Induction Immunosuppression and Renal Outcomes in Adult Heart Transplantation

Abstract

Background:

This study explores the use of induction therapy in orthotopic heart transplantation as it relates to preoperative renal function and evaluates the impact of its utilization on post-transplant outcomes.

Methods:

We conducted a retrospective analysis using the United Network for Organ Sharing database from 2000 to 2018 evaluating the initiation of de novo dialysis after transplantation. We examined the relationship between induction immunosuppression and pre-transplant estimated glomerular filtration rate with post-transplant outcomes, accounting for inter-center variability through a mixed-effects logistic regression model.

Results:

In total, 16,201 patients were included with a median age of 57 y (interquartile range 47, 63); 26% were women (n = 4222) and 28% (n = 4552) had a history of diabetes mellitus. The median estimated glomerular filtration rate (eGFR) was 67.5 mL/min (interquartile range 53.1, 86.7); 51.2% (n = 3068) of the recipients with eGFR < 60 received induction therapy compared to 42.5% (n = 4336) within the eGFR ≥ 60 group (P < 0.001). Adjusted multivariable analysis found that induction therapy was associated with de novo dialysis (odds ratio 1.25, 95% confidence interval 1.10–1.43, P < 0.001), with the most significant effect on patients with eGFR ≥ 60. Although significant, there was a weak correlation between center-level induction utilization and mean eGFR (r = −0.2, P < 0.001).

Conclusion:

In this analysis, the use of induction immunosuppression in orthotopic heart transplantation varied widely between centers and did not correlate strongly with pre-transplant eGFR. In addition, its utilization did not mitigate the risk of renal replacement therapy after transplantation and in fact was associated with increased risk even after adjusting for confounders most notably in patients with eGFR ≥ 60.

Introduction

Worsening renal function in the setting of chronic heart failure (HF) is recognized as an independent predictor of poor outcomes.^1–3 It is estimated that approximately one-fourth of the outpatient population with chronic HF has chronic kidney disease,⁴ and that pre-existing renal dysfunction with a creatinine clearance <50 mL/min doubles the 30-d mortality risk after orthotopic heart transplantation (OHT).⁵

Although the International Society for Heart and Lung Transplantation instituted a guideline change in 2016 which lowered the threshold for relative contraindications for OHT from an estimated glomerular filtration rate (eGFR) of <40 to <30 mL/min/1.73 m²,⁶ decreased renal function continues to be a major determinant of worse long-term outcomes. Thus, transplant centers face an increasing proportion of high-risk renal profile candidates listed for OHT.

Moreover, renal function also declines after OHT, with a 12% mean decrease in measured GFR within the first year.⁷ As expected, patients with poor kidney function pre-OHT are more likely to show deteriorating renal function after heart transplantation.^8,9 Even though induction immunosuppression is recommended to limit the nephrotoxicity secondary to calcineurin inhibitors (CNIs), some studies have found an increased requirement for renal replacement therapy related to the use of induction immunosuppression,¹⁰ as well as a lack of renal benefit from CNI-minimization protocols.¹¹

The objective of this study is to examine the relationship between the use of induction therapy and pre-OHT renal function and the impact on postoperative outcomes while accounting for inter-center variability.

Methods

Data source

The United Network for Organ Sharing (UNOS) registry includes deidentified information on pre-operative, intra-operative, and post-operative variables from all donors, wait-listed candidates, and transplant recipients in the United States. The University of Pittsburgh Institutional Review Board approved this study.

Study population

Adult (≥18 y) patients undergoing OHT between 2010 and 2018 were included. We excluded patients undergoing repeat OHT or multi-visceral transplants, patients with a history of end-stage renal disease on dialysis, and centers with less than 50 cumulative transplants during the study period. Recipients receiving the interleukin 2 receptor antagonist basiliximab or polyclonal anti-thymocyte antibodies thymoglobulin as induction regimen drugs were included in the induction cohort.

