Abstract
Background:
The use of induction immunosuppression for heart transplantation (HT) is debated given the uncertain benefit and potential risks of infection and malignancy.
Methods:
This is a retrospective single-center analysis of 475 consecutive HT recipients from 2003 to 2020 grouped by use of induction with basiliximab group (BG) and the no basiliximab group (NBG). Subgroup analysis by era compared pre-2016 standard-basiliximab (BX) induction and 2016–2020 with selective-BX use as part of a calcineurin-inhibitor-sparing regimen.
Results:
When adjusted for confounders (sex, age, PRA, eGFR), the BG was less likely to have acute cellular rejection (ACR) (OR.42, p< .001), but had more antibody mediated rejection (AMR) (OR 11.7, p < .001) and more cardiac allograft vasculopathy (CAV) (OR 3.8, p= .04). There was no difference between BG and NBG in the incidence of malignancies or infections. When stratified by era (pre-2016 vs. 2016–2020), ACR remained less common in the BG than the NBG (36% vs. 50%, p= .045) groups, while AMR remained more common (9.7 vs. 0% p= .005). There was no significant difference in conditional survival comparing pre-and post-2016 NBG (HR 2.20 (95% CI.75–6.43); however, both pre-2016 BG and post-2016 BG have significantly higher mortality (HR 2.37 [95% CI 1.02–5.50) and HR 2.69 (95% CI 1.08–6.71), p = .045 and.03, respectively].
Conclusion:
Basiliximab reduces the incidence of ACR but increases the risk of AMR, CAV, and may be associated with increased mortality. Mechanistic studies are needed to describe a potential T-cell-escape mechanism with enhanced humoral immunity.
Keywords: allograft rejection, basiliximab, cardiac allograft vasculopathy, heart transplant, immunosuppression induction, infection, malignancy
1 |. INTRODUCTION
The benefit of induction immunosuppression for heart transplant (HT) recipients is debated, and its routine use is not recommended by the International Society for Heart and Lung Transplantation (ISHLT) except in patients at high risk for acute rejection or in those with significant renal impairment.1
Currently, induction therapy with either monoclonal interleukin-2 (IL-2)-receptor antagonist (basiliximab) or polyclonal antilymphocyte antibodies (ATG) is used in 40%–50% of HT.2 Basiliximab is the most commonly used induction agent, providing a more selective effect than ATG via inhibition of IL-2 mediated T-cell proliferation. While intended to reduce the risk of early acute rejection, the impacts of induction regimens on longer-term outcomes such as late or chronic rejection and overall survival are not well defined.3 Studies including randomized clinical trials, case-controlled trials, and systematic reviews have shown conflicting reports on short-term benefits with decreased rates of acute rejection contrasted with higher morbidity and unclear effect on long-term outcomes.2–4
Here, we report 18 years of patient outcomes from a large transplant center to define the risks and benefits of induction with the IL-2 receptor antagonist basiliximab when used as standardized induction for all recipients and as induction in select patients requiring a calcineurin inhibitor-sparing regimen.
2 |. METHODS
This is a single-center retrospective analysis of patients undergoing heart transplantation at the University of Pennsylvania from 2003 to 2020. We aimed to report long-term outcomes throughout an evolving area of transplant care in a cohort analysis without prospective design or specific outcomes. Data were collected from the electronic health record (EHR) via database queries and manual chart review, as well as from United Network for Organ Sharing (UNOS) data. Exclusion criteria were multiorgan transplantation, retransplantation, death during index transplant admission, and bridge to transplant with a left ventricular assist device (LVAD) as our institution does not use induction therapy in LVAD patients. Since the focus was on long-term outcomes of induction therapy, mortality data are presented based on conditional survival of the index transplant admission.
Patients were divided into two groups based on induction strategy with the IL-2 receptor antagonist basiliximab at time of HT: basiliximab group (BG) and no-basiliximab group (NBG). Over the course of the early 2000s basiliximab use was increasing in our institutions for single organ HT and had become standard of care by 2007. In 2016, however, our center stopped this routine use of basiliximab induction after an internal quality analysis showed increased infection, malignancy, and a trend towards higher mortality. Since 2016, basiliximab is only used as part of a calcineurin inhibitor-sparing protocol (CNI) in recipients with baseline renal impairment.
The CNI protocol is used if eGFR is < 45 mL/min or GFR 45–59 mL/min in the setting of additional risk factors. A subgroup analysis by era compared pre-2016 with standard basiliximab induction and 2016–2020 with selective basiliximab use.
Acute cellular rejection (ACR) was defined by the 2004 ISHLT criteria and antibody medicated rejection (AMR) based on 2013 ISHLT based on review of pathology diagnostic reports. AMR was defined as AMR ≥ 1+ on histopathology. DSA were defined as MFI ≥ 1500 as reported by the institutional HLA lab. All patients are routinely placed on statin therapy post-transplant for CAV prevention. Details on the institutional immunosuppression regimen including CNI target troughs, biopsy timing schedule and CAV surveillance schedule can be found in the supplemental material.
