (*n* = 1), which were immediately treated. The difference in incidence of intraoperative surgical complication was statistically insignificant with 3.8% (*n* = 5) during the day and 1.2% (*n* = 1) during the night (*p* = 0.34). We observed a higher incidence of DGF nocturnal operations with 54.2% compared to 47.7% in the "day-time" group (*p* = 0.35). The incidence of AR was 25% for "night-time" compared to 22% for "day-time" allograft recipients (*p* = 0.57). Table 2 shows patient outcomes.


**Table 1.** Recipient and donor characteristics and operative details. Results are presented as mean and standard deviations or as absolute and relative frequencies; h—hours; min—minutes; ESRD—end-stage renal disease; \* statistically significant.


**Table 1.** *Cont.*

**Table 2.** Graft and recipient outcome, \* statistically significant.


#### *3.1. Patient and Graft Survival*

The Kaplan–Meier survival curves for patient survival (Figure 1a) and death-censored allograft survival (Figure 1a) by status are shown in Figure 1. In log-rank testing, "day-time" operation was

associated with a significantly higher risk of patient death (log-rank test 5.65; *p* = 0.017). During the first 12 months after surgery, a total of nine deaths occurred in the overall sample of 215 kidney transplants recipients (4.18%). No death occurred in the "night-time" kidney group within one year of transplantation, whereas two of the 132 "day-time" renal recipients died with a functioning transplant (one case of coronary heart disease, and one of malignancy) and seven patients died after returning to dialysis (all cases due to bacterial sepsis). Kaplan–Meier analyses demonstrated no statistically significant differences for death-censored graft survival (Figure 1b) between the "night-time" and "day-time" recipient cohorts (log-rank test 0.014; *p* = 0.907).

**Figure 1.** (**a**) Kaplan–Meier survival curve for patient survival after "day-time" and "night-time" renal transplantation (log-rank test 5.65; *p* = 0.017); (**b**) Kaplan–Meier survival curve for death censored graft survival of patients after "day-time" and "night-time" renal transplantation (log-rank test 0.014; *p* = 0.907).

#### *3.2. Early Graft Failure*

During the first three months post-operation, graft failure was noticed in seven out of 132 "day-time" allograft recipients (5.3%), and in two out of 83 (2.4%) "night-time" transplant recipients (*p* = 0.49). In the "day-time" cohort, the most common cause of graft failure was primary non-function (*n* = 3), whereas recurrent disease, sepsis and death with functioning graft were noticed in the other cases, respectively. One "day-time" renal transplant recipient suffered an invasive fungal infection, which produced an allograft vessels aneurysm leading to graft loss. From the "night-time" group, two recipients (2.4%) lost the graft during the first three months after transplantation, due to AR and graft infection.

#### *3.3. Postoperative Complications*

One or more postoperative complications occurred in 74 out of the 132 "day-time" renal transplant recipients (56%) compared to 41 of the 83 "night-time" allograft recipients (49%) during the first three months post-operation (*p* = 0.34). Each category of complication assessed by the Clavien–Dindo grading system was analysed separately against the two-time groups, of which no category was significantly different (Table 3). In particular, "night-time" and "day-time" renal recipients did not differ significantly in the incidence of postoperative complications requiring medical or surgical reintervention (Clavien–Dindo Grade IIIa/b). Table 4 shows the number of operations that were performed within each time period and the incidence of surgical complications. The most common surgical complications in both groups included haemorrhagic events requiring blood transfusions or surgical intervention (17.2%), lymphoceles (10.7%), seromas (9.7%), and wound dehiscence (7%). A statistically insignificant higher incidence of wound complications among "night-time" kidney recipients was observed. The incidence of urologic complications was higher for the "day-time" surgery, which was also statistically insignificant. Among the 12 patients with urological complications, nine (6.8%) occurred within the "day-time" group and three (3.6%) during the "night-time" group. Five patients (2.3%) were treated with interventional procedures and two (0.7%) received surgical intervention under general anaesthesia. Ureteric necrosis occurred in one "day-time" renal recipient, which was treated with ureteric re-implantation. The incidence of vascular complications within three months post-transplantation was, respectively, 4.5% for "day-time" and 2.4% for "night-time" surgery. In four cases (1.9%), an early secondary surgical intervention was required for vascular complications. Renal artery stenosis occurred in 0.9% of all recipients. Renal artery aneurysm and renal vein thrombosis occurred equally at the rate of 0.5%.