Baseline characteristics

Baseline waitlist data included age, gender, body mass index, ethnicity, etiology of HF, pulmonary vascular resistance, diabetes mellitus, the 4-variable Modification of Diet in Renal Disease eGFR, ventricular assist device (VAD) support, mechanical ventilation, class I and II panel reactive antibodies, and a pre-existing 13-point risk score for acute rejection within 1 y of OHT.¹² Donor variables included age, left ventricle ejection fraction, mechanism of death, prior cardiac arrest, diabetes mellitus status, cocaine usage, and smoking history. Donor-recipient matching variables included human leukocyte antigen mismatch, ABO (blood group system) blood type match, gender match, cytomegalovirus match, and donor-to-recipient weight ratio.

Outcomes

In order to comparatively assess the relationship between induction therapy and renal function, we stratified the cohort based on eGFR (with a cutoff of 60 mL/min/1.73 m²) and the use of induction therapy in recipients. The primary outcome was new-onset requirement of renal replacement therapy (RRT) following OHT. The secondary outcomes included rates of operative mortality, acute rejection rate, and survival at 90 d, 1 y, and 5 y.

Statistical analysis

The data are displayed as percentage for categorical variables and as mean or median for continuous variables, with standard deviation or interquartile range (IQR), respectively. Pearson’s chi-squared and Kruskal–Wallis tests were used where appropriate. All statistical tests were 2-sided, and P < 0.05 was considered significant. For the evaluation of the center variability, we used the percentage of induction therapy and the mean eGFR at a center level. Multilevel mixed-effects logistic regression models were used to estimate the relationship between induction therapy and post-OHT RRT requirement accounting for center variability. The variance partition coefficient was estimated to evaluate the proportion of variation attributable to systematic differences between center characteristics in the context of a binary dependent variable. This contribution as a percentage was calculated as [X/X+3.29] × 100, where X is the variance estimate from the random effect component of the model. Kaplan–Meier estimates were used to evaluate the longitudinal impacts of the use of induction therapy on survival. Pearson’s correlation coefficient (R) was utilized to assess the correlation between pre-OHT risk scores and the use of induction therapy. All analyses were performed with version 15 Stata Statistical Software (StataCorp LP, College Station, TX).

Results

Baseline characteristics

A total of 16,201 OHT patients were included, with a median age of 57 y (IQR 47, 63). The population was comprised of 26% women (n = 4222). The most common HF diagnosis was dilated cardiomyopathy in 52% (n = 8425); 28% (n = 4552) of patients had a history of diabetes mellitus and the median waiting list time was 106 d (IQR 10, 294). Induction immunosuppression was used in 45.7% (n = 7404), with 64.4% (n = 4771) receiving basiliximab and 35.5% (n = 2633) receiving thymoglobulin. The median eGFR was 67.5 mL/min (IQR 53.1, 86.7), with 21.2% (n = 3560) of patients having a preserved renal function above 90 mL/min, and 1.9% (n = 308) having an eGFR < 30 mL/min (Fig. 1).

Fig. 1 –

OHT during the study period stratified by eGFR.

A history of diabetes mellitus was more frequently observed in the group of eGFR < 60, both with induction and non-induction therapy, compared to patients with eGFR ≥ 60 (34% versus 33.5% versus 24.3% versus 25.1%; P < 0.001) (Table 1). Patients receiving induction therapy with eGFR < 60 were less likely to have been bridged to OHT with VAD support (24.2% versus 27.5% versus 27.7% versus 28.8%; P < 0.001).

