Lung transplant outcomes: a review of survival, graft function, physiology, health-related quality of life and cost-effectiveness

Influence of pre-operative diagnosis

An important factor affecting survival outcomes is the wide variety of disease states for which LT is performed. Each of these diseases is accompanied by a host of factors that directly impact on both short- and long-term survival outcomes. Patients with chronic obstructive pulmonary disease (COPD) have the best short- and long-term survival, whilst patients with idiopathic pulmonary fibrosis (IPF) and pulmonary hypertension fare less well. Thus, when assessing survival statistics, it is important to do so in relation to the underlying lung disease. Also of great importance is the disparity in survival for different underlying disease states whilst patients are waiting for transplantation. In this regard, patients with IPF are the most likely to die whilst waiting, whereas patients with COPD stand the best chance of surviving the potentially long time on the waiting list. Thus, it is important to view post-transplant outcomes in the context of the natural history of the disease. Having a solid understanding of these issues helps not only in deciding on the proper timing for transplantation for the individual patient, but also has implications for donor-organ allocation when faced with patients with different diseases.

Influence of the type of surgical procedure

Another observation that possibly impacts on survival is the type of transplant performed. In the most recent analysis of data from the International Society For Heart and Lung Transplantation (ISHLT) registry, there appears to be a small survival advantage for patients who have received DLT compared with SLT procedures (fig. 1⇓) 3. In a further analysis of the international registry data, Meyer et al. 4 demonstrated that the survival advantage of DLT procedures extended up to age 60 yrs for recipients with COPD. Similar findings forpatients with COPD were reported by the Washington University group, St. Louis, MO, USA, in a retrospective review of their experience over a 13-yr period 5. Much speculation exists about why DLT would confer a long-term survival advantage, but one study suggests a decreased rate of bronchiolitis obliterans syndrome (BOS) in patients who have undergone DLT 6. However, a problem with interpreting this finding, as highlighted by Hadjiliadis et al. 6, is the fact that development of OB in the transplanted single lung will probably lead to a larger decrease in forced expiratory volume in one second (FEV₁; less pulmonary functional reserve), and also because native lung hyperinflation may impinge on allograft function. In any event, the potential advantage with DLT may be overshadowed by the potential benefit of helping twice as many patients with SLT. This issue is still being debated, in particular with relation to specific diseases. Aside from septic lung diseases that absolutely warrant DLT, debate still exists for primary pulmonary hypertension as to whether SLT, bilateral lung transplantation (BLT) or HLT is most appropriate. Similarly, in Eisenmenger's syndrome secondary to a simple septal defect, debate exists as to whether HLT or BLT/SLT with repair of the defect is optimal. The ISHLT registry data 3 show no difference in outcome between SLT and BLT for primary pulmonary hypertension, and there have been no direct comparisons of LT versus HLT in this group. Interestingly, for patients with Eisenmenger's syndrome associated with a ventricular septal defect, the ISHLT/United Network for Organ Sharing (UNOS) registry data 7 show better survival with HLT compared with BLT with closure of the defect. However, there are a number of papers supporting a role for BLT and defect closure to maximise the use of available donor organs 8–10.

Survival benefit

Although the success of LT can be assessed by several different outcome measures, survival has traditionally been the most important benchmark against which transplant procedures have been judged. Survival results for LT and HLT are tracked by the ISHLT/UNOS registry database. These data, compiled from a very large cohort of patients recorded over a decade, provide physicians with the necessary information to consider a potential recipient's “chances” following transplantation. The survival curve for LT, derived from accumulated data from 1990–2000, is shown in figure 2⇓ 3. The curve is characterised by an initial steep decline over the first year, followed by a slower but steady attrition over the ensuing years. Interestingly, the more recent experience demonstrates improvement in overall survival (fig. 3⇓). However, this improvement is solely as a result of reduced mortality during the peri-operative period, with little change in the slope of the long-term survival curve. Early results were affected by a steep learning experience, particularly with regard to surgical technique, peri-operative management, antimicrobial prophylaxis and recipient selection. Over time, not only have treatment regimens been changed and refined, but the characteristics of donors and recipients chosen have also evolved. As the difficulty in obtaining donor organs hasincreased, the limits of donor acceptability have been extended. In particular, donor age has increased over the years, and this may have significant implications for post-transplant survival.

