Pulmonary arterial hypertension is a disease characterized by progressive resistance to flow across the pulmonary vascular bed, resulting in death from progressive right heart failure. The evolution of the currently approved therapies for certain pulmonary arterial hypertension functional classes has resulted in a considerable improvement in 1-year survival. However, substantial risk remains. For those patients judged to be deteriorating on therapy, bilateral sequential lung transplant or, occasionally, heart-lung transplant is the preferred option. Review of the literature suggests many relatively simple, routinely acquired clinical and hemodynamic indices that appear to predict risk of mortality and hence need for transplant. The current LAS, in sharp contrast, substantially disenfranchises patients with pulmonary arterial hypertension. It is hoped that validation of some or all of these risk factors, as is being undertaken in the REVEAL database, may permit development of a reliable, simple,and widely applicable tool for assessing mortality risk and hence need for transplant referral in our PAH patients.
The therapy of pulmonary hypertension (PH) has changed substantially in the last 10 to 15 years. Survival of patients with idiopathic PH as described in the National Prospective Registry,1 which accrued data from 1981 to 1985, was dismal with an estimated 1-, 2-, and 3-year survival of only 68%, 48%, and 34% respectively. At that time, patients failing to demonstrate a sustained response to therapy with calcium channel blockers (CCBs) had no recourse except lung transplantation. Patients with what was then referred to as primary PH represented approximately 10% of patients awaiting lung transplantation nationwide in 1996.2
With the advent of the first commercially available therapy for pulmonary arterial hypertension (PAH), epoprostenol, a substantial survival benefit was seen with 1-, 2-, and 3-year survival in idiopathic PH patients in functional class III and IV reported as 88%, 76%, and 63% respectively (Figure 1).3 The evolution of the currently approved therapies for World Health Organization (WHO) diagnostic group I PAH functional classes II-IV (prostacyclins, endothelin receptor antagonists, and phosphodiesterase inhibitors) appears to have resulted in a substantial improvement in 1-year survival (variously reported as between 92%4 and 86%5). Three- and 5-year survival, however, remains substantially less (69% and 61% respectively), demonstrating the ongoing need for novel and more effective therapies. For individuals failing therapy, lung or heart-lung transplantation currently remains an important option.
3-year survival observed in functional class III and IV patients started on epoprostenol compared to survival predicted by the NIH equation using baseline hemodynamics, P<0.001 at 1, 2, and 3 years. McLaughlin VV et al. Survival in primary pulmonary hypertension: the impact of epoprostenol therapy. Circulation. 2002;106:1477-1482. Reprinted with permission from Wolters Kluwer Health.
Historically, many PAH-treating physicians used the initiation of an intravenous prostacyclin analog as the cue to refer for transplant; however, clearly patients' response to such therapy varies. The current lung allocation scoring (LAS) system, designed by the federal government's Organ Procurement and Transplantation Network (OPTN), no longer prioritizes patients based on duration of time on the waiting list, but rather attempts to allocate scarce donor organ resources based on a balance of need (severity of illness) and presumed benefit (likelihood of survival post transplant). The initiation of a prostacyclin intravenously is generally considered to represent the strongest therapy with the greatest likelihood of success—and the ultimate solution. It may herald the beginning of a prolonged period of improvement and/or stability, rather than one of deterioration. Therefore, more reliable indices of deterioration and the need for imminent transplant are required.
Given the limitations of transplantation—which itself has a median 5-year survival of only about 50%—it becomes important for the treating clinician and the transplant physician to be able to recognize those features in a patient with PAH that predict worse outcome and the need for expeditious referral and transplantation.
The original National Prospective Registry evaluated only patients with idiopathic pulmonary arterial hypertension (IPAH) in the pre-PAH treatment era, and assessed numerous clinical factors as potential predictors of outcome. Hemodynamic values (cardiac index [CI], mean right atrial pressure [RAP], and mean pulmonary artery pressure [PAP]), functional class, presence or absence of Raynaud's phenomenon, and DLCO (diffusion capacity for carbon monoxide) appeared on univariable analysis as important predictors of survival. However, only the hemodynamic variables were reliable by multivariate analysis. Signals derived from this original small (n=194) registry restricted to IPAH patients have served as the template for evaluating potential risk predictors in the current therapeutic era.