Table 1 –

Recipient characteristics.

	eGFR ≥ 60 (n = 10,210)			eGFR<60 (n = 5991)
	Non-induction (n = 5874)	Induction (n = 4336)	P-value	Non-induction (n = 2923)	Induction (n = 3068)	P-value
Age, median (IQR)	55 (44, 62)	54 (43, 61)	<0.01	61 (54, 66)	60 (53, 65)	<0.01
Female, % (n)	23.7% (1395)	25.9% (1125)	<0.01	27.9% (815)	28.9% (887)	0.38
BMI, median (IQR)	26 (23, 30)	26 (23, 30)	0.05	27 (24, 31)	27 (24, 31)	0.40
eGFR, median (IQR)	81 (70, 98)	79 (68, 95)	<0.001	49 (41, 54)	48 (40, 54)	<0.001
Rejection score, median (IQR)	8 (6, 9)	8 (6, 10)	<0.01	7 (6, 9)	7 (6, 9)	<0.01
Race/ethnicity			<0.01			<0.001
White	62.7% (3684)	63.5% (2754)		72.5% (2118)	71.3% (2188)
Black	23.1% (1356)	24.4% (1056)		15.3% (448)	19.2% (589)
Hispanic	9.3% (544)	7.7% (332)		7.8% (228)	5.7% (175)
Diagnosis			<0.001			<0.001
Non-ischemic dilated cardiomyopathy	54.3% (3187)	55.2% (2393)		47.1% (1376)	47.9% (1469)
Ischemic cardiomyopathy	33.5% (1966)	32.8% (1424)		41.0% (1197)	40.3% (1236)
Congenital	2.2% (130)	3.3% (144)		1.6% (48)	2.1% (65)
Valvular	1.2% (73)	1.3% (57)		1.4% (42)	1.5% (46)
Hypertrophic cardiomyopathy	2.3% (136)	2.9% (125)		2.0% (58)	2.3% (72)
Restrictive cardiomyopathy	3.1% (181)	2.4% (104)		3.5% (102)	4.2% (128)
Diabetes mellitus	25.1% (1476)	24.3% (1055)	0.36	33.5% (979)	34.0% (1042)	0.70
Intra-aortic balloon pump	6.1% (357)	7.4% (322)	<0.01	4.1% (121)	4.8% (146)	0.08
Inotropes at listing	31.9% (1872)	35.0% (1519)	<0.001	31.5% (920)	36.6% (1123)	<0.001
Serum bilirubin, median (IQR)	0.7 (0.5, 1)	0.7 (0.5, 1)	0.38	0.7 (0.5, 1.1)	0.7 (0.5, 1.1)	0.07
Pre-transplant mechanical ventilation	0.7% (40)	1.1% (45)	0.05	0.9% (25)	0.7% (23)	0.56
Transfusions while listed	21.2% (1244)	23.4% (1016)	0.007	23.2% (679)	22.9% (702)	0.75
College or more	53.3% (3133)	53.8% (2332)	0.66	56.0% (1637)	59.0% (1810)	0.01
PVR, median (IQR)	2.3 (1.5, 3.3)	2.3 (1.5, 3.3)		2.2 (1.5, 3.2)	2.2 (1.5, 3.3)	0.36
CF ventricular assist device	28.8% (1690)	27.7% (1203)	0.26	27.5% (803)	24.2% (743)	<0.01
Heartmate 2	22.4% (1315)	20.2% (878)	<0.01	22.5% (658)	18.4% (565)	<0.001
Heartmate 3	1.0% (61)	1.0% (44)	0.91	1.0% (29)	0.9% (27)	0.65
HeartWare	5.3% (314)	6.5% (281)	0.01	4.0% (116)	4.9% (151)	0.07
Class I PRA >20%	5.4% (320)	6.0% (258)	0.28	5.2% (153)	5.2% (161)	0.98
Class II PRA >20%	4.3% (253)	3.3% (143)	<0.01	4.0% (116)	4.0% (123)	0.94
Days on waiting list, median (IQR)	103 (30, 289)	109 (31, 296)	0.16	113 (32, 303)	101 (30, 290)	0.09

BMI = body mass index; PRA = panel reactive antibodies; PVR = pulmonary vascular resistance; CF = chronic failure.