Bacterial infections were defined as any infection needing antibiotic treatment. Fungal infections were defined as cryptococcal, aspergillosis, candida or mucor infections. Data on CMV viremia were collected. Malignancies were defined as solid and liquid tumors other than post-transplant lymphoproliferative disorder (PTLD), melanoma, nonmelanoma skin cancer. The University of Pennsylvania Institutional Review Board approved this study (IRB protocol number 849008).
2.1 |. Statistical analysis
Continuous variables were expressed as mean ± standard deviation or median with interquartile ranges, and were compared using the unpaired t-test or Kruskal–Wallis test as appropriate. Categorical variables were compared using the chi-square test. Survival was assessed by Kaplan–Meier analyses and compared by log-rank tests based on conditional survival of the index admission. Multivariable Cox regression models were constructed adjusting for variables with significant and near-significant p-values on univariable analysis (p ≤ .1) and were performed for both the whole cohort and stratified by era. This allowed for a more parsimonious model in the setting of limited numbers of key outcomes, while still permitting adjustment for likely potential confounders. All statistical analyses were performed using Stata v.15.0 (StataCorp, LLC).
3 |. RESULTS
3.1. Patient characteristics
A total of 475 patients were included in the analysis after exclusion criteria were applied. There were n = 105 multiorgan transplants, n = 33 retransplants, and n = 82 LVAD bridge-to-transplant patients who excluded from analysis, along with n = 40 additional patients who met inclusion criteria but did not survive the index transplant admission. Out of the 475 patients, 284 (59.8%) received basiliximab induction and 191 (40.2%) did not (Figure 1).
FIGURE 1.
Study population and design. Initially, 1078 consecutive HT recipients were identified and after applying exclusion criteria, 475 were included in the analysis. Patients were grouped by no induction n = 284 (59.8%) and basiliximab induction n = 191 (40.2%). Overall, 7308 endomyocardial biopsy samples were included, 4205 in the basiliximab—and 3103 in the no induction group. A subgroup analysis was performed stratified by era for pre- and po 2016 basiliximab versus no induction groups following an institutional change away from standard use of basiliximab induction.
There was a modest, nonsignificant difference in sex composition of the subgroups, with 67.6% male sex in the BG and 75.4% male sex in the NBG (p = .07).
Patients in the BG were older (median 57 years, [IQR 49,62] vs. 54 [IQR 43, 60]; p= .004) and had a higher PRA (p= .01) when compared to NBG (Table 1). There was a small but significant difference in baseline renal function, with serum creatinine higher in the BG group (1.2 vs. 1.1, p = .002).
TABLE 1.
Study population.
| NBG n = 284 | BG n = 191 | p-value | |
|---|---|---|---|
| Age at transplant (years) | 54 (43,60) | 57 (49,62) | .004* |
| Male N (%) | 144 (75.4) | 192 67.6 | .07 |
| Race N (%) | |||
| White | 125 (70.7) | 208 (73.2) | |
| Black | 37 (19.4) | 53 (18.7) | |
| Other | 19 (10) | 23 (8.1) | .75 |
| Etiology N (%) | |||
| ICM | 54 (28.3) | 85 (28.3) | |
| NICM | 131 (68.6) | 192 (67.6) | |
| CHD | 3 (1.6) | 4 (1.4) | .9 |
| DM N (%) | 130 (68.1) | 187 (66.1) | .6 |
| Baseline creatinine (average) | 1.1 (.9,1.38) | 1.2 (.99,1.5) | .002* |
| eGFR (average) | 66.7 (33.6) | 61.3 (37.5) | .1 |
| PRA (%) | |||
| 0% | 165 (86.4) | 212 (74.7) | |
| 0%–19% | 7 (3.7) | 28 (9.9) | |
| 20%–49% | 9 (4.7) | 15 (5.3) | |
| >50% | 10 (5.2) | 29 (10.2) | .01* |
| Prior sternotomy (%) | 68 (35.6) | 69 (24.5) | .009 |
Abbreviations: BG, basiliximab group; CHD, congenital heart disease; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; ICM, ischemic cardiomyopathy; NBG, no basiliximab group; NICM, nonischemic cardiomyopathy; PRA, panel reactive antibodies; SCR, serum creatinine.
3.2 |. Rejection
Pathology results from a total of 7308 endomyocardial biopsy (EMB) samples were included in the analysis, 4205 in the BG and 3103 in the NBG. Quilty lesions were less commonly seen in EMB results from the BG compared to NBG (11.9%, vs. 18.1%, p < .001). ACR with ISHLT grades ≥ 2R was significantly more common in the NBG as compared to BG at 1-year post-transplant with 52.4% vs 38.4% (p = .001) and at 5-years post-transplant with 60.2% vs. 38.9% (p < .001) of patients, respectively, experiencing a “high-grade” cellular rejection. Most BG ACR appeared early, with all but two ACR events in the BG group being in the first year post-transplant.