Since the first reported case in the world literature of PH responding to CCB administration,6 extensive literature about vasoreactivity testing to safely and reliably identify those patients who would benefit from therapy with CCBs has been published. A positive vasodilator test followed by therapy with CCBs has been repeatedly demonstrated to identify a population of patients with improved survival.7–9 Beneficial response to vasoreactivity testing therefore clearly identifies a small patient population likely to respond for an extended period of time to a relatively simple and inexpensive therapy. Reassessment of the patients on such therapy is mandatory; however, prolonged benefit is not guaranteed. Recognition of evolving clinical features that will predict mortality is as critical in these patients as in the bulk of “nonvasoreactive” PAH patients.
Germline mutations in the bone morphogenetic protein receptor gene (BMPR2) have been detected in up to 40% of sporadic cases of PAH and in 70% of cases of familial PAH. The influence of BMPR2 mutations on presentation and outcome was reviewed in 223 patients from the French National Registry.10 Mutation carriers tended to present younger, have more severe hemodynamic abnormalities at presentation (higher mean PAP; lower CI, lower mixed venous oxygen saturation [MVO2], and higher pulmonary vascular resistance [PVR]), and were less likely to have an acute vasodilator response than noncarriers. Mutation carriers also progressed to death or transplant more rapidly and at a younger age than in noncarriers (age at death 34.4+15; vs 50.5+17.5, P=0.003). Recognized mutations in BMPR2 potentially identify a patient population at increased risk, warranting careful follow up and earlier referral.
Pulmonary arterial hypertension is a disease characterized by progressive resistance to flow across the pulmonary vascular bed, resulting in death from progressive right heart failure. It would seem reasonable, therefore, that hemodynamic criteria—PAP, RAP, CI, PVR index, and MVO2, measured directly through right heart catheterization or less accurately but indirectly through echocardiography—would reliably indicate progressive risk for mortality.
Consistent with this in the National Institutes of Health (NIH) registry, patients with lower CI (<2 L vs >4 L), higher mean PAP (>85 mm Hg vs <55 mm Hg), and higher RAP (>20 mm Hg vs <10 mm Hg) had substantially worse survival—contributing to the generation of the NIH survival prediction equation. However, in a review of hemodynamic predictors of outcome in 178 functional class III and IV patients treated with epoprostenol,14 only a RAP of >12 mm Hg independently predicted poor survival. If pretreatment and 3-month data were evaluated, failure of total pulmonary resistance (TPR) to fall by 30% relative to baseline was associated with a hazard ratio (HR) for death of 1.9 (P=0.041). In univariate analysis in the same cohort, paradoxically a mean PAP <65 mm Hg was significantly related to an increased risk of death—possibly on the basis of a failing ventricle and a falling CI.
While isolated hemodynamic variables may not reliably predict risk of death, it is clear that failure to respond to therapy with improved CI, reduction in RAP, and decrease in TPR are probably reasonable and appropriate indicators of high mortality risk and the need for referral and transplantation.
A recent review of data from the Pulmonary Hypertension Connection, a consortium of 3 university centers in Chicago, evaluated hemodynamic predictors of mortality in 576 PAH patients over the age of 18 at time of referral to the center. The intent was to reappraise the original NIH survival equation in the current era. An analysis of the subgroup of patients with IPAH, familial- or anorexigen-associated PAH (282 patients), was also performed. Based on this retrospective review, the following hemodynamic predictors of risk for mortality were identified: increased mean RAP (for every 5 mm Hg increase, HR 1.29); increased PVR (for every 5 Wood unit increase, HR 1.11); and decreased CI (for every increase in CI by 1 L/min/m2 a reduction of 25% in mortality risk) (Table 1).
Risk Factors for Mortality and/or Transplant
|Less Risk||Δcontinuous variable||More Risk|
|CI||>4||Each 1 L decrease-25% increase||≤2L|
|mRAP||≤10||≥12; each 5 mm increase; HR increase = 1.29||≥20|
|TPR with Rx||30% decrease|
|PVR||Each 5 WU increase = HR increase of 1.11|
|FC||I and II||III and IV|
|HR post Rx||≥91|
|DLCO||>64% predicted||Each 10% decrease in DLCO mortality increased||<43% predicted|
|NT ProBNP||≥50% reduction with Rx||>1400|
|Albumin||low||High to normal|
No studies have validated echocardiographic changes in the context of survival. Several studies have demonstrated that beneficial response to therapy is associated with improvement in echocardiographic evaluation of hemodynamic parameters. A recent prospective single-center study in France of PH patients11 (excluding those with congenital heart disease and patients with post capillary PH) for the first time correlated echocardiographic measurements with survival. A total of 79 patients followed for a median of 12 months had diagnostic right heart catheterization an echocardiographic study, 6-minute walk test (6MW) and functional class stratification. This review included patients with parenchymal lung disease and pulmonary embolic disease. The authors conclude that some previously recognized echocardiographic parameters were valuable predictors of mortality:
Tei index >0.98 was associated with HR of 5.41 for death (P=0.035)
TAPSE as a continuous variable: HR for death increased by 16% for each 1 mm decrease (HR 0.84; P=0.024)
Presence of a pericardial effusion: HR for death 5.18, P=0.002.
