Effect of antifibrotic agents on postoperative complications after lung transplantation for idiopathic pulmonary fibrosis

Antifibrotic agents (AFAs) are now standard‐of‐care for idiopathic pulmonary fibrosis (IPF). Concerns have arisen about the safety of these drugs in patients undergoing lung transplantation (LTx).


INTRODUCTION
Idiopathic pulmonary fibrosis (IPF) is a progressive fibrosing interstitial lung disease of unknown cause associated with a pathologic and/or radiological pattern of usual interstitial pneumonia. 1The current pathophysiologic explanation is an alveolar epitheliopathy characterized by repetitive injury.The standard of care includes the use of antifibrotic agents (AFAs), either pirfenidone or nintedanib. 2 Marketed in Europe since 2011 and in the United States since 2014, pirfenidone inhibits fibroblast proliferation, reduces myofibroblast differentiation and inhibits transforming growth factor beta-induced synthesis of collagen type I. 3 In 2014 in the United States and 2015 in Europe, nintedanib received approval for treating IPF.This intracellular tyrosine kinase inhibitor blocks growth factors involved in IPF pathophysiology.It also has antiangiogenic effects, even though the link between this effect and its antifibrotic activity remains unclear. 4hese two drugs could reduce the decline in forced vital capacity (FVC) as compared with placebo [5][6][7] and have a modest effect on mortality. 8Despite these positive results, they do not represent a curative treatment.IPF remains a fatal disease with a median survival estimated at 2-5 years from the time of diagnosis. 9Early evaluation and referral for lung transplantation (LTx) is still recommended. 10Moreover, nintedanib is now used to treat progressive fibrosing interstitial lung diseases (ILDs), 11 thus increasing the potential number of patients undergoing LTx while receiving nintedanib.
Concerns have arisen regarding the safety of these drugs in patients undergoing LTx and potential complications, particularly impaired wound healing resulting in bronchial anastomotic or skin healing issues.Because of the inhibitory action of nintedanib against vascular endothelial growth factor receptor, haemorrhagic complications might also be expected.
3][14][15][16] Available studies mostly concern small numbers of patients with low event rates.One of the largest studies, 13 performed in Australia, analysed 226 IPF patients who underwent LTx; 40 patients were under AFA treatment at the time of LTx, but only 11 were receiving nintedanib.][19] Another limitation of available studies is the low proportion of patients receiving nintedanib because of its later authorization.According to the European Medical Agency, treatment with nintedanib should be initiated based on clinical judgement of adequate wound healing only and if the anticipated benefit outweighs the potential haemorrhagic risk. 20More recently, Bendstrup et al. suggested discussing all options for the management of nintedanib before LTx but ideally to maintain the treatment. 21his lack of strong consensus in AFA treatment is explained by the insufficient data on maintaining the treatment until LTx.Reassuring first results must be confirmed by further studies.The aim of this study was to determine whether maintaining AFAs in the peri-operative period leads to increased bronchial anastomoses issues, delay in skin healing and haemorrhagic complications.

Study design
We used a multi-centre, nationwide, retrospective, observational analysis including all patients who underwent LTx for IPF at each of the six participating centres (Paris Bichat, Paris Marie Lannelongue, Paris Foch, Lyon, Marseille and Strasbourg) between 1 February 2011 and 31 December 2018.February 2011 was chosen as the date of authorization of pirfenidone use in France.
Patients who were still receiving one of the AFAs (pirfenidone or nintedanib) within 4 weeks before LTx were included in the AFA group.Patients who did not receive any AFA or who discontinued it more than 4 weeks before LTx were included in the control group.The time period of 4 weeks was defined according to previous studies. 12,22No specific duration of therapy was required for inclusion in the AFA group.Patients who had sequentially received both AFAs were included in the group corresponding to the last received treatment.Clinical outcomes were monitored up to 3 months after LTx.The current study was approved by the Institutional Review Board of the French society for respiratory medicine (Société de Pneumologie de Langue Française [CEPRO 2020-047]).Patients were informed of the use of their anonymized data and were given the opportunity to decline this use.

