The effect of intraoperative inosine infusion on transplant outcomes in deceased-donor kidney transplant recipients

Abstract

Objectives

It is unknown when inosine was first employed as a renoprotective agent in the context of kidney transplantation procedures. However, there is no clinical evidence to support a protective role of inosine. The aim of this study was to investigate the effect of inosine on graft recovery.

Patients and methods

Data related to donors and recipients were retrieved from relevant records between 2015 and 2023. A total of 1138 kidney transplant cases were identified, including 1005 recipients who received a bolus of 1000 mg inosine and 133 recipients who did not receive inosine during transplantation surgery. The endpoints of the analysis included recipient recovery after transplantation as assessed by delayed graft function (DGF), peak estimated glomerular filtration rate (eGFR) after transplantation, and unfavorable graft function recovery.

Results

Given the high dimensionality of the donor and recipient variables, propensity score weighting analyses were conducted. No significant differences in the risk of DGF (OR = 0.80 [0.52, 1.22], p = 0.301), unfavorable graft function recovery (OR = 0.95 [0.61, 1.51], p = 0.842) or peak eGFR after transplantation (β = 1.61 [−4.33, 7.56], p = 0.594) were observed between the inosine and no-inosine groups via overlap weighting analysis.

Conclusions

Intraoperative infusion of 1000 mg of inosine has no effect on graft recovery after kidney transplantation. Therefore, the practice of using inosine during kidney transplantation surgery is not supported by evidence.

Introduction

Ischemia-reperfusion injury (IRI) represents a significant challenge in kidney transplantation. IRI can lead to delayed graft function, acute and chronic rejection, and reduced graft survival [1,2]. The current use of intraoperative renoprotective agents, including antioxidants, is usually based on the dogma held by transplant surgeons. Inosine has been reported to play a protective role in inflammation and IRI [3–7]. As outlined by Kim et al. [8], inosine has been the subject of extensive investigation in animal models of diseases, including neuropsychological disorders, infectious diseases, cancer and cardiovascular diseases. Inosine has long been used as a renoprotective agent in kidney transplantation procedures [9]. In animal models, inosine markedly attenuates cardiac reperfusion injury by increasing myocardial and endothelial function in both cardiopulmonary bypass and heart transplantation surgeries [10,11]. The impact of inosine on kidney ischemia injury in canine models has yielded inconclusive results, as evidenced by the discrepancies observed in both the in situ kidney warm ischemia injury model [12,13] and the kidney warm ischemia-cold storage-autotransplantation model [14,15]. Nevertheless, there is limited clinical evidence to support the claim that inosine exerts a protective effect on graft recovery. The objective of this study was to examine the impact of intraoperative inosine infusion on graft recovery in a substantial kidney transplantation cohort.

Patients and methods

Study design and patients

In our transplant center, a total of 1323 deceased kidney transplantation cases were retrospectively retrieved from January 1, 2015, to December 31, 2023. Cases involving recipients who were under the age of 18, had undergone multiorgan transplantation, or were lost to follow-up were excluded from the analysis. Figure 1 displays the detailed process. All the donor sources were deceased citizen donors, as this has been the sole legal procedure for the procurement of solid organs from nonrelative sources for transplantation in China since 1 January 2015. All the donor death-related profiles were extracted from the China Organ Transplant Response System (COTRS), in which all donors and procured organs are numbered, matched, and allocated (https://www.cot.org.cn/). Cases allocated from other organ procurement organizations (OPOs) were excluded if the relevant donor profile was not accessible. This study is based on a clinical cohort study of kidney transplantation that was approved by the Research Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University (Institute Review Board approval: [2020]02-243-01). This research was performed in accordance with the Declaration of Helsinki. The reported clinical and research activities are consistent with the principles of the Declaration of Istanbul, as outlined in the 'Declaration of Istanbul on Organ Trafficking and Transplant Tourism'.

Figure 1.

Diagram of the process used to screen the included cases. A total of 31 cases were excluded from the graft function analysis because of early death or loss to follow-up. DD: deceased donor; Tx: transplantation.