Despite improvement in the peri-operative survival, over the long-term the slope of the survival curve continues to show the same slow attrition rate reflecting the problem of chronic rejection, for which there remains no effective therapy. This observation is disappointing in view of the ever broadening range of immunosuppressive agents available, aswell as improved understanding of pharmacokinetic and pharmacodynamic principles related to their use.

There are a number of major complications that frequently arise following LT that, unfortunately, commonly lead to significant morbidity and mortality (table 1⇓). In addition to these, a litany of other problems may also have significant impact on outcomes, including cardiovascular and gastrointestinal complications, as well as malignancy.

Data from the ISHLT registry show that infection is the leading cause of death in the first post-transplant year 3. Infection tends to be a pervasive problem that may occur as a primary or secondary process, following significant surgical complications (excessive bleeding, vascular and airway anastamotic complications, etc.) and primary graft dysfunction/severe ischaemia-reperfusion injury. Recipients with certain diseases pose a much greater risk for post-operative complications than others. For instance, patients with cystic fibrosis are at higher risk for peri-operative infection as a result of pre-transplant chronic sino-pulmonary infections. Complex congenital heart disease patients with multiple previous corrective procedures are at high risk of bleeding during the transplant procedure 8, 11. Patients with severe pulmonary hypertension treated with SLT are at an increased risk of severe ischaemia-reperfusion injury and may require prolonged post-operative mechanical ventilation 10. Similarly, recipients of older donor lungs (>50 yrs) that incur longer ischaemic times are also at increased risk of early and late graft dysfunction 3, 13. Despite this, in the current era, it isreasonable to expect 1-yr survival figures of 80–90% in uncomplicated recipients receiving “good” donor organs.

Beyond the first year, infection remains a major cause of morbidity and death. Several factors lead to the ongoing risk for infection, including the need for lifelong immunosuppression, abnormal airway clearance mechanisms, exposure of the allograft to the external environment and, therefore, to a host of potential pathogens, as well as OB. OB, the histological hallmark of chronic allograft rejection, and its clinical correlate BOS, occurs in ≤70% of recipients by 5 yrs after LT 14. This is the leading cause of death in the long term (>1 yr). Once OB is established, infection becomes the common final pathway leading to death in many of these patients.

As noted previously, the overall outcomes of LT must be assessed relating post-transplant outcomes to prospects without transplantation (so called “waiting list” mortality). Patients with IPF, cystic fibrosis and primary pulmonary hypertension have a higher mortality on the waiting list than patients with uncomplicated COPD or congenital heart disease. With a clear understanding of waiting-list mortality, it is feasible to develop models to assess the survival benefit for LT, and, indeed, these models may be important in determining appropriate organ-allocation systems. For instance, in a European database study 15, patients with pulmonary fibrosis and cystic fibrosis experienced a benefit from transplantation (compared with no transplant) at 55 and 90 days, respectively, whereas those with emphysema benefited by day 260. However, conflicting data were derived from the ISHLT registry 16, where a survival benefit could not be established for COPD, stimulating some debate about theutility of LT for this disease. The lack of agreement presumably relates to characteristics of the study populations, as well as possibly to the retrospective study designs. Thus, post-transplant outcomes must be interpreted in the light of prognosis without transplantation so that patients can be counselled appropriately, as well as for deriving a reasonable approach for allocating organs in the most judicious fashion.