Selection criteria
Data for all patients with ILD were analysed.Inclusion criteria were age ≥18 years, a diagnosis of IPF and single or bilateral LTx performed between February 2011 and December 2018.Diagnosis of IPF was assessed by each centre according to the official American Thoracic Society/European Respiratory Society/Japanese Respiratory Society/Latin American Thoracic Association guidelines after a multidisciplinary discussion involving pulmonologists, radiologists and pathologists experienced in the diagnosis of ILD. 1 The diagnosis required the exclusion of other known causes of ILD, the presence of a usual interstitial pneumonia pattern on high-resolution chest tomography (HRCT) or the combination of HRCT and surgical lung biopsy showing a usual interstitial pneumonia pattern.The pathology of the explanted lungs was reviewed, and data for only patients with usual interstitial pneumonia were included.
The exclusion criterion was a history of another solid organ transplantation before LTx.

SUMMARY AT A GLANCE
This multi-centre nationwide, retrospective study including 58 patients receiving antifibrotic agents before lung transplantation for idiopathic pulmonary fibrosis shows that their use did not increase airway, bleeding or wound post-operative complications and could be associated with reduced rates of primary graft dysfunction and 90-day mortality.

Data collection
Patient demographics and pre-transplant and posttransplant data were obtained from the medical files at each centre.
Demographics and pre-transplant data included age; sex; duration of AFA therapy; results of the most recent biologic tests, complete pulmonary function tests, arterial blood gas tests and 6-min walk distance; mean pulmonary pressure and identification of a genetic mutation associated with IPF.The presence of pulmonary hypertension was defined as mean pulmonary arterial pressure > 20 mm Hg and pulmonary vascular resistance >3 Wood units 23 at the time of placement on the LTx wait list.Transplant data included time on the wait list before LTx; type of LTx (single or bilateral); priority access (high emergency or elective); use of extracorporeal membrane oxygenation (ECMO) before, during or after LTx; duration of mechanical ventilation (defined as time to first extubation, death or weaning from mechanical ventilation in case of tracheotomy); length of stay in the intensive care unit (ICU); duration and dose of noradrenaline; and use of treatments with a potential impact on risk of post-transplant complications (immunosuppressive, antiaggregant, anticoagulant and antifungal therapies).Primary graft dysfunction (PGD) after 72 h was defined according to the International Society for Heart and Lung Transplantation guidelines. 24The same investigator reviewed blood gas and chest x-ray results to grade PGD.

Clinical outcomes
Post-transplant complications occurring within 3 months postoperatively were recorded.Bronchial anastomotic complications and skin healing disorders were assessed.For the diagnosis of bronchial anastomotic complications, the patients were retrospectively classified according to the International Society for Heart and Lung Transplantation statement 25 on the basis of the endoscopy reports.
Dehiscence was defined as any degree of separation of the bronchial wall at the level of anastomosis. 25Fistula was defined as an abnormal connection between the bronchial lumen and adjacent structures (i.e., pleural space, vessels or pericardium).Necrosis and ischemia were not evaluated because of lack of reproducibility in a retrospective design.Stenosis was recorded if endobronchial therapy (balloon dilatation, stent placement or laser therapy) was performed.Skin healing disorders included chest wall dehiscence and groin complications.Given the absence of standardized definition, these complications were defined by the requirement of either surgical revision, or use of systemic antibiotics.
We collected all significant haemorrhagic findings (i.e., haemoptysis, chest wall haemorrhage or any other haemorrhage).The need for blood transfusion, re-operation or arterial embolization was recorded.The 30-and 90-day mortality rates were recorded and causes of death were assessed from medical records.

