Primary graft dysfunction: lessons learned about the first 72 hours after lung transplantation

Abstract

Purpose of Review

In 2005, the International Society for Heart and Lung Transplantation published a standardized definition of primary graft dysfunction (PGD), facilitating new knowledge on this form of acute lung injury that occurs within 72 hours of lung transplantation. PGD continues to be associated with significant morbidity and mortality. This review will summarize the current literature on the epidemiology of PGD, pathogenesis, risk factors, and preventative and treatment strategies.

Recent findings

Since 2011, a number of manuscripts have been published that provide insight into the clinical risk factors and pathogenesis of PGD. In addition, several transplant centers have explored preventative and treatment strategies for PGD, including use of extracorporeal strategies. More recently, results from several trials assessing the role of extracorporeal lung perfusion may allow for much-needed expansion of the donor pool, without raising PGD rates.

Summary

This review will highlight the current state of the science regarding PGD, focusing on recent advances, and set a framework for future preventative and treatment strategies.

Introduction

Primary graft dysfunction (PGD) after lung transplantation is a significant source of early morbidity and mortality.(1-7) We will review recently published literature on the epidemiology, pathogenesis, risk factors, prevention, and treatment of PGD.

Definition

PGD, a form of acute lung injury occurring within 72 hours of lung transplantation, is characterized by hypoxemia and alveolar infiltrates in the allograft.(2) Early case reports of post-operative allograft dysfunction used different criteria to define the syndrome and referred to it by various names.(4, 8-12) In 2005, the International Society for Heart and Lung Transplantation (ISHLT) standardized the definition of PGD, permitting consistent classification of this clinical syndrome.(2) PGD is now graded based on PaO₂/FiO₂ (P/F) and the presence of bilateral radiographic infiltrates consistent with non-cardiogenic pulmonary edema. The definition also requires exclusion of contributing factors, including hyperacute rejection, pulmonary venous anastomotic obstruction, cardiogenic edema, and pneumonia.(2) The severity of PGD is re-assessed daily up to 72 hours after reperfusion (Table 1). Grade 3 PGD (PGD3) corresponds to a P/F of less than 200 with allograft infiltrates, on ECMO or ventilated with FiO2 above 0.5 and on iNO. (2)

Table 1.

Grading system for PGD(2)

Grade	PaO2/FiO2	Radiographic infiltrates consistent with pulmonary edema
0	>300	Absent
1	>300	Present
2	200-300	Present
3	<200	Present

Adapted from Reference(2)

PGD3 occurring at any time point after reperfusion is associated with greater alteration in plasma markers of lung injury and higher 30-day mortality compared to those with lower grades.(1) However, even among patients with PGD3, there is variability in the duration of pulmonary edema and patient mortality,(13) leading to a belief that separate endotypes may be indicative of differences in injury response. In 2015, the ISHLT began a second consensus effort to update and refine the PGD definition; therefore, updated criteria should be forthcoming addressing issues including timing of PGD onset and resolution, precision of the P/F at different levels of FiO₂, single versus bilateral transplants, and differences in grading fibrotic lung disease patients.

Epidemiology and Outcomes

Since 2005, the incidence of PGD3 at a single time point has been reported to be 7.9% and 25%, whereas the incidence of PGD3 at any time point during the first three days is approximately 30%.(1, 5, 7, 14-17) The variability is likely attributable to differences in institutional practices and risk factor distributions.

PGD is associated with significant early and late post-transplant morbidity. Regardless of the time point of grading, PGD3 has been associated with significantly longer hospital length of stay, duration of mechanical ventilation, and 90-day mortality than those with lower grades of PGD.(1, 4, 5, 14, 16, 17) PGD3 present at 48 or 72 hours after transplantation was associated with an 18% increased absolute risk of 90-day mortality(14), with increased mortality persisting at 1, 5, and 10 years after transplantation.(5, 7, 14-17) In addition, all PGD is associated with elevated risk of bronchiolitis obliterans syndrome (BOS), a form of chronic allograft dysfunction.(5, 7, 15, 16) Future mechanistic trials are needed to understand the relationship between PGD and the development of BOS.

