Everolimus Versus Mycophenolate Mofetil De Novo After Lung Transplantation: A Prospective, Randomized, Open-Label Trial

Introduction

As an established treatment for end-stage lung diseases, lung transplantation has approached a 1-year survival rate of above 80% 1. Therefore, pretransplant morbidity and chronic lung allograft dysfunction (CLAD) 2 have become the major limiting factors for long-term graft and patient survival. Along with viral infection and ischemic injury, acute rejection has been found to be a major risk factor for the development of CLAD 3. Thus, efforts have been made to improve immunosuppressive therapy so that the incidence and severity of both viral infections and acute rejections are minimized. Moreover, immunosuppressive protocols after lung transplantation seem to be more aggressive than after any other solid organ transplantations. Therefore, high exposure to immunosuppressive drugs and the effects of comedication lead to a high incidence of adverse events 4 and toxicity.

Everolimus is a proliferation signal inhibitor and has demonstrated in vitro and, in preclinical models of acute rejection, enhanced immunosuppressive potential 5. Everolimus has been used for triple immunosuppressive therapy following kidney transplantation as an alternative to cell cycle inhibitors such as azathioprine, showing less acute rejection and fewer cytomegalovirus (CMV) infections 6, 7. When used in heart transplant recipients, a trend of fewer acute rejections and a definitive advantage in avoidance of allograft vasculopathy were reported 7.

For lung transplant recipients, the efficacy of Everolimus was compared to azathioprine in maintenance immunosuppression in a multicenter prospective trial: No difference was seen in the rates of graft loss and death at 24 months follow-up. Also, a higher incidence of kidney failure was reported 8. In addition, a study by Bhorade et al compating azathioprine to sirolimus, another mTOR inhibitor, failed to demonstrate less rejections in the sirolimus arm 9.

The open question of possible advantages of Everolimus in lung transplantation encouraged us to perform a prospective clinical trial. Unlike in multicenter studies, avoiding the diversity of lung transplant protocols, a study at a high-activity single center could be beneficial. In addition, Everolimus was compared to the more contemporary cell cycle inhibitor mycophenolate mofetil. It was administered not as a fixed dose but in a trough level–based regimen and was administered relatively early after lung transplantation.

Methods

Results

A total of 190 patients after de novo lung transplantations were enrolled in a single center in a time period of 4.5 years. A flowchart of the study is displayed in Figure 1. Forty percent of the individuals undergoing lung transplantation during the study period were included. The patients were randomized so that 95 patients were in each the MMF group and the Everolimus group. Groups were well balanced according to baseline demographics (Table 1).

Ninety-seven patients (51%) completed 2 years on drug treatment, defining the per-protocol (PP) population. The most frequent reasons to discontinue were recurrent acute rejection or BOS onset (49%). Discontinuation occurred more frequently in the Everolimus group (55%) than in the MMF group (43%). Wound healing problems led to discontinuation of Everolimus in 8% of patients. Discontinuation was performed median on day 118 (interquartile 41–322).

Acute and chronic organ dysfunction

In the ITT population, the incidence of BOS was 20/95 (21%) patients in the MMF group and 13/95 (14%) in the Everolimus arm (p = 0.250; −7.3% difference, 90% confidence interval [CI]: −18.1% to 3.4%). As for the PP population, the incidence of BOS was 8/54 (15%) in the MMF group and 1/43 (2%) in the Everolimus arm (p = 0.041; 12.5% difference, 90% CI: 3.7% to 21.3%). In the Everolimus group, fewer biopsy-proven rejection episodes were observed, while the number of steroid pulse courses did not differ between groups (Table 2). Analyzing the PP population, 2-year BOS-free survival was 98% in the Everolimus group and 89% in the MMF group (p = 0.09). BOS-free survival patterns for the ITT population show that the hazard for BOS is increasing over time in both groups (Figure 2).

Table 2. Secondary outcome parameters (Chi-squared test) with respect to infection, acute rejection, and kidney function

Item	MMF occurrences (%)	Everolimus occurrences (%)	p-value
Infections, any episode—No. (%)	264 (278)	225 (237)	—
Patients with at least one infection—No. (%)
Bacterial, any	39 (41)	46 (48)	0.307
Lower respiratory tract infection	115 (121)	88 (93)	0.003
Pneumonia	21 (22)	19 (20)	0.722
Viral, any	14 (15)	12 (13)	0.673
CMV antigenemia	16 (17)	4 (4)	0.005
CMV disease	2 (2)	0 (0)	0.155
Herpes infection	7 (7)	8 (8)	0.788
Fungal infection, any	23 (24)	18 (19)	0.378
Gastrointestinal infection	13 (14)	25 (26)	0.030
Steroid pulse therapy	21 (22)	19 (20)	0.722
Steroid pulse therapy with ≥A1	16 (17)	4 (4)	0.005
GFR, 12 months < 30 – n (%)	4 (6)	7 (14)	0.148
GFR, 24 months < 30 – n (%)	4 (8)	7 (17)	0.168

• Thirteen tests (0.10/13 = 0.00769), Bonferroni correction: p-values < 0.0077 are deemed to be significant. MMF, mycophenolate mofetil; CMV, cytomegalovirus; GFR, glomerular filtration rate.