**Table 3.** Postoperative complications with Clavien–Dindo Classification. Results are presented as absolute and relative frequencies. \* If more than one occurred per case, according to patient records, the complication with the highest degree was selected (Minor I+II, Major complications IIIa-IVb, Mortality V).

**Table 4.** Incidence of surgical complications. Incidence is expressed as percentages (%) of total number (*n*) of patients.



**Table 4.** *Cont.*

#### **4. Discussion**

Over the past decade, increased understanding of the effects of shift work and sleep deprivation on neurocognitive functions and physicians health has been established [25]. A single-center study by Rothschild et al. suggested that surgical outcomes were compromised if surgeons had less than six hours of sleep per shift [19]. Traffinder et al. reported that fatigued surgeons made 20% more errors and took 14% longer to perform laparoscopic tasks [26]. On the other hand, studies have demonstrated that outcomes of surgical procedures may not be adversely affected by fatigue or disruption of the normal circadian rhythm [16–18]. Five studies with limited numbers of transplants have previously assessed this issue by focusing on the impact of night-time surgery on graft outcome or complications in patients undergoing renal transplantation [21–24]. Only one single-center study, performed by Fechner et al., demonstrated that night-time surgery carries a higher risk of adverse events and poorer outcomes, particularly driven by higher rates of vascular complications [21]. Kienzel et al. reported that, if transplantations were postponed until the next morning, the increase in CIT would decrease the long-term survival [22]. Seow et al. did not observe an adverse effect of night-time surgery on patient outcomes but highlighted surgical clinical expertise to be a crucial factor for surgical complications [23]. Several limitations need to be considered in the interpretation of the contradictory results. Most studies published to date reported great variability in the methodology and outcome measures. In addition, the definition and understanding of sleep deprivation varied widely among previous investigators. Mentioned studies are frequently single-center and reported the results of a small groups of surgeons, which limits the generalizability. In the present study, we did not find any significant impact of night-time kidney transplant surgery on outcomes including three-month and one-year patient or allograft survival, postoperative complications, DGF or AR in the first year. Our analysis revealed a variable incidence of complications among the different time groups and we could not determine any consistent trend. While the incidence of vascular, haemorrhagic and urological complications was greatest in the "day-time" operative group, wound complications occurred more often among recipients of "night-time" transplants without statistical significance. The mean CIT was slightly longer among those who underwent night-time transplant operations compared to the "day-time" cohort. We observed diminished patient survival among "day-time" renal transplant recipients compared to "night-time" allograft recipients, whereas long-term graft survival was unaffected by the time of day. With no significant difference in baseline characteristics, except for the slighter higher distribution of male donors in the "day-time" cohort, the reasons for this observation are still unclear. We have controlled for a majority of the clinically meaningful variables available to us in this data set, but it is possible that yet unidentified biologic factors could account for the difference in patient survival between the "night-time" and "day-time" cohorts. With this in mind, there is an urgent need for research in order to clarify the biological consequences of sleep disturbance and fatigue among renal transplant patients. The influence of circadian rhythmicity on physiologic functions related to renal cells, including blood pressure control and homeostasis regulation, is a well-studied phenomenon [27–30]. Evidence suggests that reduced sleep duration and disturbed circadian rhythms may increase sympathetic nervous system stimulation, increase blood pressure, and impair metabolic regulation [31–33]. Thus, misalignment of intrinsic circadian rhythms with environmental time may contribute to poor kidney functioning and renal injury among kidney transplant recipients and donors. Future study is required to clarify this issue. There may be several possible explanations for the lack of 'night-time effect' on outcomes after renal transplantation in this study. Recent studies have demonstrated that there is inter-individual variability in vulnerability to cognitive deficits from sleep loss and the ability to sustain effective neurocognitive performance [34,35], suggesting a reason why there were no differences between the "day-time" and "night-time" cohorts in our study. Van Dongen et al. reported differences in endogenous regulatory processes among individuals, which may affect their tolerance for shift work and cognitive performance during work shifts [36]. Performance adaption across successive shifts has been observed [37,38]. Leff et al. suggested improvement in technical procedural skills across remaining night shifts may be due to ongoing learning or adaption to chronic fatigue [37]. When considering the impact of nocturnal shift work on surgical performance, it is essential to also consider the effects of societal and environmental forces that may contribute to the biological consequences of circadian misalignment. It is known, that there is a detrimental effect of noise inside the operating room on the performance of surgeons and anaesthesiologists [39]. The exposure to excessive operating room noise and distractions during the main day-time business hours may impair cognitive skills. Other factors influencing the performance of a surgeon, such as leadership and communication may be at least as important as technical skills and the number of hours slept [40]. In addition, the use of caffeine and periods of short naps may mitigate the potential risks associated with sleep deprivation [41]. A study of this nature has some limitations, primarily through its retrospective design. The small overall number of patients and individual complications in our cohort might weaken the conclusions of our pilot-study and limits the power to detect differences. To assess severity, we additionally categorised all postoperative complications using the Clavien–Dindo classification system. Although this system has been proven to be reproducible and applicable with minimal interobserver variability, it has some limitations [42]. Data regarding a surgeon's subjective perception of fatigue, resting time and quantification of sleep deprivation were not available and could not be included in the analysis. It is further possible, that transplant surgeons perform day-time procedures beginning at 8 am after being 'on-call' overnight. With that in mind, one may argue whether the classification based on time group selections assumes that day-time surgeons are well rested, and perform better than night-time surgeons regardless of their overall workload. We cannot lose sight of other potential variables such as the effect of procurement-related organ lesions on renal transplant outcome. Data concerning surgeons' fitness before procurement were not available. Further investigation is needed aiming to record errors during organ procurement related to surgeons' fatigue.