Data sources and variables

All the donor profiles were sourced from the COTRS. Recipient profiles were obtained from electronic medical records. Regarding inosine usage during transplantation surgery, inosine was typically administered as a single additional 100 mg:2 mL*10 ampoules inosine injection. Furosemide was administered according to the recipient urine output after transplantation. The variables in the donor dataset included donor age, sex, body mass index (BMI), terminal serum creatinine level, cause of death, donor death category, donor history of hypertension and diabetes, and cold ischemia time. The interval between donor cardiac death and vessel unclamping during the transplantation operation was estimated as the cold ischemia time, with the initial cold perfusion in donors disregarded. All the donated kidneys were stored under static cold conditions and transported. No warm ischemia injury occurred during implantation because vessel anastomosis was performed under cold conditions, with the kidneys surrounded by crushed ice. The expanded criteria donor (ECD) was defined in accordance with Rao's definition [16]. The recipient variables, including the year of transplantation, the recipient peak panel reactive antibody (PRA) level, age, sex, BMI, history of transplantation and diabetes, dialysis vintage, preoperative creatinine level, initial calcineurin inhibitor (CNI) usage, and induction therapy, were adjusted for in the analysis. Furthermore, the year of transplantation was identified as an independent risk factor for delayed graft function (DGF) because of the increasing acceptance of marginal kidneys over time. The proportion of ECDs was 13.83% during the 2015–2017 period, 22.82% during the 2018–2020 period and 35.46% during the 2021–2023 period (not included in the table). Given that the concept of ECD is a comprehensive derivative of donor age, terminal serum creatinine level, cause of death and hyper­tension history, it was deemed unnecessary to include this repetitive variable in the analysis, in addition to the aforementioned factors.

Outcomes

The primary outcomes of this study were delayed graft function (DGF), graft function (peak eGFR) within the first 12 months, and unfavorable graft function recovery. DGF was defined as the need for dialysis during the first week after transplantation [17]. Graft function was defined as the peak eGFR of the recipient during the follow-up surveillance period. Cases that were censored (death after surgery or loss to follow-up) within the first month were excluded from the analysis, as the eGFR is unstable and ambiguous in such cases. The eGFR was calculated with the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2009 equation [18]. Unfavorable graft function recovery was defined as a stable eGFR that failed to reach a level of 45 mL/min/1.73m².

In this study, we adopted an arbitrary binary outcome of 'unfavorable graft function recovery', which means that the stable eGFR of the recipient could not reach 45 mL/min/1.73m². The cutoff was adopted on the basis of previous publications indicating that recipients with stable eGFRs below 45 mL/min/1.73m² had significantly greater graft loss and risk of death. Furthermore, the risk of graft failure increased with decreasing eGFR. Previous reports [19–22] on the risk of graft failure based on 3-month or 12-month posttransplant eGFRs indicated that the risk of graft failure increased with decreasing eGFR. In accordance with the classification of CKD stages, the risk of graft loss increased steeply with eGFRs below 30 mL/min/1.73m² and exhibited a plateau increase with eGFRs between 30 and 44 mL/min/1.73m². There was an imperceptible increase with eGFRs in the range of 45 to 59 mL/min/1.73m². This result aligns with the classification of CKD stages on graft failure reported in previous studies [23]. It is therefore reasonable to propose that a classification cutoff of 45 mL/min/1.73m² should be adopted. The peak eGFR is another surrogate endpoint for the evaluation of graft function potential. We did not utilize traditional eGFRs at specific time points, such as 3, 6 or 12 months because defining an eGFR at these time points is subjective, given the numerous test outcomes that exist at approximately these times. Selecting any fluctuating result could introduce bias. Furthermore, complications such as rejection, infection or other issues commonly occur at these time points, which can affect the eGFR results. While the peak eGFR represents the graft's potential in the recipient, it is an objective and unique outcome. In fact, the peak eGFR and 3-month, 6-month and 12-month eGFRs are strongly correlated. Therefore, we used the peak eGFR, which is free of dialysis or hypometabolism, such as infection or malnutrition after transplantation, as an objective index for evaluating graft recovery. In summary, we used DGF, peak eGFR after transplantation and unfavorable graft function recovery to evaluate the recovery of the graft from different perspectives. DGF indicates the recovery process of the graft; the peak eGFR indicates the extreme potential of the graft in the recipient's body; and unfavorable graft function recovery was used to binarily classify the ultimately recovered graft function status.