The median donor age was 30 y (IQR 23, 41); 51.8% (n = 8407) of recipients were cytomegalovirus-matched, and the median allograft ischemic time was 3.1 h (IQR 2.3, 3.7). Donors for recipients with reduced renal function tended to be older (31 versus 30 y, P < 0.001), regardless of induction therapy utilization. In contrast, patients who received induction therapy had longer ischemic times (3.2 versus 3.0 h, P < 0.001) and were further from the transplant center (78 versus 89 miles, P < 0.001), irrespective of the renal function (Table 2).

Table 2 –

Donor characteristics.

	eGFR ≥ 60 (n = 10,210)			eGFR < 60 (n = 5991)
	Non-induction (n = 5874)	Induction (n = 4336)	P-value	Non-induction (n = 2923)	Induction (n = 3068)	P-value
Age, median (IQR)	30 (22, 40)	30 (22, 40)	0.22	32 (24, 42)	31 (23, 41)	<0.01
LV ejection fraction, %, median (IQR)	60 (55, 65)	60 (55, 65)	0.31	60 (56, 65)	60 (56, 65)	0.87
Mechanism of death			<0.01			<0.01
Anoxia	29.9% (1756)	26.6% (1152)		31.6% (923)	26.8% (822)
Stroke	18.7% (1097)	18.8% (817)		19.5% (570)	19.9% (610)
Head trauma	48.8% (2865)	51.4% (2229)		46.6% (1362)	50.5% (1548)
CNIs tumor	0.7% (42)	0.5% (23)		0.2% (6)	0.6% (16)
Other	1.9% (113)	2.6% (112)		2.1% (62)	2.3% (61)
Cardiac arrest	8.5% (497)	6.4% (278)	<0.001	7.9% (232)	6.2% (189)	<0.01
Diabetes mellitus	3.6% (209)	3.5% (152)	0.89	4.5% (132)	3.6% (111)	0.07
HLA mismatch (>3)	75.9% (4461)	80.9% (3506)	<0.001	75.6% (2211)	79.4% (2436)	<0.001
Gender match	75.2% (4415)	76.5% (3.316)	0.13	75.3% (2200)	75.5% (2316)	0.84
CMV match	52.0% (3053)	50.7% (2200)	0.22	52.8% (1542)	52.5% (1612)	0.87
Donor-recipient weight ratio, mean (SD)	1.02 (0.24)	1.02 (0.23)	0.16	1.01 (0.23)	1.01 (0.23)	0.91
Ischemic time, median (IQR)	3.0 (2.3, 3.7)	3.2 (2.4, 3.8)	<0.001	3.0 (2.3, 3.7)	3.2 (2.5, 3.9)	<0.001
Distance donor hosp to transplant center (nautical miles), median (IQR)	77 (11, 251)	86 (12, 275)	<0.01	78 (12, 251)	94 (13, 299)	<0.01

CMV = cytomegalovirus; LV = left ventricle; HLA = human leukocyte antigen; SD = standard deviation; CNIs = calcineurin inhibitors.

Renal function and induction therapy

Overall, 36.9% (n = 5991) of recipients had an eGFR < 60 mL/min; these patients were significantly older (median age 60 versus 54 y; P < 0.001) and demonstrated a relatively higher proportion of women in contrast to recipients with an eGFR ≥ 60 mL/min (28.4%, n = 1702 versus 24.7%, n = 2520; P < 0.001). Furthermore, 51.2% (n = 3068) of the recipients with eGFR < 60 received induction therapy compared to 42.5% (n = 4336) within the eGFR ≥ 60 group (P < 0.001). In both groups, basiliximab was the induction drug most commonly used (65.4% versus 63.8%; P = 0.15). Although there was a tendency toward a higher induction utilization in centers with recipients with lower mean eGFR (Fig. 2), center-level data showed a weak correlation between induction therapy utilization rates with both the mean eGFR (r = −0.21) and the percentage of patients with eGFR < 60 (r = 0.07).