AMR events with ISHLT grades ≥ pAMR1 were significantly more common in BG versus NBG, occurring in 7.6 versus.5% of patients, respectively, at 1-year post-transplant (p < .001) and in 20.4% versus 3.6% of patients at 5-years post-transplant (p < .001). There was no significant difference in patients hospitalized for rejection treatment (either AMR or ACR) between BG and NBG at 1- and 5-years post HT (p = .9 and p = .8, respectively).
When adjusted for significant and near-significant univariable confounders (sex, age, PRA, and GFR), the BG was still less likely to have ACR (OR.42, 95% CI.0,.28, p < .001) and still had more AMR during the first post-transplant year (OR 11.7, 95% CI 1.5, 89.6, p = .02). There remained no difference between groups in hospitalization for acute rejection. Adjusted results are summarized in Table 2.
TABLE 2.
Rejection, de-novo DSA and graft function outcomes for BG versus NBG and stratified by era.
| NBG n = 284 | BG n = 191 | p | MVA (gender, age, eGFR, PRA) | |
|---|---|---|---|---|
| ACR ≥ 2R, N (%) | 115 (60.2) | 108 (38.9) | <.001* | BX OR.42, 95% CI.0–.28, p < .001* |
| pAMR ≥ 1+, N (%) | 1 (.5) | 20 (7.6) | <.001* | BX OR 11.7, 95% CI 1.5–89.6, p.02* |
| Quilty Lesion, N (%)$ | 561 (18.1) | 499 (11.9) | <.001* | |
| Treatment, N (%) | 34 (18.2) | 47 (17.6) | .9 | |
| De novo DSA year 1, N (%) | 17 (9) | 39 (14.7) | .07 | |
| De novo DSA year 5, N (%) | 6 (5.2) | 32 (17.9) | .002* | BX OR 1.5, 95% CI.8–3, p.2 |
| CAV (ISHLT ≥ 2) year 5, N (%) | 3 (2.7) | 15 (9) | .036* | BX OR 3.8, 95% CI.06–13.7, p.04* |
| LVEF 1 year, average | 60 (58,65) | 65 (60,65) | .01* | |
| LVEF 5 year, average | 65 (60,70) | 65 (60,65) | .08 | |
| ACR ≥ 2R (%) | NBG pre-2016 | BG pre-2016 | p | |
| N (%) | 75 (67.6) | 69 (41.1) | <.001* | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 40 (50) | 39 (35.5) | .045* | |
| pAMR ≥ 1+ (%) | NBG pre-2016 | BG pre-2016 | ||
| N (%) | 1 (.9) | 10 (6.3) | .03* | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 0 (0) | 10 (9.7) | .005* | |
| De novo DSA | NBG pre-2016 | BG pre-2016 | ||
| N (%) | 0 (0) | 19 (11.7) | <.001* | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 17 (22.1) | 20 (19.4) | .7 | |
| LVEF | NBG pre-2016 | BG pre-2016 | ||
| Average | 60 (60,65) | 65 (60,65) | .02 | |
| NBG post-2016 | BG post-2016 | |||
| Average | 62.5 (57,65) | 65 (60,65) | .3 |
Abbbreviations: ACR, acute cellular rejection; AMR, antibody medicated rejection; BG, basiliximab group, BX-BG; CAV, cardiac allograft vasculopathy; DSA, donor specific antibodies; LVEF, left ventricular ejection fraction; NBG, no basiliximab group.
3.3 |. Donor specific antibodies (DSA)
There was no difference in the rate of de-novo DSA formation at 1-year post HT between BG and NBG (p = .07), while significant differences were observed at 5-years post-HT with 17.9% of patients in the BG developing DSA versus 5.2% in the NBG (p = .002). This significant difference in DSA was not seen between BG and NBG in the adjusted analysis (Table 2).
3.4 |. Cardiac allograft vasculopathy (CAV)
Aggressive CAV, defined as ≥ ISHLT grade 2 at 5-years post-HT, was significantly more common in the BG, with 9% of patients developing CAV versus 2.7% in the NBG (p = .036). When adjusted for confounders, the increased risk of aggressive CAV in the BG versus NBG persisted (OR 3.8, 95% CI.06, 13.7, p = .04) (Table 2).
3.5 |. Left ventricular ejection fraction
There was no clinically relevant difference in the median left ventricular ejection fraction (LVEF) at 1- and 5-year post-HT between BG and NBG (Table 2).
3.6 |. Infection
There were no significant differences in rates of bacterial infections, fungal infections, or CMV viremia at 1-year post-HT between the BG and NBG (Table 3).
TABLE 3 |.
