Between 2005 and 2018, 1100 LTx were undertaken, with institutional recipient management and donor selection protocols described previously (15, 16). All recipients received standard triple immunosuppression with tacrolimus or ciclosporin (pre-2008), azathioprine or mycophenolate mofetil and corticosteroids. All individuals prescribed EVE were considered for inclusion. Excluded were those: lost to follow up, early discontinuation (duration of therapy <3 months) or previous sirolimus therapy.

Recipients on EVE were matched 1:1 with those who remained on CNI-based immunosuppression based on: transplant date, procedure, age at transplant, sex, and underlying diagnosis. After matching, survival and time to CLAD outcomes were calculated from the date of EVE commencement. If a LTR who remained on CNI-based immunosuppression did not survive to the date of starting EVE, they were excluded from further analysis. The final cohort consisted of 182 LTR (91 EVE recipients with 91 CNI controls).

Acute cellular rejection was defined as changes on transbronchial biopsy of ≥ International Society for Heart and Lung Transplantation (ISHLT) Grade 2, or in the absence of a biopsy an otherwise unexplained drop in lung function treated with intravenous corticosteroid (17, 18). Acute antibody-mediated rejection was diagnosed and managed according to Alfred Hospital protocols (19).

All LTR living within 300 km of our centre underwent indefinite long-term follow-up with regular spirometry. Spirometry at time of starting on EVE, 1-year following change and at time of diagnosis of CLAD was investigated for its impact on survival and time to CLAD.

CLAD was defined as an irreversible decline in forced expiratory volume in 1 s (FEV₁) to <80% of baseline (the mean of the two best post-LTx measurements, obtained at least 3 weeks apart with or without a decline in forced vital capacity [FVC]) (20). The phenotypes of CLAD were defined as either BOS (FEV₁/FVC <0.7 and FVC ≥80% predicted baseline FVC at CLAD onset) or RAS (FEV₁/FVC ≥0.7 and FVC <80% predicted baseline FVC at CLAD onset) (21, 22). Whilst total lung capacity (TLC) is not routinely undertaken and was not available on all recipients to allow its use in the definition, we utilised the spirometric definition detailed above to define RAS as detailed in the published consensus guidelines (21, 22). Declines in lung function/CLAD were treated according to the standard protocols of the time (18). For this analysis CLAD status, staging and phenotype were redefined as per ISHLT criteria (20, 23).

Induction therapy with the IL-2 receptor blocker, basiliximab, was given as a CNI sparing agent to LTR who were identified pre-transplant as being at higher risk of developing post-LTx renal dysfunction (n = 73). Subsequent strategies for LTR with renal impairment involved CNI reduction (n = 47) or elimination (n = 44); control of hypertension, diabetes, and cholesterol; and initiation of EVE (22). For LTR receiving EVE in combination with a CNI for a renal indication, further increases in serum creatinine would warrant eventual withdrawal of the CNI.

CMV prophylaxis, monitoring and treatment strategies are described elsewhere (24). Immediate post-transplant prophylaxis for all patients at risk of CMV infection received at least 7 days of intravenous ganciclovir followed by valganciclovir for a duration determined by risk category. Severe CMV infection or CMV reactivation was defined as >10,000 IU/mL in the blood or CMV >50,000 IU/mL in the bronchoalveolar lavage (BAL).

EVE was utilized in the setting of failure of first-line immunosuppressive strategies, e.g., significant renal impairment, CNI-neurotoxicity, malignancy (25, 18). EVE was prescribed with or without a CNI, determined by the degree of CNI intolerance. As per unit protocol, EVE was typically commenced at a dose of 0.25–0.5 mg twice daily with halving of the CNI dose. If EVE was to be used in conjunction with a CNI (minimization strategy) a trough concentration of 3–5 ng/mL for EVE and 4–6 ng/mL for tacrolimus would be targeted. If CNI cessation was planned (elimination strategy), a trough concentration of 5–7 ng/mL for EVE was targeted. Whenever EVE was utilized as part of a CNI elimination strategy, the CNI was ceased when EVE trough concentration was ≥3 ng/mL.

Continuous variables were summarized using means and standard deviations (SD) or medians and interquartile ranges (IQR) wherever appropriate. Categorical variables were expressed as counts and percentages. Overall survival was defined as the time from the date of starting EVE to the date of death or last follow-up. Time to CLAD was calculated from the date of starting EVE to the date of diagnosis of CLAD.