Statistical analysis

Categorical variables are presented as counts and percentages, with group comparisons conducted via Pearson's chi-square test with Yates' continuity correction. Continuous variables are expressed as the means and standard deviations (for normally distributed data) or medians and interquartile ranges (for skewed data), and comparisons were conducted via t tests or Wilcoxon rank-sum tests. Given the multitude of variables associated with both donors and recipients, as well as the relatively limited number of events within each group, propensity score weighting analysis was performed to investigate the treatment effect [24]. Two widely-used weighting methods have been employed, namely, Heman stabilized inverse probability treatment weighting (sIPTW) and overlap weighting (OW) [25]. The sIPTW was calculated as follows: Pt/PS for the treatment group (inosine) and (1-Pt)/(1-PS) for the control group (no inosine). The OW was calculated as follows: 1-PS for the treatment group (inosine) and PS for the control group (no inosine). Pt indicates the treatment rate of the entire cohort, whereas PS indicates the propensity score for treatment. The propensity score for inosine usage was calculated with a number of variables, including donor age, donor sex, donor BMI, donor death category, donor diabetes, donor hypertension history, donor death causes, donor terminal creatinine level, cold ischemia time, recipient transplant year, recipient peak PRA, recipient sex, recipient age, recipient BMI, recipient transplant history, recipient diabetes, recipient dialysis time, recipient creatinine level before transplantation, recipient induction therapy, and recipient initial CNI category. To investigate the potential correlation between inosine usage and recipient DGF, peak eGFR recovery, and unfavorable graft function recovery, generalized linear models with clustered-robust standard errors were employed. These models accounted for the synchrony effect in posttransplant recovery of the mate kidneys [26,27]. The donor ID was set as the cluster variable. All the statistical analyses were performed with Stata 16.1 IC (StataCorp, College Station, TX) and R version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). A two-sided P value of less than 0.05 was considered to indicate a statistically significant difference.

Results

Baseline demographics

A total of 1138 kidney transplantation records were ultimately included from our center from 2015 to 2023. Most of the recipients had routinely received an inosine injection before unclamping the graft vessels, while 133 recipients did not receive any inosine. The screening process is shown in Figure 1. The characteristics of the two groups are listed in Table 1. The inosine group was significantly different from the no-inosine group in terms of donor age, donor BMI, number of cold ischemia hours, donor death category (DCD/DBD), donor terminal creatinine level, kidney transplantation year, and initial CNI category. The cases in which inosine was not administered mainly occurred in the last three years. Previously, inosine was utilized as a standard practice in kidney transplantation. However, the use of inosine was gradually discontinued because of the absence of evidence demonstrating its efficacy. The decision to use inosine was made at the discretion of transplant surgeons; thus, the follow-up periods were significantly shorter in the no-inosine group. Additionally, the distributions of transplant outcomes were significantly different. The no-inosine group presented a higher DGF rate (49/133 vs. 210/1005), higher unfavorable graft function recovery rate (36/125 vs. 180/982), and a lower peak eGFR after transplantation (62.6 ± 28.5 vs. 68.8 ± 25.4).

Table 1.

Baseline demographics before and after stable inverse probability treatment weighting (sIPTW).