Infectious and malignancy complications for BG versus NBG and stratified by era.
| NBG | BG | p | MVA (gender, age, eGFR, PRA) | |
|---|---|---|---|---|
| Bacterial infection N (%) | 53 (32.2) | 99 (38.2) | .2 | BX OR 1.2 (95% CI.8–1.8) p.4. |
| Fungal infection N (%) | 11 (5.8) | 22 (8) | .4 | BX OR 1.24 (95% CI.6–2.7) p.6 |
| CMV viremia N (%) | 28 (15.6) | 55 (20.6) | .4 | BX OR for 1.34 (95% CI.8–2.7) p.3 |
| Malignancy 1 year N (%) | 19 (10.2) | 27 (9.9) | .9 | BX OR.9 (95%CI.5–1.8) p.8 |
| Malignancy 5 year N (%) | 14 (18.4) | 18 (16.2) | .7 | |
| Bacterial infection | NBG pre-2016 | BG pre-2016 | p | |
| N (%) | 37 (34.3) | 66 (39.5) | .4 | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 16 (28.6) | 33 (35.9) | .4 | |
| Fungal infection | NBG pre-2016 | BG pre-2016 | ||
| N (%) | 4 (3.6) | 10 (6) | .4 | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 7 (6.9) | 12 (11.1) | .6 | |
| CMV viremia | NBG pre-2016 | BG pre-2016 | ||
| N (%) | 19 (17.9) | 33 (19.8) | .7 | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 9 (12.6) | 22 (22) | .09 | |
| Malignancy | NBG pre-2016 | BG pre-2016 | ||
| N (%) | 17 (15.6) | 24 (14.4) | .8 | |
| NBG post-2016 | BG post-2016 | |||
| N (%) | 2 (2.6) | 3 (2.8) | .9 |
Abbreviations: BG, basiliximab group, BX-BG; CMV, cytomegaly virus; NBG, no basiliximab group.
Amongst patients who were excluded from the analysis for mortality during the index admission there was no significant difference in cause of death from an infectious etiology with 44.4% in the BG and 34.8% in the NBG (p.46).
3.7 |. Malignancy
There was no difference in malignancies between BG and NBG at 1-year post HT and 5-years (Table 3).
3.8 |. Maintenance immunosuppression regimen
Maintenance immunosuppression differed between BG and NBG at 1-year post HT (p = .017). BG and NBG had a similar rate of triple immunosuppression with CNI, mycophenolate mofetil (MMF) and corticosteroid with 49.6 and 44.4%, respectively (p = .28). Rates of two-drug regimens were also similar between groups, though there were differences in the type of two-drug regimen, specifically with the combination of CNI plus MMF being more common in the NBG than in BG (40.6% vs. 27.4%, p = .004). The combination of CNI plus corticosteroid without MMF was more common in the BG than in NBG, but not significantly so (Table S1).
3.9 |. Subgroup analysis: Stratified by era for pre-2016 and 2016–2020 for 1-year outcomes
3.9.1 |. Rejection, DSA, and CAV
Regardless of era, ACR remained less common in BG versus NBG (pre-2016: 41.1% vs, 67.6%, p < .001; and post-2016: 35.5% vs. 50%, p = .045). AMR was also more common in BG versus NBG for both eras: AMR rates between groups pre-2016 were 6.25% versus.9% (p = .03), and AMR rates between groups post-2016 were 9.7% versus 0% (p = .005). Hospitalization and treatment for acute rejection were not significantly different in either era (Table 2). Denovo DSA were significantly more common in the pre-2016 era in the BG versus NBG (11.7% vs. 0%, p < .001), but were not significantly different in the post-2016 era (19.4% vs. 22%, p = .7) (Table 2).
3.9.2 |. Left ventricular ejection fraction
There was no clinically significant difference in LVEF between the groups (Table 2).
3.9.3 |. Infections
There were no significant differences in bacterial and fungal infections as well as CMV viremia when stratified by era (Table 3).
3.9.4 |. Malignancy
There was no difference in malignancies between the groups for any of the compared time frames (Table 3).
3.9.5 |. Maintenance immunosuppression
In the pre-2016 era, maintenance immunosuppression at 1-year post HT was significantly different between BG and NBG (Table S1).
3.9.6 |. Renal function
Renal function assessed by eGFR (using the Modification of Diet in Renal Disease (MDRD) equation5) was not different at baseline or 1-year post HT between the groups including the percentage change from baseline (Table S2). In the post 2016-CNI sparing protocol, the BG had a significantly lower eGFR of 60.5 mL/min/1.73 m2 versus post-2016 NBG with 80.9 mL/min/1.73 m2 (p = .02), which was not seen at 1-year post-HT with CNI sparing (BG 47.4 mL/min/1.73 m2 vs. post-2016 NBG 47.5 mL/min/1.73 m2, p = .99). However, the BG had significantly less decline in eGFR than the NBG with −17.9% versus post-2016 NBG −36.7% (p = .016). When comparing the CNI sparing versus all NBG groups, there was no difference in baseline eGFR, eGFR at 1-year, and percentage change from baseline to 1-year post-HT (Table S2).
3.9.7 |. Patient survival
Overall mortality based on conditional survival from the index admission at 1- and 5-years post-HT was higher in the BG group with 7% in the BG and 2.6% in the NBG and 19.7% in the BG versus 11.5% in the NBG, respectively. Cox survival analysis showed significantly higher unadjusted mortality in the BG with HR 1.9 (95% CI 1.0, 3.5, p.037). However, the adjusted analysis did not show a significant difference in survival between BG and NBG (HR 1.8, 95% CI.97, 3.4, p.06) (Figure 2, Table 4).