#### **5. Concluding Remarks and Future Directions**

To date, there are very few reports on the effect of night-time surgery on renal transplant outcomes. We, therefore, believe that the initial results from this pilot-study are a welcome addition to the urological literature and provide encouragement for further analysis. We concluded that night- time kidney surgery does not carry a higher risk of adverse events and poorer outcome among patients undergoing renal transplantation. Consequently, kidney transplantation should be immediately performed regardless of the time of the day, with the known adverse effects of prolonged CIT. However, in order to fully assess the effects of sleep deprivation and circadian rhythm disturbance on surgical

performance in kidney transplantation, prospective research involving larger cohorts is needed. Therefore, among other things, a transparent evidence-based assessment of the level of fatigue, shift intensity and sleep quality in medicine, especially in the field of surgery, is required. Moreover, systems-based interventions, as well as individual coping strategies and experiences that mitigate the effects of fatigue and disruption of the circadian rhythms, should be taken into consideration. In addition, there is a need for future research focusing on the impact of sleep displacement and circadian misalignment on renal functioning among recipients and donors in the field of kidney transplantation.

**Author Contributions:** F.F., K.B. and L.L. designed the study; N.S. analyzed the data and wrote the manuscript; A.B., R.P., T.S. and N.B. drafted and revised the paper; all authors approved the final version of the manuscript.

**Acknowledgments:** We acknowledge support from the German Research Foundation (DFG) and the Open Access Publication Fund of Charité–Universitätsmedizin Berlin.

**Conflicts of Interest:** The authors have no conflict of interest to disclose.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