They also, however, described some new prognostic factors predictive of mortality:
Mean PAP >49 (HR for death 3.94, P=0.012); as a continuous variable for each 1 mm increase in mean PAP increased HR for death by 4%; P=0.04
Diastolic PAP (dPAP) >29; HR for death 4.97, P=0.006; as a continuous variable, each 1 mm increase in dPAP increased HR for death by 6% (HR 1.06, P=0.01)
Abnormal end diastolic septal curve: HR for death 5.33; P=0.024
IVC diameter >20 mm with respiratory variation of diameter <50%; HR for death 3.39 (P=0.018)
While intriguing, these data remain to be validated prospectively.
Measurement of arterial blood gases (ABG) is a fairly routine part of the evaluation and diagnosis of patients with PAH. In patients in the NIH registry, “hypoxemia and hypocapnea were almost an invariable finding.”12 This was originally attributed to restrictive physiology; however, there is evidence for increased ventilatory drive caused by low cardiac output and reduced tissue oxygen delivery. A 12-year retrospective study on patients with IPAH13 looked at the impact of ABG at rest and with exercise alone and with other variables on survival. The baseline PaO2 at rest and the alveolar-arterial oxygen gradient (PA-aO2) had no significant impact on survival. In contrast, baseline (pre-PAH treatment) and 3-month post-treatment PaCO2 less than 32 mm Hg were correlated with significantly reduced survival (Figures 2 and3 and Figure 3, P>0.001). While serial measurements of PaCO2 were not undertaken in this study to assess change in response to therapy, this may be a useful indicator of patients at increased risk.
Impact of PaCO2 on survival measured at baseline and after 3 months of therapy in patients with idiopathic pulmonary hypertension. Hoeper MM et al., Eur Resp J. 2007;29:944-950. Reprinted with permission from the European Respiratory Society.
Impact of PaCO2 on survival measured at baseline and after 3 months of therapy in patients with idiopathic pulmonary hypertension. Hoeper MM et al., Eur Resp J. 2007;29:944-950. Reprinted with permission from the European Respiratory Society.
The 6MW is a submaximal exercise test that can be performed even by debilitated patients in varying degrees of heart failure. It has been useful in patients with advanced left heart failure and is a prognostic indicator for that disease. The 6MW has been used as a surrogate for functional improvement, and as an end point in many therapeutic trials in patients with PAH. It has been repeatedly shown to predict survival in PAH,12,14-15 though the absolute values discriminating high risk patients have varied between study populations (<332 m; <380 m). In contrast, recent data suggest that a subset of younger, thinner patients is able to walk more than 450 meters at baseline.16 In these individuals, outcome is more reliably predicted by their functional class and by hemodynamic variables than by their walk distance.
At the time of the original NIH registry, at diagnosis most patients were in NYHA functional class III or IV. Data from the earliest studies of epoprostenol in such patients clearly showed that the worse the functional class at initiation of therapy, the worse the prognosis was likely to be, and more importantly, in the current therapeutic era, failure to improve from functional class III or IV to I or II was associated with substantially worse outcome (Figures 4 and5).3,14 Worse functional class as a predictor of mortality was also validated as a predictor of mortality in multivariate analysis performed by the previously referenced PHC survival analysis. Deteriorating functional class or failure to improve from a high-risk functional class is therefore a reliable predictor of higher pretransplant mortality.
Impact of functional class on survival: prior to therapy (a) and following 3 months of epoprostenol therapy in patients with idiopathic pulmonary arterial hypertension (b). Reprinted from J Am Coll Cardiol, 40(4), Sitbon O et al, Long-term intravenous epoprostenol infusion in primary pulmonary hypertension: prognostic factors and survival, pp 780-788, copyright 2002, with permission from Elsevier.