Causes of mortality

Causes of morbidity

Pulmonary denervation

The surgical techniques of LT involve complete transection of the pulmonary autonomic nerves at the level of the distal trachea in the case of HLT, and at the level of the major bronchi in isolated and sequential BLT. Whether pulmonary reinnervation occurs over time in humans has not been directly studied. However, it would be anticipated that complete reinnervation would be unlikely in the context of human allotransplantation since no attempt is made to approximate nerve sheaths, which may vary substantially in anatomic position in donor and recipient structures. Indirect evidence for the persistence of denervation in human LT recipients has been derived from the observation of bronchial hyperresponsiveness to inhaled methacholine in these patients 23–27. This has been hypothesised to be the result of denervation hypersensitivity of muscarinic receptors. Furthermore, HLT recipients studied for 36 months post-operatively show a markedly decreased cough response to inhalation of nebulised distilled water, despite preservation of the laryngeal cough response, providing further evidence for the persistence of pulmonary denervation 28.

Control of breathing

Resting minute ventilation, tidal volume, respiratory frequency and mean inspiratory flow rate are normal after HLT 29, 30. Furthermore, HLT recipients have normal ventilatory drive during quiet breathing 31, 32. These findings suggest that pulmonary neurogenic mechanisms play little, if any, role in the regulation of resting ventilation.

During sleep, Shea et al. 30 found the level, pattern and variability of breathing during sleep in HLT recipients to be similar to normal subjects. Furthermore, there was no evidence for sleep-disordered breathing or nocturnal desaturation in the transplant recipients. In contrast, Sanders et al. 33 reported that HLT recipients tend to have a more rapid breathing frequency than normal subjects during sleep. However, to what extent the presence of restrictive ventilatory defects could account for these findings could not be ascertained in this latter study.

The preservation of an essentially normal level and pattern of ventilation in the presence of chronic pulmonary denervation suggests that pulmonary neural mechanisms are relatively unimportant in the normal regulation of these activities during relaxed wakefulness and sleep.

Although the level of ventilation at rest is normal in healthy HLT recipients, there is evidence for hyperventilation during exercise 32, 34, 35. The explanation for exercise hyperventilation is unknown; however, it may reflect an inadequate cardiocirculatory response to exercise associated with low mixed venous oxygen saturation and high mixed venous partial pressure of carbon dioxide tension (P_CO2) 34.

Gas exchange

Regional pulmonary blood flow normally decreases in response to local alveolar hypoxia, a phenomenon known as hypoxic pulmonary vasoconstriction. The precise mechanisms regulating hypoxic pulmonary vasoconstriction are uncertain. However, two possibilities exist. The first is that hypoxic pulmonary vasoconstriction is neurally regulated via the autonomic nervous system. Alternatively, hypoxic pulmonary vasoconstriction could result from local factors responsible for vasoregulation, such as chemical mediators. As such, pulmonary denervation resulting from LT represents an interesting model with which to evaluate the relative importance of these two postulated mechanisms.

Following human HLT, Robin et al. 36 observed intact hypoxic pulmonary vasoconstriction, and concluded that neural mediation, although perhaps important for modulating its intensity, is not essential for the phenomenon of hypoxic pulmonary vasoconstriction.

Several studies have reported normal levels of dead space ventilation at rest and during exercise after successful HLT 30, 37. This is not surprising, since the distribution of both ventilation and perfusion have been shown to be normal in HLT recipients with pulmonary denervation, at least in the resting state 38. After SLT for obstructive lung disease, ∼80% of both the ventilation and the perfusion are distributed to the transplanted lung 39. Although a similar pattern of perfusion distribution is seen in restrictive SLT recipients, the ventilation to the transplanted lung in these latter patients is only 60–70% of the total 39, 40. In the presence of acute rejection of a single lung graft, blood flow may be redistributed away from the transplanted lung 40, which can result in substantial ventilation-perfusion mismatching and, importantly, hypoxaemia, especially in patients with underlying pulmonary vascular disorders 41.