CsA exposure was intentionally reduced in the Everolimus group (Figure 3). A total number of 34 patients (15 MMF, 19 Everolimus) were excluded from the study for suspected recurrent rejection. It was believed to be unsafe to continue with cyclosporine-based immunosuppression, and patients were switched to Tacrolimus.

Posttransplant infections

Secondary outcomes are displayed in Table 2. Lower respiratory tract infections were significantly less frequent in the Everolimus-treated patients. CMV infections significantly increased in the MMF group (Table 2 and Figure 4). Fungal infections (all tracheobronchial) did not differ between groups.

Discussion

This is a large prospective randomized study with Everolimus and reduced calcineurin inhibitors (CNIs) de novo after lung transplantation. The substitution of a cell cycle inhibitor by a proliferation signal inhibitor and reduced CNI proved to be safe. No evidence of statistically significant reduction in the incidence of BOS and acute rejection was observed with treatment using the ITT analysis.

The original hypothesis of a reduction of BOS incidence by 15% within 2 years still seems reasonable; however, only about 50% of patients completed the study. The estimated dropout rate was more than doubled and therefore proves that the hypothesis failed. The analysis of the per-protocol cohort revealed a possible benefit for the Everolimus group (1/43) versus 8/54, p = 0.041. However, the higher rate of discontinuation in the Everolimus arm may have introduced a selection bias for the per-protocol analysis.

A similar result was seen in the most recent multicenter trial 14. By comparing similar immunosuppressive protocols, Glanville et al. could also not show any significant differences in BOS incidence in the ITT group 14. This study was also underpowered with a comparable dropout rate of more than 50%. However, in their per-protocol analysis, advantages were also seen for the Everolimus group for a potential reduction of BOS incidence. The high dropout rates in both studies stress the general challenge of these investigations.

Important observations were made in the secondary endpoints of our study. Use of Everolimus starting 28 days after lung transplantation is safe and allows further refinement of immunosuppressive protocols for this challenging patient population. Biopsy-proven acute rejections were even less frequent in Everolimus-treated patients 14. This finding was confirmed by Glanville et al.

Compared with older studies, two important changes were made in Glanville's study and in our protocol: The administration of Everolimus was not given as a fixed dose but in a trough level–adjusted manner 14-16. In addition, both protocols reduced the CsA exposure (22–28% in our study and 25–50% in Glanville's study) intentionally in the Everolimus arm 14-16. Lowering the CsA levels in the Everolimus arm did not lead to a higher rate of acute rejection (AR). The importance of the CsA reduction was seen in five early patients in this study without immediate reduction who developed thrombotic angiopathy 15. Once this serious adverse event became evident, additional cases could be avoided by ensuring the timely reduction of CsA exposure in all patients.

As a single-center, open-label study over a prolonged period of time, changes in patient management may have caused some bias. The unfamiliar Everolimus use in particular may have caused a bias of discontinuation of Everolimus in the early phase of the study and a potential hesitation of decreasing CsA levels in the Everolimus arm. This is an unproven hypothesis but may present a serious limitation of the study. It would explain, for example, the discrepancy of biopsy-proven AR and the number of suspected ARs (the number of steroid treatment episodes) in the Everolimus arm (Table 2). A total of 34 patients were withdrawn from the study because of suspected or proven recurrent rejection. This was seen as an important safety step for the study and was ultimately a limitation as well. These patients were treated with a Tacrolimus-based protocol, which may be more effective in lowering AR and subsequently also BOS 16. An additional limitation is that only patients with a beneficial perioperative course were selected for the study. It remains unknown if patients with primary graft dysfunction or other early perioperative limitations would show similar results. As the high dropout rate is the main limiting factor, the results of this study should be interpreted with caution. Moreover, the agreement of reviewers on the onset of BOS was only 64.4%, indicating the importance of external review but also the difficulty of the parameter. Despite this, the high similarity in the results and the dropout rate seen in the multicenter trial of Glanville et al. and our findings is striking 14.