Univariate and multivariate analyses for overall survival and time to CLAD were performed using Cox proportional hazards regression with results reported as hazard ratios (HR) and 95% confidence intervals (95% CI). Variables with a p < 0.05 on univariable analyses or those deemed clinically relevant were considered for inclusion in the multivariable models.

To account for any possible imbalance between groups due to differences in baseline demographics and the evolution of treatments, propensity scores were included as an additional covariate in the regression models. Propensity score matching was also used to reduce selection bias from confounding factors between the EVE or CNI-based immunosuppression group. The individual propensities for being in the EVE group were estimated with the use of a multivariable logistic regression model that included date of transplant, age at transplant, sex, azithromycin and CMV reactivation as the predictor variables. All calculated p-values were two-tailed and a p < 0.05 indicated statistical significance. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, United States).

Changes in estimated glomerular filtrate rate (eGFR) over time (time of LTx, EVE commencement and 1-year post) was assessed using linear mixed models fitting main effects for time, group (EVE or CNI) and their two-way interactions.

The study was approved by the Alfred Hospital (252-12, 30 May 2012) and Monash University Ethics Committees (252-12, 19 April 2017).

Baseline demographics are described in Table 1. The most common indication for starting EVE was renal impairment (79%), followed by malignancy (8%), neurotoxicity (7%), intolerance to CNI (4%) and recurrent CMV (2%). The median time from LTx to initiation of EVE was 334 days [IQR: 155-604], with the median time of follow up for all LTR included being 1881 days [IQR: 993–2970].

The median survival for the entire cohort was 1881 days [IQR: 993–2970], (EVE: 1869 days [IQR: 910–3185] vs. CNI: 1944 days [IQR: 1196–2850]). One, three- and five-year survival was 80.2%, 68.1% and 59.9%, respectively. There was no difference in overall mortality between the groups (48% EVE vs. 45% CNI, p = 0.648) (Figure 1).

The baseline renal function between the two groups was comparable at the time of LTx (p = 0.478). Estimated GFR declined significantly from the time of LTx to the introduction of EVE for all included LTR (from 84.4 to 56.1 mL/min/1.73 m², p ≤ 0.0001). At the time of commencement of EVE, there was a significant difference in estimated glomerular filtrate rate (eGFR) between the two groups (EVE, 44.5 mL/min/1.73 m² vs. CNI, 67.7 mL/min/1.73 m², p ≤ 0.0001). At 1-year follow up, the eGFR in the EVE group was significantly lower than the CNI group (EVE, 56.2 vs. CNI, 64.1 mL/min/1.73 m², p = 0.03) (Figure 4). The changes in eGFR assessed over time between the two groups and their two-way interactions demonstrated significant interaction between group and time (p < 0.0001), suggesting that the EVE and CNI groups behaved differently over time (Figure 4).

Demographics of the two EVE strategies (CNI minimization and CNI elimination) are described in Table 6. The CNI elimination strategy was not associated with poorer survival (p = 0.158) or a faster time to CLAD (p = 0.944). When comparing the eGFR changes over time, there is no statistical difference between the two groups (p = 0.498).

There was no difference between the groups with regards to cause of death (Table 7). In particular, no difference in death due to CLAD (31.9% EVE vs. 29.6% CNI, p = 0.74). Additionally, whilst EVE was utilized primarily as a CNI-sparing agent to preserve renal function, there was no difference in mortality from renal failure (2.2% EVE vs. 1.1% CNI, p = 1.00).

The primary aim of our study was to determine whether LTR who started on EVE had poorer survival outcomes compared to those that remained on CNI-based immunosuppression. Overall survival, including CLAD related mortality was similar to those who remained on CNI-based immunosuppression, suggesting a similar trajectory for both groups. Our study demonstrated that EVE is a safe and effective option when prescribed for CNI intolerance, primarily nephrotoxicity or neurotoxicity as part of maintenance immunosuppressive regimens.

There has been some reluctance in using EVE as part of immunosuppressive regimens in the perioperative period due to challenges with the risk of wound dehiscence as well as destabilizing immunosuppression. While we would recommend EVE for stable long-term LTR as part of a maintenance immunosuppression regimen in LTR with CNI-induced nephrotoxicity, we would not recommend starting on EVE in the perioperative period.