Fig. 2 –

Scatter plot of induction therapy utilization and eGFR at a center level. Panel A shows correlation with mean eGFR. Panel B shows correlation with the percentage of eGFR ≤60 mL/min. Each circle represents an OHT center.

For maintenance immunosuppression, tacrolimus was used in 92.3% (n = 9428) and 90.8% (n = 5438) of patients in the eGFR ≥ 60 and eGFR < 60 cohorts, respectively (P < 0.001). Overall, no differences were observed in cyclosporine utilization between eGFR groups (4.3% versus 4.8%; P = 0.23). However, more patients received induction therapy and tacrolimus versus only tacrolimus as immunosuppression strategy (91.8% versus 89.7%; P = 0.005), as well as cyclosporine plus induction versus only cyclosporine (5.6% versus 3.9%; P = 0.002) within the eGFR < 60 group. Likewise, in the eGFR ≥ 60 group more patients (5.6%, n = 241) received induction therapy and cyclosporine in contrast to cyclosporine only (3.5%, n = 203) (P < 0.001). No difference was observed in tacrolimus utilization among eGFR ≥ 60 (Supplementary Table 1).

Post-orthotopic heart transplantation outcomes

In total, 9.8% (n = 1596) of patients required RRT post-OHT, which was more frequently observed in patients with reduced renal function: 6.2% (n = 222), 8.2% (n = 546), 12.7% (n = 723), and 34.1% (n = 105) with eGFR ≥ 90, 89–60, 59–30, and <30 mL/min, respectively (Supplementary Table 2). Likewise, there was a differential distribution in the RRT rates associated with the utilization of induction therapy in recipients with eGFR ≥ 60 (8.7% versus 6.7%; P < 0.001), which was not significant within the eGFR < 60 group (14.2% versus 13.4%; P = 0.33). Patients within the eGFR < 60 group had lower rates of treated acute rejection within 1-y post-OHT compared to patients with an eGFR ≥ 60 mL/min (14.5% versus 17.8%; P < 0.001).

Among patients with an eGFR < 60, there was an increased length of stay (16 d, IQR 11–24 versus 14 d, IQR 10–21; P < 0.001), increased 30-d mortality (4.7% versus 2.8%; P < 0.001), and increased 90-d mortality (7.1% versus 4.1%; P < 0.001) compared to patients with an eGFR > 60. Furthermore, patients receiving induction therapy had lower 30-d mortality (2.3% versus 3.2% in eGFR ≥ 60; 4.1% versus 5.3% in eGFR < 60) and lower 90-d mortality (4.0% versus 4.9% in eGFR ≥ 60; 6.4% versus 7.7% in eGFR < 60), in both eGFR groups (P < 0.001) (Table 3).

Table 3 –

Post-transplant outcomes by renal function and induction therapy.

Post-transplant outcomes	eGFR ≥ 60 (n = 10,210)			eGFR < 60 (n = 5991)
	Non-induction (n = 5874)	Induction (n = 4336)	P-value	Non-induction (n = 2923)	Induction (n = 3068)	P-value
Post-OHT dialysis, % (n)	6.7% (392)	8.7% (376)	<0.001	13.4% (391)	14.2% (437)	0.33
Post-OHT stroke, % (n)	2.6% (150)	2.7% (116)	0.70	2.7% (78)	3.0% (92)	0.44
Pacemaker, % (n)	3.1% (180)	3.0% (130)	0.85	2.8% (82)	3.4% (105)	0.20
LOS, median (IQR)	14 (10, 20)	15 (10, 22)	<0.001	15 (11, 24)	16 (11, 25)	0.01
1-y acute rejection requiring treatment, % (n)	17.1% (1006)	18.6% (808)	0.04	14.3% (417)	14.7% (450)	0.66
Acute rejection, % (n)	22.0% (1292)	20.8% (901)	0.14	18.3% (535)	19.0% (582)	0.51
30-d mortality, % (n)	3.2% (189)	2.3% (98)	0.004	5.3% (155)	4.1% (127)	0.03
90-d mortality, % (n)	4.9% (286)	4.0% (173)	0.03	7.7% (226)	6.4% (197)	0.04
1-y survival, % (n)	92.8% (5453)	92.9% (4027)	0.53	88.4% (2584)	90.0% (2762)	0.01
5-y survival, % (n)	85.2% (5007)	85.5% (3709)	0.40	81.1% (2371)	82.6% (2534)	0.06