FIGURE 2.
1-year mortality basiliximab versus no induction. Crude survival was higher in the NBG when compared to BG. However, the adjusted analysis did not show a significant difference in survival between BG and NBG with HR of 1.8 (95% CI.97–3.4) p.06.
TABLE 4.
Crude patient survival and adjusted/unadjusted models.
| NBG | BG | p | |
|---|---|---|---|
| 1- year mortality N (%) | 5 (2.7) | 20 (7) | .034* |
| 5 - year mortality N (%) | 14 (11.5) | 40 (19.7) | .05 |
| 1-year mortality | NBG pre-2016 | BG pre-2016 | |
| N (%) | 1 (.9) | 9 (5.3) | .05 |
| NBG post-2016 | BG post-2016 | ||
| N (%) | 4 (5) | 11 (9.6) | .24 |
| BG versus NBG | Unadjusted Cox analysis | ||
| 1.9 (95% CI 1.0–3.5) | p.034* | ||
| BG versus NBG | Adjusted Cox analysis$(age, gender, eGFR, PRA) | ||
| HR 1.8 (95% CI.97–3.4) | p.06 | ||
| Adjusted Cox Analysis (era, age, gender, eGFR, PRA) | |||
| NBG / pre-2016 | Reference | ||
| NBG / post-2016 | HR 2.20 (95% CI.75–6.43) | .150 | |
| BG / pre-2016 | HR 2.37 (95% CI 1.02–5.50) | .045* | |
| BG / post-2016 | HR 2.69 (95% CI 1.08–6.71) | .034* |
Abbreviations: BG, basiliximab group; eGFR, estimated glomerular filtration rate; NBG, no basiliximab group; PRA, panel reactive antibodies.
Stratified by era there was a trend towards higher 1-year mortality in the pre-2016 BG with 5.3% as compared to the NBG with.9% (p = .05). Comparing post-2016 BG and NBG, there was no significant difference with 9.6% versus 5% 1-year mortality, respectively (p = .24). In an analysis adjusted for era, age, sex, eGFR, PRA and when choosing the pre-2016 NBG as a reference group, there was no significant difference in survival comparing pre-and post-2016 NBG group (HR 2.2, 95% CI.8–6.4, p = .15); however, both pre-2016 BG and post-2016 BG had significantly higher mortality when compared to the pre-2016 NBG (HR 3.4, 95% CI 1.0, 5.5, p = .045 and HR 2.69, 95% CI, 1.1–6.7, p = .034) (Figure 3, Table 4).
FIGURE 3 |.
Survival stratified by era and induction. No significant difference in survival comparing pre-and post-2016 NBG groups was found when choosing the pre-2016 NBG as a reference group. Both pre-2016 BG and post-2016 BG had significantly higher mortality when compared to the pre-2016 NBG (p = .045 and.03).
4 |. DISCUSSION
This report represents one of the largest and longest-followed observational cohorts of HT recipients receiving the IL-2 inhibitor basiliximab. Our main findings are as follows: (1) Adjusting for confounders (sex, age, PRA, eGFR) and stratified by era, the BG was less likely to have ACR but had more AMR at 1-year post HT and more CAV at 5-years post HT; (2) basiliximab had no effect on graft function, malignancies, and infections; (3) both pre-2016 BG and post-2016 BG have higher mortality based on conditional survival when compared to the pre-2016 NBG, but no difference was found between post-2016 NBG and the BG from either era.
The goals of using induction are to achieve early, high-intensity immunosuppression in patients deemed high risk for rejection, to allow for the delayed introduction of CNI in patients with renal insufficiency, and to facilitate earlier corticosteroid weaning. However, the use of induction strategies for HT is debated and long-term effects ill defined.3 A 2005 investigation suggested a survival benefit with the use of induction in patients at high risk for rejection, though the use of induction in patients at lower risk led to increased mortality.6 A subsequent 2013 Cochrane review showed no significant differences in mortality or CAV.4 And more recently, a single-center study showed lower ACR rates with basiliximab but no difference in survival when compared to no induction.7 ISHLT registry studies found slightly worse early outcomes with basiliximab compared to ATG and no-induction regimens, with higher risk of graft failure-related deaths when compared to no induction, and higher risk of early rejection events when compared to ATG.3 There was also a signal for higher long-term mortality in the basiliximab group related to graft failure when compared to ATG.8 Finally, a 2015 UNOS registry analysis showed no difference in survival for ATG versus basiliximab, and no difference between either method when compared to no induction.9 This led to the conclusion that survival might be more dependent on individual risk rather than on use or type of induction regimens, with apparent risk trade-offs resulting in a net-neutral effect on long-term outcomes. However, a more recent UNOS analysis (2017–2000) again raised concern for increased mortality with basiliximab use when compared to no induction without an effect on rejection.10
We studied long term effects on basiliximab induction and thus chose to base our analysis on conditional survival of the index transplant admission. While not significant, we observed a trend towards higher all-cause mortality in the adjusted comparison between BG and NBG (p.06). Notably, in our subgroup analysis stratified by era, both pre-2016 BG and post-2016 BG have significantly higher mortality when compared to the pre-2016 NBG. This finding was not replicated when comparing post-2016 NBG to the BG groups. It is notable that in our data, patient mortality in the pre-2016 NBG is quite low: ∼1% at 1-year and ∼6% at 5-years. Data collection was performed using automated EHR data queries, UNOS data, and manual chart review. However, there may be errors in data from the early portion of this dataset when EHR systems were being implemented. Overall, based on the mixed statistical results in our data, and based on the historical body of literature with conflicting findings, we find it difficult to reach confident conclusions about the effects of basiliximab on all-cause mortality.