	Before weighting				After weighting
Donor variables	No (n = 133)	Yes (n = 1005)	P	SMD	No (n = 142.48)	Yes (n = 980.28)	P	SMD
Age (years)	47.1 ± 12.3	43.8 ± 14.2	0.011 &	0.248	44.5 ± 11.5	44.1 ± 14.1	0.803 &	0.030
Sex (male/female)	104/29	778/227	0.926*	0.019	110.4/32.1	763.0/217.3	0.952*	0.009
BMI (kg/m2)	24.0 ± 2.9	23.2 ± 3.3	0.008 &	0.257	23.3 ± 2.6	23.2 ± 3.27	0.851 &	0.022
Cold ischemia time (hours)			<0.001*	0.791			0.648*	0.186
<8 h	27	381			58.3	360.5
8 ∼ 11.9 h	34	339			47.6	329.1
12 ∼ 15.9 h	9	117			18.4	111.9
≥16 h	7	64			4.3	62.0
unknown	56	104			13.9	116.9
Death category (DCD/DBD)	32/101	470/535	<0.001*	0.489	68.4/74.0	440.2/540.1	0.660*	0.063
Diabetes (yes/no)	5/128	42/963	>0.999*	0.021	6.9/135.5	39.5/940.8	0.811*	0.041
Terminal serum creatinine level (mg/dL)	1.60 [0.96, 3.13]	1.40 [0.88, 2.36]	0.010 #	0.272	1.49 [0.86, 2.36]	1.41 [0.88, 2.44]	0.615 #	0.095
Donor hypertension history (yes/no)	38/95	287/718	>0.999*	<0.001	38.3/104.2	280.1/700.2	0.772*	0.037
Donor death causes			0.720*	0.076			0.959*	0.044
Others	11	76			9.4	74.7
Trauma	45	370			52.4	365.2
CVDs	76	535			80.7	540.3
Recipient variables
Transplant year			<0.001*	0.937			0.645*	0.122
2015 ∼ 2017	12	272			39.6	251.2
2018 ∼ 2020	28	454			66.0	421.9
2021 ∼ 2023	93	279			36.8	307.2
Peak PRA > 0% (yes/no)	14/119	134/871	0.443*	0.087	14.4/128.1	125.9/854.4	0.617*	0.086
Age (years)	43.6 ± 12.5	42.8 ± 11.0	0.447 &	0.067	41.9 ± 11.4	42.8 ± 11.0	0.578 &	0.077
Sex (male/female)	97/36	696/309	0.443*	0.081	96.7/45.7	682.7/297.6	0.796*	0.038
BMI (kg/m2)	22.7 ± 4.1	22.3 ± 3.7	0.187 &	0.117	22.3 ± 4.1	22.3 ± 3.8	0.994 &	0.001
Transplant history (yes/no)	9/124	32/973	0.066*	0.165	4.7/137.8	34.0/946.3	0.932*	0.008
Diabetes (yes/no)	23/110	137/868	0.313*	0.101	22.0/120.4	135.7/844.6	0.764*	0.046
Dialysis months			0.502*	0.146			0.432*	0.210
Preemptive	14	94			8.4	91.8
1 ∼ 6 months	23	229			42.2	215.6
7 ∼ 12 months	33	217			26.9	217.1
>12 months	63	465			64.9	455.8
Pretransplant creatinine level (mg/dL)	10.23 [8.28, 12.33]	10.76 [8.51, 13.16]	0.092 #	0.136	10.74 [9.06, 12.34]	10.72 [8.47, 13.10]	0.756 #	0.042
Induction therapy			0.719*	0.048			0.625*	0.060
ATG/ALG	120	892			123.3	867.5
IL2-R	13	113			19.2	112.8
Initial calcineurin inhibitors			0.037*	0.228			0.391*	0.140
Tac	124	868			116.9	853.7
CsA/SRL	9	137			25.6	126.6

Abbreviations: BMI, body mass index; ECD, expanded criteria donor; SCD, standard criteria donor; DCD, donor after cardiac death; DBD, donor after brain death; CVDs, cerebrovascular diseases; PRA, panel reactive antibody; ATG/ALG, anti-thymus globulin/anti-lymphocyte globulin; IL2-R, interleukin-2 receptor antagonist, basiliximab (Simulect^®); Tac, tacrolimus; CsA, cyclosporin A; SRL, sirolimus.

For continuous variables, the results are presented as the means ± standard deviations for normally distributed data and medians [interquartile ranges, Q1 and Q3] for skewed data. *, Pearson's chi-square test with Yates' continuity correction; #, Wilcoxon rank-sum test; &, unpaired t-test.

Propensity score analysis

The propensity score was calculated using the aforementioned variables. Figure 2 illustrates the propensity score for each group. Following sIPTW, the balance of these two groups demonstrated a favorable outcome with SMDs below 0.1, with the exception of two variables, which presented SMDs above 0.2. To minimize the potential for bias due to extreme propensity scores, we employed overlap weighting for the analysis. Figure 3 shows the balance achieved before and after propensity score weighting. The OW method demonstrated superior balance to sIPTW.