Statistical analysis
Continuous variables are presented as mean and standard deviation if normal distribution and were compared by Student t-test.Continuous variables with non-normal distribution are reported as median (interquartile range) and were compared by Mann-Whitney test.Categorical variables are presented as number and proportion and were compared by chi-square test or Fisher exact test as appropriate.To determine risk factors associated with complications (bronchial, haemorrhagic or cutaneous), mortality and PGD, univariate and multivariate analyses were performed with a Generalized Linear Model.The multivariate model included variables with p < 0.2 on univariate analysis or variables known to be associated with the outcomes from the literature.Backward selection was applied to identify the final multiple variable model, estimating odds ratios (ORs) and 95% confidence intervals (CIs).Survival curves were plotted with the Kaplan-Meier method and a log-rank test was used to assess differences across survival curves.We also performed a complementary analysis by excluding patients who stopped AFAs between 1 and 28 days before LTx, from the AFA group.All data analyses involved two-sided tests, with p < 0.05 considered statistically significant.Statistical analyses involved using R 4.1.1(R Foundation for Statistical Computing, Vienna, Austria) and GraphPad Prism V8 (GraphPad Software).

Study group
Between February 2011 and December 2018, a total of 1987 LTx procedures were performed in the six participating centres, including 205 patients with IPF; 58 (28.2%) of these had received AFAs in the month before LTx.The type of AFA was pirfenidone in 37 (18%) patients and nintedanib in 21 (10.2%)(Figure 1).The median duration of AFA therapy before LTx was 13.8 (5.6-24) months.The control group consisted of 147 patients: 115 had no AFA therapy and 32 had discontinued AFA therapy more than 28 days before LTx.

Pre-transplant characteristics
Pre-transplant characteristics are summarized in Table 1 and Table S1 in the Supporting Information.At the time of LTx, vital capacity was lower in the control group than AFA group, but total lung capacity, carbon monoxide diffusion capacity and gas exchange were similar, as was the frequency of pulmonary hypertension.The two groups did not differ in previous anticoagulant or antiaggregant therapy.

Clinical course of LTx
The main characteristics appear Table 2 and Table S2 in the Supporting Information.Median time on the wait list was 31 (10-87) days.Bilateral LTx was performed in 125 (61%) patients, and 57 (27.8%) underwent LTx following a high emergency allocation of the graft.The two groups did not differ significantly in length of ICU stay or use of ECMO or vasopressors.Median duration of mechanical ventilation and use of vasopressors was lower in the AFA than control group (48 [24-192] vs. 96 [24-252] h, p = 0.03 and 66 [24-144] vs. 48 [24-192] h, p = 0.02, respectively).The rate of PGD grade ≥1 was lower in the AFA than control group (n = 15, 25.9%, vs. n = 63, 42.9%, p = 0.02).The difference was still significant for PGD grade 3 (p = 0.04).The two groups did not differ in fungal colonization ( p = 0.34).surgery and embolization were required in 39, 34 and 7 patients, respectively.The two groups did not differ in skin healing complications (p = 0.70).

Survival
The 90-day mortality was lower in the AFA than control group (6.9% vs. 18.4%, p = 0.046; Table 3, Figure 2).F I G U R E 2 Kaplan-Meier curves for cumulative survival.Data for the survival analysis were censored at 90 days.

Complementary analysis
We performed a complementary analysis by excluding patients who stopped AFAs between 1 and 28 days before LTx, from the AFA group (Table 1, Table S1 in the Supporting Information).On comparing the control group and the patients who maintain AFA up to the day of LTx, we found the same results regarding postoperative complications, with no difference between the two groups in anastomotic, skin cicatrisation and haemorrhagic complications.However, the difference in mortality rate between the treated and control groups did not reach statistical significance (p = 0.064).