We observed a significant reduction in ACR ≥ 2R for BG patients both overall and when stratified by era. This finding is consistent with results from both randomized controlled trials using IL-2 receptor antagonist11 and several aforementioned observational studies.6,7 A 2018 meta-analysis also showed lower rates of moderate-severe rejection with an induction regimen, though in this analysis it was ATG that conferred lower ACR risk while basiliximab did not.2 Findings regarding allograft rejection are not unanimous; however, with a large, aforementioned UNOS review finding no clear differences with induction.9 Overall, the historical data on the benefit of induction on rejection outcomes is not conclusive, with differing study designs, small patient numbers, and varying drugs creating challenges in interpretation. We believe our findings contribute to a preponderance of evidence suggesting that at least for ACR, induction-based regimens may have a modest benefit.
In light of the above conclusion, perhaps the most unexpected findings of our study are the increased rates of AMR and CAV with basiliximab use. While the higher rate of AMR and CAV ISHLT ≥ 2 in the BG was not associated with differences in graft function at 5-years as assessed by LVEF, one wonders whether these underlying, later-occurring manifestations of alloimmunity (as compared to ACR) may have contributed to the signals for decreased survival among BG patients. This finding may also explain the several studies suggesting long-term survival benefit of ATG over basiliximab, as ATG has been shown to decrease coronary plaque progression as interrogated by IVUS.12 The underlying mechanism of how basiliximab reduces ACR rates in exchange for increased humoral alloimmunity in the form of DSA, AMR, and CAV will require a deeper exploration of tissue-level biology. Could there be an immune escape mechanism with enhanced early humoral immunity and late complications? Basiliximab IL2 antagonism impacts not only effector T-cells and cytotoxic T-cells. Anti-inflammatory, tolerogenic regulatory T-cells are also impacted by basiliximab, and these cells have a purported role in suppressing chronic humoral immune responses.13 Thus, it may be that the lack of specificity in IL2 antagonism reduces short-term T-cell mediated rejection, but at the expense of developing a stable population of regulatory T-cells which help mitigate chronic, humoral immune activation. This type of unanticipated consequence from using an overly blunt immunosuppressive tool is not unprecedented. A similar phenomenon was recently observed using pan-B-cell depleting rituximab induction therapy, which unexpectedly led to accelerated CAV development in longer-term follow-up.15
Maintenance immunosuppression at 1-year post HT was significantly different between BG and NBG. Dual therapy with MMF and CNI was significantly more common in the NBG while triple therapy with MMF, CNI and corticosteroid was more common in the BG, which might have affected the rejection outcome. Notably, when basiliximab usage changed in 2016 with induction only occurring as part of a CNI-sparing protocol, there were no differences in maintenance immunosuppression at 1-year post HT. It is conceivable that the overall lower usage of MMF-based regimens in BG patients may have affected rejection outcomes. However, given that the rates of ACR and AMR for BG patients relative to NBG patients were essentially unchanged in the post-2016 era when rates of MMF usage were quite similar (84.2% vs. 80.5%), maintenance immunosuppression regimen alone does not appear to completely explain our observed rejection differences. We did find a notable change in the de-novo DSA outcome between eras, with de-novo DSA more common in the pre-2016 BG when compared to the NBG. However, the very low percentage recorded in the pre-2016 NBG raises concern for under detection, and there were no significant differences in the adjusted analysis of DSA incidence between BG and NBG across the full study cohort. Thus, this outcome also does not appear to be clearly driven by differences in the pre-2016 maintenance immunosuppression alone, even though it may be a contributing factor. Overall, our assessment is that at least some of the observed differences in de-novo DSA between eras is a result of changes in testing frequency, in the definition of positive DSA, and in the technical aspects of DSA analysis.
Our analysis did not reveal a significant difference between BG and NBG for infectious or malignancy outcomes. Similar results on infection-related outcomes have been shown in contemporary analyses.2,4,19 A 2020 ISHLT registry analysis had shown more infection-related deaths with induction, but this was not confirmed following adjustment for confounders.3 There is more conflicting data on malignancies with induction therapies. Some studies have reported no increase in malignancy rates with ATG,2,4,3 though to our knowledge, no prior studies have reported increased malignancies with basiliximab induction.