Within 8 weeks after uncomplicated HLT, gas exchange tends to be normal 37. Diffusing capacity for carbon monoxide (D_L,CO) has been variably reported as being normal 32, 42 or mildly reduced 43, 44 in HLT recipients. Recipients have normal oxygenation 3 months after DLT 45. After SLT, oxygenation improves substantially from pre-operative values 45; however, a mildly elevated alveolar-arterial oxygen difference in these latter patients is probably the result of the contribution to gas exchange provided by the remaining diseased native lung.

Oxygenation remains normal in HLT recipients during peak incremental 32, 46 and steady-state 37 exercise. Furthermore, the change in D_L,CO with increasing levels of exercise in these patients is similar to that observed in normal subjects 44. Oxygen saturation with peak incremental exercise is also normal in DLT recipients, as well as in SLT recipients with underlying obstructive lung disease 46–48. However, mild exercise desaturation may be seen in SLT recipients with restrictive parenchymal or pulmonary vascular disorders 39, 46.

Airway function

Following HLT, Glanville et al. 24 reported a combination of nonspecific bronchial hyperresponsiveness, absence of bronchoconstriction with respiratory manoeuvres and the absence of bronchodilatation with deep inspiration following induced bronchoconstriction, and postulated that the results were consistent with a lack of normal pulmonary innervation. Other groups have also reported evidence for moderate bronchial hyperresponsiveness to methacholine 23, 26, 27, 49 and histamine 26 challenge after HLT and LT. This hyperresponsiveness has been observed in the absence of substantial transbronchial biopsy evidence for airway inflammation 26, 27, and has been found to be stable over time 25. These patients do not exhibit exercise-induced bronchoconstriction 50, and do not generally present with a clinical syndrome suggestive of asthma. Several investigators have proposed that enhanced bronchial responsiveness to methacholine is a result of denervation hypersensitivity of airway smooth muscle muscarinic receptors 25, 27, 49. In contrast, Herve et al. 51 did not find increased nonspecific bronchial responsiveness to methacholine or isocapnic dry air hyperventilation in HLT and DLT recipients with normal lung histology. The explanation for these differing findings is not apparent. Although the presence or development of bronchial hyperresponsiveness to a variety of nonspecific stimuli (methacholine, hypotonic saline, hyperventilation, exercise) has been reported by a number of centres to be a predictor for the subsequent development of OB in LT recipients, this finding has not been universal nor of sufficient predictive value to be clinically useful as a screening test for this condition 23, 25, 42, 50–52.

Ventilatory sensation

Several lines of investigation have suggested that pulmonary afferent neural information can play a direct role in the mediation of dyspnoea 53–56. Alternatively, vagal mechanisms may contribute to dyspnoea indirectly, by modulating the pattern of ventilation and, hence, the level of central respiratory output and pattern of respiratory muscle activation 57. Pulmonary denervation, an inevitable outcome with LT, provides a unique opportunity to study the role of pulmonary autonomic nerves in the mediation of dyspnoea in humans.

Banner et al. 23 reported that HLT recipients undergoing methacholine bronchial provocation testing experienced a “tight” sensation in the chest, associated with the feeling that it was difficult to take a deep breath, similar to the sensations commonly reported by normal subjects undergoing bronchial challenge testing. The investigators concluded that pulmonary innervation is not essential for this sensation. However, they did point out that their findings do not rule out a possible important role for receptors in the remaining native trachea above the level of the anastomosis. The sensation of breathlessness during exercise 32, 58, as well as detection of inspiratory resistive loads 59, is also normal following HLT. These observations indicate that pulmonary neurogenic mechanisms do not appear to contribute significantly to the perception of breathlessness in normal humans.

Pulmonary function

At the time of donor identification, much attention is devoted to assuring that the physical size match between donor and recipient thoraces is fairly close. However, as a result of limited donor availability, precise donor-to-recipient size matching is not always feasible. Fortunately, the thorax has remarkable plasticity, and the rib cage and diaphragm are able to easily compensate for smaller donor organs. Transplant teams have, in general, tried to avoid utilising oversized organs, since post-operative compressive atelectasis with poor ventilation in a patient with already impaired host defenses could result in localised infectious complications. In the situation of substantial size discrepancies between donor and recipient, several centres have had success with pneumoreduction procedures to the donor lungs at the time of transplant 60.