Pathogenesis

PGD represents the end result of multiple deleterious mechanisms provoked by donor brain death, mechanical ventilation, procurement, storage, and ischemia reperfusion injury (IRI). IRI refers to sterile inflammation that occurs after substrate supply is restored following a period of absent blood flow.(18) While many mechanisms are at play, recent studies highlight the importance of the interplay between innate immune activation, epithelial cell injury, endothelial cell dysfunction, and cytokine release(19) (Table 2).

Table 2.

Human translational studies exploring the pathogenesis of PGD

Pathway	Human translational studies	Reference
Epithelial cell injury	Higher blood and BAL levels of RAGE were associated with PGD	(20-22)
	Preexisting or de novo antibodies against type V collagen were associated with lower P/F ratio and PGD	(23, 24)
	Elevated levels of CC16 were associated with higher odds of PGD	(25)
	Higher donor levels of HMGB1 correlated with lower P/F before and after transplantation.	(26)
Endothelial dysfunction	Grade 3 PGD was associated with higher VEGF levels than lower grades of PGD	(27, 28)
	Increased pretransplant endothelin-1 levels were associated with PGD	(29)
	Angiopoietin-2 levels were elevated in PGD	(30)
Chemotaxis and cytokines	Higher plasma levels of MCP-1, IP-10, and IL-2R were associated with PGD	(31)
	Elevated serum IL-8 levels were associated with PGD	(32, 33)
	Higher blood and BAL levels of IL-6 were associated with grade 3 PGD	(34)
Innate immunity	Upregulation in the expression of genes involved in inflammasome and TLR pathways were found in BAL of patients with grade 3 PGD	(35)
	TLR4 mutations associated with innate immune hyporesponsiveness were associated with a lower PGD risk	(36)
	Higher plasma PTX3 levels were seen in those with grade 3 PGD	(37)
	Single nucleotide polymorphisms associated with higher levels of PTX3 were associated with grade 3 PGD	(38)
	Innate lymphoid cells were isolated from BAL of lung transplant recipients	(39)

Adapted from Reference(19)

BAL, bronchoalveolar lavage; CC16, plasma clara cell secretory protein; ET-1, endothelin-1; HMGB1, high-mobility group box-1; IP-10, interferon-inducible protein; MCP-1, monocyte chemotactic protein-1; P/F, PaO₂/FiO₂; PAI-1, plasminogen activator inhibitor-1; PGD, primary graft dysfunction; PTX3, long pentraxin-3; RAGE, receptor for advanced glycation end products; TLR, toll-like receptor; VEGF, vascular endothelial growth factor

Innate immune activation

Upregulation of innate immune pathways, including toll-like receptors (TLR) and NOD-like receptor (NLR) inflammasome signaling, has been demonstrated in vivo animal models following IRI and in human subjects with PGD.(35-38, 40-47) Animal models suggests that IRI occurs in two phases mediated in part by a bimodal response by the innate immune system with donor macrophages and lymphocytes inducing the early phase and recipient monocytes, T-cells, and neutrophils inducing the late phase.(40-47)

A family of innate immune cells, innate lymphoid cells (ILCs) have recently been recognized for their role in inflammation regulation, barrier protection and repair, and host defense. While ILCs have lymphoid morphology, they lack antigen receptors and can be subdivided based on cytokine expression and function.(39, 48, 49) Group 2 ILCs (ILC2) produce type 2 T-helper cell-associated cytokines and have been isolated in both human blood and lung, suggesting a trafficking ability. In animal models, ILC2 cells appear to play an important role in airway epithelial integrity and airway remodeling.(39) Isolation of ILC2 cells in BAL of transplant recipients warrants further exploration of the role of ILCs and other innate immune cells in PGD.(39)