DISCUSSION
This study examined the postoperative outcome after LTx for IPF according to the preoperative use of AFAs.We found no difference between the AFA and control groups in frequency of airway complications, skin healing issues and bleeding complications, but the rates of PGD and 90-day mortality were significantly decreased in patients who were still receiving AFA within 4 weeks before LTx.
In 2016, concerns arose about the risk of anastomotic airway complications, skin healing issues and bleeding associated with the preoperative use of AFAs in patients undergoing LTx. 26 The potential side effects of these agents led some transplant teams to withdraw such therapies at the time of placement on the wait list for LTx.This is an important issue with immediate consequences for the care of patients with IPF considered for LTx.Sustaining AFAs is beneficial to preserve the best possible respiratory function before LTx and may help prevent acute exacerbation or death in IPF patients while on the wait list.As a reminder, according to the Organ Procurement and Transplantation network/Scientific Registry of Transplant Recipients 2021 Annual Data Report, restrictive lung disease was the lung diagnosis associated with the highest mortality rate while on a wait list. 27However, the risk of potential post-operative complications in patients receiving AFAs could have a negative impact on the pre-transplant benefit of AFA therapy.
In our study, we did not find any increase in rate of bronchial anastomosis complications, skin healing issues and bleeding that could be attributed to the preoperative prescription of AFAs in IPF patients.6][17][18][19]22,26,28,29 None of the previous studies highlighted an increase in complications.Besides the small number of patients receiving AFA (as in our study) and the fact the type of AFA was pirfenidone in most cases, another limitation of these previous studies focusing on complications potentially related to AFAs was the small overall event numbers.For example, the rate of anastomotic airway complications after LTx ranged from 2% to 18% in the International Society for Heart and Lung Transplantation consensus statement 25 and reached 33% in some series. 30,31However, several studies of the effect of AFA therapy before LTx reported bronchial anastomotic issues in less than 1%-3.2% of cases. 12,15,22Moreover, the definition of the type of airway complication was not always precise in previous studies. 12,14,22Finally, regarding the theoretical bleeding risk related to nintedanib, only a few studies focused on haemorrhagic complications.
Although an influence of pirfenidone and nintedanib on wound healing could have been expected because of their mechanisms of action, the lack of observed effect could be explained by the short half-lives of these agents (2.4 h and 10-15 h for pirfenidone and nintedanib, respectively).The time of 4 weeks between the cessation of treatment and the LTx, although arbitrary, was chosen according to previous studies. 12,22This choice of 4 weeks is arguable because of no evidence of a persistent effect of AFAs after cessation.4][15] To exclude any interference with our results, we performed a complementary analysis excluding all patients who stopped AFA less than 4 weeks before LTx and did not find any difference regarding post-operative complications in this additional analysis.
An interesting point is the lower mortality rate in the AFA than control group, which was not expected.Although most previous studies did not find any difference in mortality, Mackintosh et al. 13 reported lower 1-year mortality in the AFA than control group.However, we need to be cautious about the beneficial effect of AFA on the 90-day mortality because this benefit was not found in the subgroup of patients who maintained AFA up to the day of LTx.We hypothesize that the discrepancy between the two results could be related to a lack of power of the complementary analysis.Recently, several studies suggested the protective role of pirfenidone on LTx outcome. 32,33In a rat model, pirfenidone alleviated lung ischemia-reperfusion injury. 32Furthermore, a clinical retrospective human study found lower duration of mechanical ventilation and lower incidence of PGD in patients who underwent single LTx while receiving pirfenidone as compared with untreated patients. 33It could be hypothesized that AFA use is associated with less severe PGD and in turn to better survival.These reassuring data, combined with previous results, suggest that maintaining AFAs up to LTx may be beneficial for patients.This is of interest because AFAs may attenuate pre-transplant decline in vital capacity and reduce the risk of an acute exacerbation of IPF while waiting for LTx. 34he main limitation of the study is its retrospective design.Therefore, we had an unavoidable missing data issue.Furthermore, we did not match treated patients with controls, and control patients had lower vital capacity than AFA patients.These differences were considered by performing a multivariate analysis including vital capacity as a covariate.Beside better FVC, the AFA group had a shorter duration of surgery and although statistical significance was not reached, this group was also characterized by a lower proportion of priority access and by a longer 6-min walking distance.The fact that the AFA group seemed to be more fit overall might have played a role in the better 90-day mortality we observed in this group.Another unavoidable limitation was the difference in care of patients between centres.We tried to reduce this bias by collecting mostly objective data.This latter point is also a strength, allowing us to extrapolate the results more reliably.Lastly, although our cohort is one of the largest on this subject, another obvious limitation of the study is the quite low number of patients receiving AFA.
A strength of the study is that we collected information from six major French LTx centres in order to include a higher number of patients.Thus, we were able to gather a quite large cohort of IPF patients undergoing LTx while receiving AFAs, with a substantial number receiving nintedanib.Another strength is the use of a precise definition of airway anastomotic problems.Furthermore, we not only focused on airway complications but also analysed haemorrhagic and skin healing complications.All the included patients had confirmed IPF after multidisciplinary discussion, and the pathological analysis of the explants showed a usual interstitial pneumonia pattern in all cases.