Figure 2.

The Distribution of propensity scores for each group revealed the presence of extreme values.

Figure 3.

The Baseline balance across the inosine and no-inosine groups was assessed both before and after propensity score weighting. The results demonstrated that overlap weighting (OW) provided a more precise balance than stable inverse probability treatment weighting (sIPTW). A standard mean difference (SMD) less than 0.1 was considered to indicate an optimal balance between the two comparative groups.

Graft recovery

In this study, three endpoint events were used to evaluate the recovery of the transplant kidneys. These events covered the recovery process and the final recovery status of the graft. The three endpoints included DGF, peak eGFR after transplantation and unfavorable graft function recovery (stable eGFR < 45 mL/min/1.73m²). According to the univariable analysis (unweighted results), the use of inosine significantly reduced the incidence of DGF (OR = 0.45, 95%CI [0.31, 0.67], p < 0.001). However, after propensity score weighting, no significant correlation was observed between inosine usage and DGF incidence in either the sIPTW analysis (OR = 0.89, 95%CI [0.50, 1.57], p = 0.688) or the OW analysis (OR = 0.80, 95%CI [0.52, 1.22], p = 0.301). Similarly, in the unweighted analysis, inosine use was associated with superior peak eGFR recovery (β = 6.50, 95%CI [1.06, 11.95], p = 0.019) and a reduced risk of unfavorable graft function recovery risk (OR = 0.55, 95%CI [0.36, 0.85], p = 0.007). After propensity score weighting, no significant correlation was identified between inosine use and either peak eGFR recovery or unfavorable graft function recovery. The detailed results are presented in Table 2.

Table 2.

Effect of intraoperative inosine administration on DGF, peak eGFR and graft function category (propensity score weighting).

	Unweighted		sIPTW		OW
	Effect and 95% CI	P value	Effect and 95% CI	P value	Effect and 95% CI	P value
DGF*	0.45 [0.31, 0.67]	<0.001	0.89 [0.50, 1.57]	0.688	0.80 [0.52, 1.22]	0.301
Peak eGFR #	6.50 [1.06, 11.95]	0.019	4.10 [-4.23, 12.43]	0.334	1.61 [-4.33, 7.56]	0.594
Unfavorable graft function recovery*	0.55 [0.36, 0.85]	0.007	0.72 [0.36, 1.44]	0.354	0.95 [0.61, 1.51]	0.842

Unfavorable graft function recovery indicated a stable eGFR < 45 mL/min/1.73m²; sIPTW, stabilized inverse probability treatment weighting; OW, overlap weighting.

The propensity score of inosine usage was calculated via donor age, donor sex, donor BMI, donor death category, donor diabetes history, donor hypertension history, donor death causes, donor terminal serum creatinine level, cold ischemia time, transplant year, recipient peak PRA, recipient age, recipient sex, recipient BMI, recipient transplant history, recipient diabetes history, recipient dialysis months, recipient pretransplant creatinine level, recipient induction therapy, and recipient initial calcineurin inhibitor treatment.

All the models were analyzed by setting the donor ID as the cluster variable.

*Logistic model with clustered-robust standard error was applied, effect size was presented as odds ratio and 95% confidence interval; #, linear model with clustered-robust standard error was applied, effect size was presented as β and its 95% confidence interval.