Following LT, after initial falls, post-operative values for total lung capacity (TLC) approach the recipient's predicted value at 1 yr post-transplant, irrespective of the size of the donor lungs, suggesting that characteristics of the chest wall are the major determinants of post-operative lung volume, and not the donor lung size or compliance 61, 62. The stable pulmonary function observed after HLT has been reported to be consistent with a mild restrictive ventilatory pattern 43. Lung allograft compliance has been found to be normal, with a low TLC best correlated with maximum inspiratory mouth pressure, suggesting that the restriction early post-transplant could be related to respiratory muscle weakness (from malnutrition, deconditioning or steroid therapy) or phrenic nerve dysfunction 42. Lung volumes after SLT will depend, to an extent, on the physiological properties of the remaining native lung. In SLT recipients with pulmonary fibrosis, vital capacity increased from a mean pre-operative value of 43% predicted to 69% predicted at 1 yr 40. Virtually all of the improvement was seen at 3 months post-operatively, after which, plateaus were seen in most patients. In the case of SLT for obstructive lung diseases, mild elevations in TLC persist, albeit substantially reduced from pre-operative values 63 and FEV₁ is expected to rise to 50–57% predicted 64–67. Levine et al. 68 reported no substantial functional advantage to performing a right, as compared with a left, SLT in patients with obstructive lung disease.

After DLT, the original pathological process has little impact on post-operative lung function. Between 6–9 months post-transplant, FEV₁, forced vital capacity (FVC) and TLC are generally in the low normal range 45, 63 with an average FEV₁ of 78–85% and FVC of 66–92% predicted 65, 67, 69.

Exercise performance

Following LT, there are substantial improvements in exercise capacity. Within the first 3 months post-transplant, a 60–75% increase in the mean distance walked in 6 min has been reported 45, with a subsequent plateau in performance. Peak oxygen consumption (V'_O2) values in the range of 40–60% predicted have been reported in the first year after HLT, SLT and DLT 39, 46–48, 70. Maximum symptom-limited incremental cardiopulmonary exercise following uncomplicated SLT or DLT is characterised by: 1) a low maximum V'_O2; 2) an absence of abnormal gas exchange, ventilatory or cardiac limiting factors; and 3) an early onset of the anaerobic threshold 46. The majority of LT recipients report leg fatigue, rather than dyspnoea, as the reason for terminating maximal exercise, with leg work capacity being a strong predictor of peak exercise performance 71. Evidence points increasingly to the role of skeletal muscle dysfunction as the most important factor resulting in reduced maximal exercise capacity following LT.

Oxygen saturation generally remains >90% in exercising LT recipients 47, 48, and ventilatory function does not appear to limit exercise in most HLT, SLT and DLT recipients in the absence of complications 72.

Following LT, heart rate increases as predicted during incremental exercise and reaches an adequate level (60–70% of age-predicted maximum) to meet the demands of the final workload of the exercise test 47, 48. Cardiac output is generally sufficient to maintain the workloads reached and, thus, cannot explain the early onset of anaerobic threshold and reduced peak V'_O2 following LT. However, peripheral circulatory parameters may play a role in exercise limitation.

Similar to deconditioned individuals, LT recipients typically show an early onset of the anaerobic threshold at submaximal levels of V'_O2 46. This may indicate either a limitation of delivery of oxygen to the periphery (i.e. a cardiac and/or vascular limitation), or abnormal uptake and use of oxygen at the level of the muscle. As a result, there is greater dependence on anaerobic metabolism for energy, greater production of metabolic byproducts, such as lactate and a decreased ability to buffer and remove these byproducts.

Impaired oxidative capacity of skeletal muscle of LT recipients has been documented in a number of studies, and may be an important factor limiting exercise in these individuals. Skeletal muscle dysfunction may be due to factors related to the pre-transplant condition, severe deconditioning or, alternatively, to immunosuppressant medications 73.