We found no statistically significant increase in the incidence of CLAD, an accelerated progression to CLAD or a tendency towards a specific CLAD phenotype with EVE. Not unexpectedly, the greater the ventilatory reserve at the time of starting EVE (FEV₁ and FVC), the less likely the recipient would progress to CLAD.

Traditionally outcomes in LTR with RAS tend to be worse than those with BOS with shortened survival (28). We found on multivariate analysis, patients with BOS unexpectedly had a greater risk of death compared to those with RAS. The potential benefit of EVE in LTR with RAS and the potential mechanisms for these findings warrant further investigation. The development of CLAD has been suggested to be due to chronic fibroblast activation and EVE is known to downregulate fibroblast activity (10). A potential mechanism could be the antifibrotic activity of EVE influencing the chronic fibroblast activation present in CLAD.

Other benefits of EVE may be due to its impact on angiogenesis. Angiogenesis is a complex process in the transplanted organ and involves cellular proliferation, vascular remodelling, and endothelial activation. EVE has anti-angiogenic effects in vitro that prevent cellular proliferation, vascular remodelling, and endothelial activation with potential benefits in chronic allograft rejection (9). EVE may indeed have promoted a stabilizing influence on pulmonary function once CLAD was established. These findings also reassure us that there was no long-term clinically significant underlying EVE lung toxicity (29).

On multi-variate analysis, as a time-dependent variable, mycophenolate containing immunosuppressant regimens were associated with improved survival. At our institution, mycophenolate is utilized as a second line agent in sensitized recipients or LTR who are commenced for the management of ACR. Azzola et al. demonstrated that EVE and mycophenolate were the two most potent antifibroproliferative drugs at concentrations achieved clinically (30). Low doses of EVE and mycophenolate can achieve at least 50% inhibition of fibroblast proliferation at therapeutic doses (30). The combination of mycophenolate and EVE may provide a synergistic benefit in LTR with CLAD due to their antifibrotic activity warranting further investigation.

The most common indication for starting on EVE in our cohort was renal preservation due to CNI-related nephrotoxicity. Preservation of renal function is paramount post-LTx as nephrotoxicity is associated with significant morbidity and mortality (31). EVE provides an alternate immunosuppressive agent in the LTx setting with potentially less nephron loss over time from reduced long-term exposure to CNIs (3, 4). Other studies in LTx have investigated the change in renal function parameters following the introduction of EVE (3, 4). Similarly, to these studies we found no increase in renal related mortality with the introduction of EVE as part of maintenance immunosuppressive regimens.

Although on univariate analysis, LTR who started EVE for renal preservation appeared to progress to CLAD faster than those that did not, this was not borne out in multivariate or survival analyses. A possible explanation for the faster time to CLAD is the contribution of renal impairment to lung function decline. It is known that LTR with chronic kidney disease demonstrate abnormalities in lung function including obstructive and restrictive ventilatory defects, and impaired diffusing capacity (32).

Additionally, it is likely that those started on EVE-based regimens due to renal impairment had a period of subtherapeutic CNI levels with the aim of stabilizing renal function in the short-term. These potential prolonged periods of low CNI levels may have led to the immunological risk of periods of suboptimal immunosuppression contributing to an increased risk of CLAD. We would suggest that in those with CNI intolerance, it may be prudent to consider starting EVE earlier rather than continue with subtherapeutic levels of CNI and the risk of inadequate immunosuppression.

Our study has several limitations. Firstly, the findings are retrospective, reflect a single center experience and the EVE prescribing patterns of this center have evolved since 2005. Immunosuppression regimens and target levels would be adjusted over time according to unit protocol and modified further according to response and tolerance. In addition, the study cohort was heterogenous, with indications for EVE use including nephrotoxicity and neurotoxicity and regimen strategies that varied from CNI minimization to elimination.

Secondly, as our center does not routinely monitor TLC this could not be utilized in the definition of RAS and therefore spirometric measures as previously detailed were incorporated into our criteria for RAS diagnosis (21, 23).

EVE-based maintenance immunosuppression can be successfully and safely be utilized when CNI minimization or elimination is required. Most importantly, our analyses demonstrated that starting EVE does not increase the risk of death or accelerate the progression to CLAD.