LOS = length of stay.

Mixed-effects logistic regression for post-orthotopic heart transplantation renal replacement therapy

With each center as a random parameter, the standard deviation for the center effect for RRT was 0.53 (95% confidence interval [CI] 0.44–0.64). After adjusting for year of OHT and recipient risk factors as fixed parameters, the utilization of induction therapy was highly associated with RRT requirement in the post-OHT period with an odds ratio (OR) of 1.25 (95% CI 1.10–1.43; P < 0.001) in an average center. Other variables also related to our primary outcome included transplant year, recipient gender, eGFR, Hispanic ethnicity, congenital heart disease diagnosis, UNOS 1a status, pre-OHT left VAD, blood transfusion while listed, prior cardiac surgeries, and pre-OHT infections (Table 4). The variance partition coefficient was 7.8% suggesting there is low center variation in RRT rates attributable to other unmeasured center characteristics.

Table 4 –

Mixed-effects logistic regression model for renal replacement therapy post-OHT.

Variables	OR (95% CI)	P-value
Induction therapy	1.25 (1.10, 1.43)	<0.001
Transplant year	1.06 (1.03, 1.08)	<0.001
Age	1.01 (0.94, 1.16)	0.45
Female	0.85 (0.75, 0.96)	0.01
Heart failure diagnosis	Dilated as reference
Ischemic cardiomyopathy	1.03 (0.92, 1.16)	0.5
Congenital	2.82 (2.1, 3.8)	<0.001
Valvular	1.43 (0.93, 2.1)	0.9
Hypertrophied	1.60 (1.15, 2.21)	<0.01
Restrictive	1.35 (1.01, 1.81)	0.04
Ethnicity	White as reference
Black	1.08 (0.95, 1.23)	0.22
Hispanic	1.34 (1.09, 1.34)	<0.01
eGFR category	>90 mL/min/1.73 m2 as reference
60–89 mL/min	1.34 (1.13, 1.59)	<0.01
30–59 mL/min	2.24 (1.89, 2.67)	<0.001
<30 mL/min	8.53 (6.3, 11.42)	<0.001
Status 1A	1.15 (1.02, 1.30)	<0.01
Diabetes	1.08 (0.95, 1.22)	0.23
LVAD	1.14 (1.01, 1.29)	0.03
Transfusion while listed	1.38 (1.21, 1.56)	<0.001
Intra-aortic balloon pump while listed	1.20 (0.95, 1.53)	0.12
ECMO support while listed	0.59 (0.26, 1.31)	0.19
Mechanical ventilation while listed	1.28 (0.92, 2.02)	0.29
Prior cardiac surgery	1.44 (1.28, 1.62)	<0.001
Pre-OHT infection	1.32 (1.11, 1.56)	<0.01
Days in waiting list	1.00 (0.99, 1.01)	0.25
Random effect parameter	Standard deviation (95% CI)
Transplant center	0.53 (0.44, 0.64)	–

ECMO = extracorporeal membrane oxygenator.