Epithelial cell injury

Numerous markers of epithelial cell injury have also been associated with PGD in both animal and human studies, including receptor for advanced glycation end products (RAGE), type V collage, and plasma clara cell secretory protein. (20-25, 50-52) The role of traumatic brain injury (TBI) in the development of epithelial cell injury and IRI has been of increasing interest as the majority of lungs used for transplantation are procured from patients who have suffered TBI. High-mobility group box protein 1 (HMGB1) is a danger-associated molecular pattern released from necrotic neurons after TBI as well as alveolar macrophages after IRI. HMGB1 binds TLR4 and RAGE, inducing cytokine release and tissue damage.(26, 45) Animal models suggest that HMGB1-induced RAGE activation may contribute to IRI via activation of nuclear factor KB (NF-KB) in epithelial cells and production of IL-17 by natural killer T cells.(26, 45) Elevated levels of HMGB1 from brain-dead donors were associated with lower P/F before and after human lung transplantation.(26, 53) Therapeutic strategies to block RAGE and decrease epithelial cell injury should be explored in recipients of lungs from brain-dead donors post-TBI.

Endothelial dysfunction

Endothelial dysfunction in lung IRI typically results mostly from disruption and restoration of blood flow with minimal contribution from tissue hypoxia and reoxygenation since the lung does not rely on perfusion for oxygenation.(18) Changes in blood flow during ischemia lead to closure of inward rectifying potassium channels and subsequent endothelial cell depolarization.(54-56) This depolarization results in production of reactive oxygen species (ROS) and nitric oxide,(57-60) increasing cell adhesion molecules for leukocyte adhesion and extravasation and activating NF-KB and other transcription factors.(18, 59, 61) Subsequent reperfusion results in hyperpolization of endothelial cells and activation of a similar cascade as ischemia, the end result of which is further oxidative injury.(62) Future potential targets to modify endothelial activation and decrease PGD may include maintenance of lung perfusion with extracorporeal lung perfusion, administration of potassium channel agonists, and inhibition of ROS production.(18, 63)

In addition to inducing oxidative injury, disruption in blood flow to the lung may disrupt the endothelial cell barrier and induce vascular remodeling. Sphingosine 1-phosphate (S1P) controls endothelial cell tight junction formation and prevents chemotaxis. S1P supplementation prior to reperfusion decreased inflammatory cytokines and improved oxygenation in animal models.(64, 65) Several animal and human studies have demonstrated an association between IRI and mediators of vascular permeability and angiogenesis including vascular endothelial growth factor and angiopoietin-2.(27-30, 66-68) Further trials are needed to explore the role of IRI on endothelial integrity and vascular remodeling.

Chemotaxis and cytokines

In many solid organ models, IRI has been shown to stimulate the release of proinflammatory cytokines as well as chemokines involved in migration of recipient immune cells including IL-1β, IL-6, IL-8, IL-11, interferon (IFN)-gamma, and tumor necrosis factor (TNF) α.(44, 69-78) These findings have been confirmed in human lung observational studies and may represent future therapeutic targets.(31-34)

Clinical risk factors

Many groups have studied PGD risk factors (Table 3).(6, 11, 12, 14, 27, 79-94) Variability in reported risk factors may be attributed to single center studies, differences in peri-operative management between centers, and use of non-standardized PGD definitions; however, several donor, recipient, and procedural risk factors have consistently emerged.

Table 3.