Previously in a subgroup analysis, patients with renal impairment did not have a mortality reduction with induction.3 In our cohort, baseline serum creatinine was statistically, but not clinically different, whereas no difference was detected in eGFR between BG and NBG at baseline and during post-HT year 1. In the subgroup analysis of post-2016 CNI sparing protocol, the BG had a significantly lower baseline eGFR but not at year-1 post HT, given a significantly less decrease in eGFR from baseline. Notably, while the baseline GFR was significantly lower at time of transplant in the post-2016 BG the absolute value was 60.5 mL/min. This relatively high GFR likely reflects some degree of hemodynamic stabilization and GDMT optimization during the wait-list time and is also reflective of a broad inclusion of patients in the CNI sparing pathway (see supplemental material). However, the lesser degree of GFR reduction over the first post-transplant year in the BG supports the renal protective effect of a CNI sparing protocol, though this did not affect long-term survival.
In summary, the most notable finding of this report is that induction with basiliximab decreased ACR rates during post-HT year 1, potentially increased the rates of later-occurring AMR and downstream CAV, and offered no survival benefit over no induction. As we move towards more personalized management approaches, future studies will need to better prospectively identify high-ACR risk patients who might benefit from induction while avoiding generalized usage that may predispose low-ACR risk patients to harm. Conventional clinical risk factors may be insufficient to achieve the degree of personalized risk stratification that clinicians need to make these choices. Continued investment in improving the speed and depth of HLA sequencing along with advances in tissue-level phenotyping to identify high-risk patients and organ matches will likely be crucial to achieving this type of personalized medicine solution.
The nonregistry analysis presented in this manuscript permitted assessment of a large amount of highly granular, directly sourced, patient-level data. However, while our cohort is large, well-phenotyped, and was followed over many years, there are limitations to this single-center retrospective cohort that should be considered when interpreting our findings. We cannot rule out unmeasured confounders or issues affecting the accuracy of primary source data entry. As discussed, the 5-year and pre-2016 NBG mortality was unexpectedly low, raising the possibility of inaccurate primary records in the older portion of the dataset that could not be addressed despite our efforts to use multiple data collection approaches. Despite our best efforts to clarify the status of patients lost to follow up, we cannot rule out that the 5-year mortality is underestimated. Furthermore, overall rates of DSA and AMR were low, likely arising from protocols in the earlier years of our cohort which were less vigilant about surveillance DSA testing and immunofluorescent staining. Despite these potential limitations, it is important to note that these issues in mortality data and AMR detection should have been distributed largely at-random across the cohort. Thus, while the absolute incidence of these events may not be fully accurate, the difference between the BG and NBG groups may still be an accurate reflection of the relative risks of induction versus no induction. It is also important to note that our study does not sufficiently reflect the potential impact of changes of the U.S. organ allocation system in 2018 and the benefit and risks of induction strategies in these patients, especially those bridged to transplant with temporary mechanical circulatory support. This merits investigation in future studies.
5 |. CONCLUSION
Induction with the IL-2 receptor antagonist basiliximab in HT recipients reduced the incidence of ACR but increased rates of AMR and CAV and did not offer any survival advantage, calling into question the overall benefit of induction except in select cases. Further clinical studies are needed to identify those patients who might particularly benefit from the early-ACR reducing effects of basiliximab, and further mechanistic studies are needed to understand the mechanism of enhanced humoral activation with IL2-receptor antagonism.
Supplementary Material
ACKNOWLEDGMENTS
Nosheen Reza is supported by the National, Heart, Lung and Blood Institute of the National Institute of Health under Award Number K23HL166961. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
- 1.Velleca A, Shullo MA, Dhital K, et al. The International Society for Heart and Lung Transplantation (ISHLT) guidelines for the care of heart transplant recipients. J Heart Lung Transplant. 2023;42(5):e1–e141. doi: 10.1016/j.healun.2022.10.015. published Online First: 20221220. [DOI] [PubMed] [Google Scholar]
- 2.Briasoulis A, Inampudi C, Pala M, Asleh R, Alvarez P, Bhama J. Induction immunosuppressive therapy in cardiac transplantation: a systematic review and meta-analysis. Heart Fail Rev. 2018;23(5):641649. doi: 10.1007/s10741-018-9691-2. published Online First: 2018/03/14. [DOI] [PubMed] [Google Scholar]