When evaluating the induction therapy effect on the predicted probabilities of RRT within each renal function category, we observed that the most significant effect occurred in recipients with eGFR ≥ 60 mL/min in contrast to non-induction therapy. That is, the utilization of induction therapy was associated with a 0.29 (95% CI 0.01–0.58; P = 0.04) and 0.32 (95% CI 0.13–0.52; P = 0.001) increase in the predicted probabilities of RRT in patients with eGFR ≥ 90 mL/min and 60–90 mL/min compared to non-induction patients, respectively, but no significant effect was observed in patients with eGFR < 60 (Fig. 3).

Fig. 3 –

Marginal effect difference of induction therapy on post-OHT renal replacement therapy stratified by eGFR categories.

Survival estimates by estimated glomerular filtration rate and induction therapy

Overall, mortality rates at 1- and 5-y post-OHT were 8.49% (n = 1375) and 15.92% (n = 2580), respectively. Patients with reduced renal function had lower survival rates, with 5-y survival of 70% in patients transplanted having an eGFR < 30 (Supplemental Fig. 1). In terms of induction therapy and renal function, survival at 1 y was superior in patients who received induction therapy in contrast to those that did not among recipients with eGFR < 60 (88.1% versus 90.0%; P = 0.01). Kaplan–Meier survival analysis showed no significant impact of induction therapy on 5-y survival regardless of the renal function group (Fig. 4).

Fig. 4 –

Kaplan–Meier survival estimates stratified by eGFR and induction therapy utilization. *Group 1 eGFR ≥ 60 without induction; Group 2 eGFR ≥ 60 with induction; Group 3 eGFR < 60 without induction; and Group 4 eGFR < 60 with induction.

Discussion

The congregation of several risk factors in chronic HF patients makes renal dysfunction a prevalent factor and a major determinant in cardiac surgery outcomes. Furthermore, acute kidney injury is common during the early post-OHT period, between 25% and 76% in some series,^8,13 and increases the risk of mortality in OHT recipients.⁵ Aside from the hemodynamic consequences from low cardiac output states, immunosuppressive medications play an important role in the patho-physiology of renal dysfunction, motivating the use of an induction therapy modality to allow for early CNI withdrawal and the use of CNI-free immunosuppression strategies.^14,15 Moreover, highly sensitized recipients might benefit from induction therapy,¹⁶ although controversy still exists regarding the survival benefit of this immunosuppressive modality.¹⁷ In consequence, the balance between acute rejection risk and immunosuppression-related adverse events complicates decision making in clinical practice.

Study findings

The principal finding of this analysis was that induction therapy did not mitigate the risk of postoperative RRT in those with reduced baseline eGFR, and was associated with increased RRT risk in those with preserved eGFR. One explanation for these findings may be that although a benefit of induction therapy is the avoidance of nephrotoxic agents in the early postoperative period, some centers may continue to use these agents in the early period in addition to induction immunosuppression to minimize rejection risk. One theoretical example of this practice could be a young female with postpartum cardiomyopathy who is sensitized and undergoing OHT with compromised baseline renal function. Many centers would weigh the risks of further renal function decline versus the high risk of rejection in this patient and may err on the side of induction therapy with early initiation of CNIs as well. Although a higher proportion of recipients had received tacrolimus or cyclosporine as maintenance immunosuppression within the eGFR < 60 group, the timing, dosing, and levels of these medications are unfortunately not available in the UNOS registry.

Furthermore, since only a small proportion of OHT recipients had an eGFR < 30, with the majority of the population (63%) having an eGFR ≥ 60, these percentages likely reflect the common practice to consider patients with end-stage renal dysfunction for simultaneous heart-kidney transplantation rather than isolated heart transplantation. Likewise, corresponding to the International Society for Heart and Lung Transplantation guidelines, patients with reduced renal function were more commonly found to receive induction therapy in our analysis, where there was a tendency of higher induction utilization rates among centers with lower mean eGFR in the United States with relatively low variability (7%) in the requirement for RRT between centers attributable to other unmeasured variables. Although ideally this would suggest a discernible pattern of resource utilization in terms of induction immunosuppression across the country, we found instead this was a weak correlation, without a recognizable pattern of higher induction therapy utilization among centers with increased proportion of patients with an eGFR < 60.