CONCLUSION
In this large, multi-centre, cohort study, previous use of antifibrotics did not increase the probability of postoperative airway, bleeding and skin healing complications after LTx.Further studies are needed to assess the effect of antifibrotics on PGD and 90-day mortality.Our data strongly suggest the safety of sustaining AFA therapy after placing IPF patients on a wait list for LTx.
The most common causes of death were infection (n = 11) and haemorrhage (n = 4), including one bronchovascular fistula.All the cases of death by infection were observed in the control group and 3 out 4 patients who died from haemorrhage belonged to the control group.Logistic regression modelling of 90-day mortality included the following factors in the model: treatment group, age, body mass index, vital capacity, priority access to surgery, bilateral LTx, duration of surgery, PGD grade 3, induction treatment, use of ECMO (pre-, peri-and post-transplantation) and acute rejection.AFA therapy was the only protective factor associated with 90-day mortality in the adjusted model (OR 0.25; 95% CI

a
Patients who did not receive any AFA or who discontinued it more than 4 weeks before LTx.b Based on chi-square test or Fisher test when sample size was small (<5).c Includes gastrointestinal bleeding, epistaxis, tracheotomy bleeding, haemoperitoneum and haemothorax.
T A B L E 1 Pre-transplant characteristics of the study cohort.Perioperative clinical course.
Note: Continuous values are presented as mean ± SD or median (range).Abbreviations: 6MWD, 6-min walk distance; AFA, antifibrotic agent; aPPT, activated partial thromboplastin time; DLCO, diffusion capacity for carbon monoxide; LTx, lung transplantation; MD, missing data; mPAP, mean pulmonary arterial pressure; NTN, nintedanib; PCO 2 , partial pressure of carbon dioxide; PFD, pirfenidone; PH, pulmonary hypertension; PO 2 , partial pressure of oxygen; PT, prothrombin time; TLC, total lung capacity; VC, vital capacity. a Patients who did not receive any AFA or who discontinued it more than 4 weeks before LTx.bComparison between control and treated groups.cPatientswhoreceivedAFAtherapyat any time before LTx (within 4 weeks before LTx in the AFA group, by definition).Airway complications, including anastomosis dehiscence and fistulae, occurred in 55/205 (26.8%) patients: 38 (25.9%) episodes in the control group and 17 (29.3%) in the AFA group (Table3, TableS3in the Supporting Information).Fistula was observed in 28 patients and was bronchopleural in most (n = 25).The two groups did not significantly differ in overall airway complications (p = 0.91).Stenosis was observed in 20 patients during the study period: 16 in the control group and 4 in the AFA group, with no significant difference (p = 0.39).Overall, the frequency of haemorrhagic complications was not higher in the AFA than control group (p = 0.12).Haemorrhagic complications were observed in 66 (32.2%) patients, most (51 of 66) concerning the chest wall.Among these 66 episodes, blood transfusion,T A B L E 2 Post-operative complications and survival in control and AFA groups T A B L E 3Note: Dehiscence include bronchial dehiscence and fistula.Abbreviations: AFA, antifibrotic agent; ECMO, extracorporeal membrane oxygenation; MD, missing data.