Clinical risk factors for PGD

Category	Risk factors for Primary Graft Dysfunction
Donor inherent variables	Age>45 or <21yo
	African American race
	Female sex
	Smoke exposure
Donor acquired variables	Prolonged mechanical ventilation
	Aspiration
	Head trauma
	Hemodynamic instability after brain death
Recipient variables	Female gender
	Body mass index>25
	Pulmonary arterial hypertension
	Pulmonary hypertension secondary to parenchymal lung disease
	Idiopathic pulmonary fibrosis
	Sarcoidosis
	Elevated pulmonary arterial pressure at time of surgery
Operative variables	Single lung transplantation
	Prolonged ischemic time
	Use of cardiopulmonary bypass
	Packed red blood transfusion > 1L
	High reperfusion FiO2 > 0.4
	Use of Euro-Collins preservation solution

Adapted from References (6, 11, 12, 14, 32, 79-83, 87, 89, 90, 93-100)

Donor-related clinical risk factors

Reported donor risk factors include age, African American race, female gender, and history of tobacco exposure (Table 3).(6, 14, 80-83, 99) Smoke exposure is hypothesized to increase oxidative injury and thus IRI.(101) In donor lungs deemed unsuitable for transplantation, pulmonary edema and IL-8 were higher in donors who currently smoked although alveolar fluid clearance was lower in donors with greater than 20-pack year history compared to those with a less than 20-pack year history, suggesting a possible dose-related effect.(102) The impact of donor smoke exposure on PGD risk is difficult to ascertain as individual studies use different smoking cut-offs and smoking status is obtained from surrogates. Further studies are needed to clarify the role of donor smoke exposure on PGD. A recent study using registry data suggested that although donor smoke exposure was associated with worse recipient outcomes, the survival probability after receiving such an allograft still exceeded that of remaining on the waiting list.(103)

Donor-acquired risk factors, including prolonged mechanical ventilation, aspiration, and trauma, are potential risk factors without definitive studies demonstrating consistent associations with PGD.(81) Efforts are needed to standardize donor management and to optimize donor-recipient matching.

Recipient-related risk factors

Reported recipient risk factors for PGD include obesity, pulmonary hypertension, and diagnoses of idiopathic pulmonary fibrosis or sarcoidosis (Table 3). (6, 11, 14, 79, 80, 82, 89, 97, 98, 100) After adjustment for multiple risk factors, obese and overweight recipient body mass index (BMI) were independent predictors of PGD. Higher levels of leptin, which regulates adipose tissue mass and has inflammatory properties, were also associated with increased PGD risk and mortality.(97) This supports previous studies showing obesity increases the risk of acute lung injury.(104, 105) While increasing adiposity is associated with PGD, it is unclear how best to assess body composition given a recent study demonstrating that BMI is a poor measure of adioposity.(106) Following identification of a reliable measure of body composition, future studies should evaluate the role of adipokines in cytokine release and PGD.

Peri-operative risk factors

The operative risk factors for PGD reported in multi-center cohort studies are single lung transplant procedure type, prolonged ischemic time, cardiopulmonary bypass (CPB) use, greater than 1L packed red blood cell transfusion volume, use of Euro-Collins preservation solution, and reperfusion FiO₂ greater than 0.4 (Table 3).(12, 14, 82, 87, 89, 90, 95, 96) Oversized allografts have been associated with a decreased risk of PGD3 and improved survival in bilateral lung transplant recipients, especially in non-COPD recipients.(107, 108)

Significant operative variability exists across transplant centers making it difficult to identify consistent operative risk factors. Confounding by disease severity or treatment indication further complicates the study of operative risk factors. Future trials should distinguish between emergent and planned use of cardiopulmonary bypass, evaluate the effect of lower reperfusion FiO₂ on PGD, and explore the mechanisms by which the above operative factors increase PGD risk.

Prevention and treatment

Several agents have been investigated to prevent PGD.(109) While some observational trials of inhaled nitric oxide (iNO) suggest improved clinical outcomes, randomized control trials failed to show a definitive effect of iNO on PGD prevention when used routinely.(110-116) Soluble complement receptor 1 inhibitor, plasminogen activating factor antagonist, and exogenous surfactant demonstrated beneficial effects on surrogates of PGD including A-a gradient.(117-120) A trial of aprotinin failed to detect an effect on PGD risk, and was stopped early out of concern for potential renal toxicity.(121) Clearly, additional trials on PGD prevention are needed.