Further work to define which patients benefit the most from induction therapy in the current era of OHT is prudent. This should incorporate all available desensitization strategies, evaluate adherence to CNI minimization protocols, and utilize pre-OHT prediction tools estimating rejection risk and renal failure risk in order to weigh the relative risks and benefits associated with induction therapy.

Prior studies

Multiple prior studies have identified predictors of postoperative renal dysfunction in cardiac surgery, including older age, female sex, valve replacement procedures, and longer cross clamp and cardiopulmonary bypass times.^9,18 Nevertheless, pre-operative renal dysfunction is the single most predictive factor for post-operative RRT. Boyle et al.¹⁹ reported an RRT incidence of 5.8% during the early post-operative period following OHT. Among the variables they found to be associated with RRT were pre-operative serum creatinine level, albumin level, insulin-dependent diabetes, and cardiopulmonary bypass time.

An analysis using UNOS data from 2000 to 2010 found an RRT incidence of 7.7%, in which congenital heart disease, reduced pre-operative creatinine clearance, mechanical ventilation prior to OHT, and prolonged ischemic time were the most highly associated factors for post-OHT renal dysfunction requiring RRT.²⁰ In contrast to our analysis, diabetes mellitus was not found to be associated with RRT and induction therapy was not assessed as a potential risk factor. Additionally, the incidence of RRT has increased by 3% over the past decade. This finding can be explained in part because patients undergoing OHT over the past 10 y tend to be older than previously (2000–2010: mean age 52.2 versus 2010–2018: mean age 53.9 y) and by the fact that left VADs have been used more frequently as a bridge-to-transplant approach (2000–2010: 23% versus 2010–2018: 27%), although the latter trend may be reversed following the allocation policy change in October 2018.²¹

These findings are also supported by the analysis from the Nationwide Inpatient Sample by Nadkarni et al.,²² in which they found a temporal trend of increased hospitalization secondary to acute kidney injury (9.7%–32.7%) as well as hospitalization requiring dialysis (1.6%–2.3%) post-OHT, from 2002 to 2013, attributable to changes in demographics and comorbidities. Furthermore, they reported a lower proportion of diabetes mellitus in patients requiring dialysis versus those without.

Amin et al., in an analysis using UNOS data from 2006 to 2015, compared basiliximab, thymoglobulin, and no induction and found that patients receiving induction immunosuppression more frequently required dialysis during their OHT hospitalization (8.8% versus 9.3% versus 6.6%; P < 0.001) and more frequently required chronic dialysis after OHT (3.9% versus 2.2% versus 2.4%; P < 0.001), respectively. In our analysis, we did not observe any differences in the requirement for RRT based on the use of basiliximab or thymoglobulin.

Limitations

This study is subject to biases inherent in a retrospective analysis. We relied upon the data fidelity of a multicenter registry, which is susceptible to missing and incorrect data entry. Additionally, the lack of granular information during the post-operative period related to adherence to maintenance immunosuppression; immunosuppressant levels, timing, and dosing; and complications affecting renal function could not be accounted for and represent potential confounders. Likewise, there are other patient factors that may weigh into the decision-making process regarding induction versus no induction therapy as well as institutional preferences that were not adjusted for in the analysis.

Conclusions

In this analysis of 16,201 OHT recipients, we demonstrate that the utilization of induction therapy is highly associated with RRT requirement post-OHT, particularly in patients with normal pre-operative renal function. Furthermore, there was a weak correlation between induction utilization and mean eGFR among transplant centers, supporting the notion that there is wide variability observed in induction therapy utilization nationwide. More studies are needed to identify patients who would derive the most benefit from induction therapy, incorporating both the risk of rejection and the risk of postoperative renal failure in OHT.