Wang et al. 74 demonstrated that the skeletal muscle of patients following LT had lower proportions of type-I (oxidative) fibres, as well as reduced oxidative enzyme concentrations. This reduction in proportion of type-I fibres and oxidative enzymes is also present pre-transplant and remains low following the procedure, at least in the short term (3 months) 75. These findings are in line with the consistent observation of reduced maximum V'_O2 and early anaerobic threshold in lung transplant recipients, and support a possible role for chronic, possibly irreversible, deconditioning as a contributing cause of the impaired oxidative capacity seen in LT recipients.

In keeping with the histological and biochemical changes seen in the skeletal muscle of LT recipients, skeletal muscle metabolism during exercise has also been found to be abnormal. Evans et al. 70 reported lower resting intracellular pH of the quadriceps muscle and greater exercise increase in lactate concentration with incremental leg exercise in these individuals. This is associated with a shorter endurance time and lower peak V'_O2. These findings indicate a greater reliance on anaerobic metabolism, possibly a result of poor oxygen uptake and/or utilisation by the muscle. Tirdel et al. 76 found evidence for a defect in the ability of the skeletal muscle to take up and utilise oxygen rather than a problem with oxygen delivery in LT recipients, a finding similar to that seen in patients with metabolic myopathies. This suggested that LT recipients have a defect at the level ofthe mitochondria, which results in a reduced ability for working muscle to extract oxygen. Reduced capacity of the quadriceps to extract oxygen has also been reported in patients with cystic fibrosis 77 and COPD 78 pre- and post-transplant, indicating that muscle dysfunction antedates thetransplant and is not completely alleviated in the post-transplant period, once again highlighting the possible role of chronic deconditioning on impaired exercise performance post-transplant.

An additional factor to consider is that LT recipients chronically take a number of immunosuppressive medications that may also impact on muscle function. Cyclosporin, a calcineurin inhibitor, affects the ability of capillaries supplying skeletal muscle fibres to adequately dilate, thereby reducing oxygen delivery to working muscle 70, 79, which could impair the oxidative capacity of exercising muscle. Mitochondrial function may also be impaired in transplant recipients due to a cyclosporin effect on mitochondrial respiration 80, 81. Corticosteroids can also result in muscle fibre atrophy and abnormalities in muscle fibre shape and size 82; however, the morphological and biochemical changes associated with chronic corticosteroid use are not the same as those observed following LT, making this an unlikely culprit.

Of interest, recipients of heart 83, kidney 84 and liver transplants 85 have been found to have exercise profiles similar to LT recipients (reduced maximum V'_O2, early anaerobic threshold, absence of abnormal ventilatory or cardiac limitations), suggesting a common factor (e.g. immunosuppressive medications) may be responsible for the aerobic impairment.

In patients with COPD, exercise training can result in improvement in the oxidative capacity of the quadriceps muscle, which is associated with a delay in the onset of lactate acidosis and an increase in peak V'_O2 86–88. Hence, one could anticipate that adaptation to training at the level of skeletal muscle should be possible following transplantation with resultant improvements in exercise capacity. Few studies have examined the effects of exercise training following LT. Aerobic exercise training has been shown to result in modest improvements in peak exercise performance in LT 89 and HLT recipients 90, although the subjects remained with significant aerobic impairment post-training. The implications for impaired exercise performance following LT remain unclear, since recipients unquestionably experience substantial improvements in functional capacity and QoL 3. Nevertheless, identification of factors contributing to exercise limitation is important in order to optimise rehabilitative strategies for these patients.

In conclusion, although LT recipients demonstrate substantial improvements in exercise capacity compared to their pre-transplant condition, significant impairment in exercise capacity persists, largely related to reduced capacity for skeletal muscle oxidative metabolism. Severe chronic deconditioning, as well as drug effects, are probably important contributing factors. Exercise training may provide a means by which skeletal muscle function can be improved in the post-transplant period.