Therapy for PGD remains largely supportive and is heavily influenced by acute respiratory distress syndrome (ARDS) management strategies. While most centers use low-stretch ventilation, tidal volumes are frequently selected based on recipient characteristics with little consideration given to donor characteristics.(122-124) While there no proven role for iNO in the prevention of PGD, it has been used as salvage therapy for severe allograft dysfunction following transplant and may be useful in patients with refractory hypoxemia post-transplant.(125-127)

Veno-arterial (VA) extracorporeal membrane oxygenation (ECMO) has been studied for refractory hypoxemia and hemodynamic instability after lung transplant.(128-134) Given the link between CPB use and PGD, several groups have evaluated replacing CPB intraoperatively with VA-ECMO with mixed results.(135-137) VA-ECMO has been associated with shorter duration of mechanical ventilation and ICU/hospital length of stay, and lower transfusion requirements, but no statistically significant difference in 90-day mortality.(137) With improvement in ECMO technology, including high performance membranes and coated circuits, veno-venous (VV) ECMO has been increasingly used with similar post-transplant outcomes and survival as VA-ECMO.(131, 135) Several groups have evaluated the use of extracorporeal life support (ECLS) with VV, VA, and arterio-venous ECMO as a bridge to lung transplantation. Future research is needed to identify optimal candidates, standardize management strategies, and understand the impact of ECLS on PGD risk.(138-146)

Future directions

Several commercial devices using extracorporeal lung perfusion technology have been created and differ by pump type, flow type, mobility, perfusate, and presence of an internal ventilator. Extracorporeal lung perfusion has been associated with increased lung utilization, similar or lower risk of PGD, and equivalent 30-day, 1-year, and 3-year survival compared to standard criteria donors.(147-159) Small case series have described similar incidence of acute rejection and respiratory infection up to 1 year after transplant.(150, 153) Use of these technologies for donation after circulatory determination of death have demonstrated similar survival as donation after neurological determination of death.(151) A portable extracorporeal lung perfusion device was created to minimize ischemic time. While case reports of this portable extracorporeal perfusion system are promising,(160, 161) the results of randomized control trials are pending.(162, 163) Extracorporeal lung perfusion may eventually allow safe expansion of the donor pool and serve as a vehicle for testing targeted therapeutics to improve organ quality and decrease PGD risk. Several multicenter prospective trials are underway to further evaluate this technology, with PGD as an endpoint.

Conclusions

PGD is associated with poor short and long-term outcomes. The ISHLT definition has improved our ability to study this clinical syndrome, but further refinement of the definition is still needed and should be forthcoming. Improved understanding of risk factors and pathogenesis will identify potential therapeutic targets to modify the risk of PGD. As the acceptable recipient pool likely expands secondary to increasing ECLS use, continued safe expansion of the donor pool must follow and may be possible given the promising results seen with extracorporeal lung perfusion.

Key points.

• While the annual number of lung transplantations is increasing, primary graft dysfunction following lung transplantation remains a significant source of short- and long-term morbidity and mortality.

• The pathogenesis of primary graft dysfunction remains poorly understood, but is likely the end result of multiple deleterious mechanisms provoked by donor brain death, mechanical ventilation, procurement, storage, and ischemia reperfusion injury.

• Better understanding of the role of established clinical risk factors, as well as the emerging data on the importance of innate immunity, epithelial cell injury, endothelial cell dysfunction, and cytokine production in the development of primary graft dysfunction is necessary in order to identify potential high-risk populations and therapeutic targets.

• Future trials should focus on identifying appropriate patients to bridge to transplant with extracorporeal life support and expanding the donor pool with extracorporeal lung perfusion without increasing the risk of primary graft dysfunction.