Discussion

The efficacy of inosine during kidney transplantation surgery was investigated in this study. These results demonstrated that inosine has no protective effect on graft recovery. Inosine, a product of adenosine produced by adenosine deaminase, is routinely used in kidney transplantation in some transplantation centers because of its potential protective effect against IRI. Previous studies have yielded conflicting results with respect to the effects of inosine on kidney ischemia injury and preservation [12–15]. Maggio et al. [14] demonstrated that inosine facilitated renal function recovery in a canine kidney autotransplantation model subjected to 30 min of normothermic ischemia and 24 h of cold storage. In this study, the final serum creatinine levels were comparable. Rothwell et al. [13] demonstrated that inosine protects against 90 min of in situ warm ischemia injury in canine kidneys; the serum creatinine levels were significantly lower in the inosine-treated group. Additionally, pathology revealed that the glomeruli and tubules were well preserved in the inosine-treated group. In contrast, Okiye's study [12] demonstrated that inosine did not increase renal function in dog or rat kidneys subjected to a 60-min normothermic ischemia model. Similarly, Fitzpatrick et al. [15] reported no effect of inosine on renal function in a dog kidney autotransplantation model subjected to 15 min of warm ischemia and 12/24 h of cold preservation with Eurocollins solution. While the precise effect of inosine remains a topic of debate, several studies have investigated the potential mechanism by which inosine may influence ischemia injury. Lewandowski et al. [5] demonstrated a protective role of inosine on glycolysis and that it improved the energy status in a rabbit ischemic myocardium model. Furthermore, inosine has been shown to exert a cytoprotective effect in HepG2 cells exposed to oxygen-glucose deprivation [4]. In a heart transplantation model, inosine improved myocardial and endothelial function in the early reperfusion stage, at least in part, by modulating the peroxynitrite-poly (ADP-ribose) polymerase pathway [10]. Additionally, inosine ameliorates neuroinflammation after stroke by reducing inflammasome activation [3]. The protective role of inosine depends on preserving the cellular adenosine triphosphate (ATP) content and accelerating the restoration of the cellular ATP pool during renal hypoxia-reoxygenation injury [28]. The selective A₃ adenosine receptor plays a pivotal role in the protective effect, and this effect is independent of adenosine deaminase [29].

However, the protective effect of inosine against oxidative stress in humans remains unknown. The current literature only reports the use of inosine in animal models, with no related clinical studies identified. Despite inosine being reported to have neurorestorative, anti-inflammatory, immunomodulatory and cardioprotective effects in basic research, it is clinically indicated for leukopenia, thrombocytopenia, hepatitis diseases, heart diseases and neurologic diseases. The rationale for the use of inosine during transplantation is intuitive on the basis of its potential protective role. In this study, our results did not support the intraoperative usage of inosine for improving graft recovery. The statistical power of the study may be a contributing factor to this outcome. Increasing the sample size may facilitate the attainment of more reliable results. Furthermore, a previous study on gene expression in transplanted lung tissues demonstrated a negative correlation between the level of inosine and the length of ICU stay and ventilation time in lung transplantation cases [30].

This study has several limitations. First, it is a retrospective study, which introduces the potential for selection bias. A total of 143 cases were lost to follow-up, which represents a potential source of bias. This is because all of the censored cases survived after transplantation, and recipients who died within a short follow-up duration were included in the analysis. This indicates that the loss to follow-up was nonrandomized. Second, the imbalance in the sample in this study weakens the robustness of the results. Only approximately ten percent of the cohort stopped using inosine during transplantation surgery. Third, the dosage used in this study was based on tradition, and there are no public sources for guidance. The dosage used in the preclinical animal model was 100 ∼ 200 mg/kg, whereas in this study, the dose used was 1000 mg/person. It is unknown whether a dose-response relationship exists.

In conclusion, the findings of this study indicate that the administration of a single injection of 1000 mg inosine does not confer a protective effect on the incidence of delayed graft function, estimated glomerular filtration rate recovery, or unfavorable graft recovery in kidney transplantation.

Funding Statement

This work was supported by the National Natural Science Foundation of China [82370756, 82170771] and the Science and Technology Planning Project of Guangzhou [2024B03J0172].

Author contributions

Hengjun Xiao and Ning Na conceived and designed this study. You Luo, Yang Zhang, Zuofu Tang, and Jinhua Zhang collected the data. You Luo and Yang Zhang analyzed the data. You Luo drafted the manuscript. Ning Na and Hengjun Xiao supervised and revised the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Statement of ethics

This study was approved by the Research Ethics Committee at the Third Affiliated Hospital of Sun Yat-sen University (IRB approval number: [2020]02-243-01). Following ethical review, access to the donor profile was permitted. The requirement for informed consent was waived by the Ethics Committee, as all the inpatients had provided explicit authorization for the use of their medical data for scientific research on admission. This research complied with the Declaration of Helsinki. The reported clinical and research activities are consistent with the principles of the Declaration of Istanbul, as outlined in the 'Declaration of Istanbul on Organ Trafficking and Transplant Tourism'.

Data availability statement

All the data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.