Another important observation is the reduction of CMV antigenemia in the Everolimus-treated population. This finding was confirmed after other solid organ transplants, like kidney 17 and heart 6. In addition, in the study of Glanville et al, a higher incidence of opportunistic infections was described in the mycophenolate arm when compared to the Everolimus arm, including CMV infections. In a post hoc analysis in another clinical trial, this finding was also seen when sirolimus was compared to azathioprine 18. In addition, we found fewer lower respiratory tract infections in the Everolimus group. It may be speculated that maintaining a better humoral immune response associated with Everolimus, which allows for significant rises in antigen-specific IgG antibodies, plays a role in mitigating CMV viremia 19 and lower respiratory tract infection.

A critical finding of severe leucopenia (<2/nL), which may be associated with comedication of MMF in the presence of anti-CMV prophylaxis using valganciclovir 20, could be prevented in Everolimus-treated patients and therefore prevent a potentially life-threatening complication after lung transplantation. Significantly less leucopenia and fewer episodes of diarrhea were also reported in Glanville's trial 14.

One of the adverse events of greatest concern is the development of renal insufficiency early after lung transplantation. Despite the fact that our study population showed intact kidney function at the time of enrollment, a remarkable drop in GFR was seen in both arms during the first 6 months after transplant. Thereafter, the values remained stable but unimproved. This indicates that current immunosuppressive protocols in addition to other nephrotoxic comedication impair kidney function significantly in the early phase after lung transplantation. A decrease of GFR by almost 50% was previously reported after lung transplantation 21, and 25% of patients are reported to show renal dysfunction (defined as an increase in creatinine) within 1 year after lung transplantation 22. It seems to be standard practice in many programs to start acting on CsA levels when serum creatinine levels start to rise. With this knowledge, we suggest using GFR as a parameter to prevent an early decrease in kidney function. Through the use of trough level–adjusted Everolimus exposure, we were able to avoid excess morbidity in the Everolimus group. However, the overall decline of kidney function remains worrisome.

In contrast to the Glanville study, we used CsA trough levels and not C₂ target levels as suggested 14, 23, 24. This was decided for logistical and compliance issues to obtain CsA levels at the same time Everolimus trough levels were measured. It can be speculated that including C₂ target levels in the trial could have enhanced efficacy 25 and renal protection.

In many studies, Everolimus has been used to reduce the nephrotoxicity of immunosuppressive protocols: A protocol that lowers CNI levels in the presence of Everolimus was suggested to improve kidney function 26. Given the fact that the time of highest nephrotoxicity coincides with the time of highest incidence of AR, a trade-off in efficacy and nephrotoxicity seems to determine the course. A novel approach was published and studied a quadruple approach of low-dose CNI, Everolimus, MMF and steroids to improve kidney function after thoracic transplantation: the Nordic Everolimus (Certican) Trial in Heart and Lung Transplantation (NOCTET) trial 27. Improvements in GFR were found as well as acceptable efficacy and safety data. However, the impact of this approach was limited to patients with a baseline GFR between 20 and 60 mL/min/1.73 m² and time since transplant of less than 4.6 years 28.

The findings of this study as well as the NOCTET trial led to the initiation of a multicenter trial in Germany to use a quadruple low-dose maintenance regimen including Everolimus in comparison to standard triple-drug maintenance therapy in order to ensure efficacy and avoid kidney injury in lung transplant recipients.

In summary, all clinical studies so far have failed to prove an advantage of proliferation signal inhibitors to reduce the incidence of BOS after lung transplantation. All studies demonstrated a high withdrawal rate from the clinical protocol, leaving the studies underpowered to prove the primary endpoint in an intention-to-treat analysis. In addition, our study as well as others required external review to determine the onset of BOS. Difficulty to pinpoint the onset of BOS as well as the current discussion about different forms of chronic allograft dysfunction question the role of BOS to be a useful primary endpoint parameter.

However, when per-protocol patients were analyzed, advantages and also specific risks in Everolimus patients were seen. It seems that no standard protocol works for the majority of patients, even in lung transplant patients highly selected for clinical trials. Individualized approaches still have to be used in most lung transplant recipients. The strength of Everolimus may be another option in the armamentarium for this purpose. Other advantages such as the avoidance of AR, CMV and lower respiratory tract infection and the absence of leukopenic episodes are hallmarks of this therapy. The risks of wound healing problems, kidney injury and thrombotic microangiopathy need to be addressed by the awareness and consequent reduction of CNI comedication. A switch from creatinine to GFR-based 29 follow-up protocols may be useful to protect kidney function. Quadruple immunosuppression may play a future role after lung transplantation.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.