- 3.Nozohoor S, Stehlik J, Lund LH, Ansari D, Andersson B, Nilsson J. Induction immunosuppression strategies and long-term outcomes after heart transplantation. Clin Transplant. 2020;34(7):e13871. doi: 10.1111/ctr.13871. published Online First: 2020/04/10. [DOI] [PubMed] [Google Scholar]
- 4.Penninga L, Moller CH, Gustafsson F, Gluud C, Steinbruchel DA. Immunosuppressive T-cell antibody induction for heart transplant recipients. Cochrane Database Syst Rev. 2013;(12):CD008842. doi: 10.1002/14651858.CD008842. pub2 [published Online First: 2013/12/04. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Levey AS, Coresh J, Greene T, et al. Expressing the modification of diet in renal disease study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem. 2007;53(4):766–772. doi: 10.1373/clinchem.2006.077180. published Online First: 20070301. [DOI] [PubMed] [Google Scholar]
- 6.Higgins R, Kirklin JK, Brown RN, et al. To induce or not to induce: do patients at greatest risk for fatal rejection benefit from cytolytic induction therapy? J Heart Lung Transplant. 2005;24(4):392–400. doi: 10.1016/j.healun.2004.01.002. published Online First: 2005/03/31. [DOI] [PubMed] [Google Scholar]
- 7.Rudzik KN, Rivosecchi RM, Palmer BA, et al. Basiliximab induction versus no induction in adult heart transplantation. Clin Transplant. 2023;37(5):e14937. doi: 10.1111/ctr.14937. published Online First: . [DOI] [PubMed] [Google Scholar]
- 8.Ansari D, Lund LH, Stehlik J, et al. Induction with anti-thymocyte globulin in heart transplantation is associated with better longterm survival compared with basiliximab. J Heart Lung Transplant. 2015;34(10):1283–1291. doi: 10.1016/j.healun.2015.04.001. published Online First: 2015/06/20. [DOI] [PubMed] [Google Scholar]
- 9.Whitson BA, Kilic A, Lehman A, et al. Impact of induction immunosuppression on survival in heart transplant recipients: a contemporary analysis of agents. Clin Transplant. 2015;29(1):9–17. doi: 10.1111/ctr. 12469. published Online First: 2014/10/07. [DOI] [PubMed] [Google Scholar]
- 10.Bellumkonda L, Oikonomou EK, Hsueh C, Maulion C, Testani J, Patel J. The impact of induction therapy on mortality and treated rejection in cardiac transplantation: a retrospective study. J Heart Lung Transplant. 2022;41(4):482–491. doi: 10.1016/j.healun.2022.01.008. published Online First: . [DOI] [PubMed] [Google Scholar]
- 11.Beniaminovitz A, Itescu S, Lietz K, et al. Prevention of rejection in cardiac transplantation by blockade of the interleukin-2 receptor with a monoclonal antibody. N Engl J Med. 2000;342(9):613–619. doi: 10.1056/NEJM200003023420902. published Online First: 2000/03/04. [DOI] [PubMed] [Google Scholar]
- 12.Azarbal B, Cheng R, Vanichsarn C, et al. Induction therapy with antithymocyte globulin in patients undergoing cardiac transplantation is associated with decreased coronary plaque progression as assessed by intravascular ultrasound. Circ Heart Fail. 2016;9(1):e002252. doi: 10.1161/CIRCHEARTFAILURE.115.002252. published Online First: 2016/01/10. [DOI] [PubMed] [Google Scholar]
- 13.Pilat N, Farkas AM, Mahr B, et al. T-regulatory cell treatment prevents chronic rejection of heart allografts in a murine mixed chimerism model. J Heart Lung Transplant. 2014;33(4):429–437. doi: 10.1016/j. healun.2013.11.004. published Online First: 2014/01/29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lopez-Abente J, Martinez-Bonet M, Bernaldo-de-Quiros E, et al. Basiliximab impairs regulatory T cell (TREG) function and could affect the short-term graft acceptance in children with heart transplantation. Sci Rep. 2021;11(1):827. doi: 10.1038/s41598-020-80567-9. published Online First: 2021/01/14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Starling RC, Armstrong B, Bridges ND, et al. Accelerated allograft vasculopathy with rituximab after cardiac transplantation. J Am Coll Cardiol. 2019;74(1):36–51. doi: 10.1016/j.jacc.2019.04.056. published Online First: 2019/07/06. [DOI] [PubMed] [Google Scholar]
- 16.Newell KA, Asare A, Sanz I, et al. l. Longitudinal studies of a B cell-derived signature of tolerance in renal transplant recipients. Am J Transplant. 2015;15(11):2908–2920. doi: 10.1111/ajt.13480. published Online First: 2015/10/16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bigot J, Pilon C, Matignon M, et al. Transcriptomic signature of the CD24hiCD38hi transitional B cells associated with an immunoregulatory phenotype in renal transplant recipients. Am J Transplant. 2016;16(12):3430–3442. doi: 10.1111/ajt.13904 [DOI] [PubMed] [Google Scholar]
- 18.Trouvin A-P, Jacquot S, Grigioni S, et al. Usefulness of monitoring of B cell depletion in rituximab-treated rheumatoid arthritis patients in order to predict clinical relapse: a prospective observational study.Clin Exp Immunol. 2015;180(1):11–18. doi: 10.1111/cei.12481 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mehra MR, Zucker MJ, Wagoner L, et al. A multicenter, prospective, randomized, double-blind trial of basiliximab in heart transplantation. J Heart Lung Transplant. 2005;24(9):1297–1304. doi: 10.1016/j.healun.2004.09.010. published Online First: 2005/09/07. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.