Long-term survival based on functional class status at the time of epoprostenol therapy initiation. McLaughlin VV et al., Survival in primary pulmonary hypertension: the impact of epoprostenol therapy. Circulation. 2002;106:1477-1482. Reprinted with permission from Wolters Kluwer Health.
Resting heart rate and reduction in heart rate both predict likelihood of mortality due to left heart disease.17 To evaluate the prognostic importance of resting heart rate in patients with PAH, 140 patients with IPAH were evaluated with walk tests and EKGs as part of their routine follow up.18 In addition to the previously reported observation that patients with higher mortality were likely to have worse walk tests (mean 326±120 vs 404±132 P<0.001) and were more likely to be functional class IV (20/54 vs 10/86, P<0.01), nonsurvivors were much more likely to have elevations in baseline heart rate relative to survivors (90±17 vs 76±13, P<0.001). Further analysis of heart rate over time in these patients revealed that while resting heart rate >82 at baseline prior to treatment was associated with significant mortality risk, post-treatment elevations of heart rate persistently over 91-92 showed a similar and significant increase in mortality (Figures 6 and7).
Decreasing survival with higher resting heart rate prior to treatment in idiopathic pulmonary arterial hypertension. Reprinted from Am J Cardiol, 102(3), Henkens IR et al, Relation of resting heart rate to prognosis in patients with idiopathic pulmonary arterial hypertension, pp 1451-1456, copyright 2009, with permission from Elsevier.
Impact of higher resting heart rate on survival after 1 year of therapy in idiopathic pulmonary arterial hypertension. Reprinted from Am J Cardiol, 102(3), Henkens IR et al, Relation of resting heart rate to prognosis in patients with idiopathic pulmonary arterial hypertension, pp 1451-1456, copyright 2009, with permission from Elsevier.
A mildly restrictive pattern on routine pulmonary function test has been generally recognized in patients with PAH, presumably on the basis of cardiomegaly and tethering of the pulmonary parenchyma by inelastic pulmonary vessels. However, in PAH uncomplicated by parenchymal lung disease, only DLCO appears to be useful in assessing the patient's mortality risk. Patients in the NIH registry had reductions in DLCO relative to predicted normals. A recent review of 408 WHO diagnostic group I patients demonstrated that by multivariate analysis (controlling for age, connective tissue disease [present vs absent], functional class, oxygen use, evidence of parenchymal abnormalities on chest CT, spirometry, PVR, MVO2, and labs [creatinine, albumin, and hemoglobin]), that the lowest tertile for DLCO was associated with poor survival and that with every 10% decrease in DLCO mortality increased (Figure 8).19
Impact of DLCO at enrollment on survival. Reprinted from J Heart Lung Transplant, 29, Chandra et al, Carbon monoxide diffusing capacity and mortality in pulmonary arterial hypertension, pp 181-187, copyright 2010, with permission from Elsevier.
B-type natriuretic peptide (BNP) is a cardiac hormone produced and released from cardiac myocytes in response to pressure and volume overload induced ventricular stretch. BNP is released as a prohormone, and then cleaved into an active peptide and an inert N-terminal fragment. Thus, measurement of both the BNP and the inactive N-terminal proBNP are markers of BNP production. Higher (than median of 168 pmol/lL) levels of NT-proBNP have been associated with lower survival in a heterogeneous group of patients with precapillary PH.20 Absolute values of NT-proBNP greater than 1400 (usual ULN <125 pg/mL) may also predict PH with increased mortality.21 Perhaps more importantly, if patients demonstrate a decrease in BNP following treatment, this appears to correlate with improvement in other parameters (6MW, CI, mean PAP).22 A review of patients treated for PAH demonstrated that a decrease in BNP of >50% during the first 3 months of epoprostenol therapy was strongly predictive of event-free survival (P=0.003).23
Of particular note is the utility of NT-proBNP as a predictor of survival in patients with scleroderma. In addition to elevations in excess of 395 pg/mL being predictive of the diagnosis of PAH in this patient population, baseline and serial changes are highly predictive of survival with a 10-fold increase in NT-proBNP predicting a 3-fold increase in mortality.24 This suggests that marked elevation of absolute values and directional trends may be useful in risk stratification for augmented therapy and transplant referral.
Renal dysfunction (both elevated BUN and creatinine and proteinuria) is frequent in PAH. Renal dysfunction is a predictor of mortality and morbidity in non-PAH cardiovascular disease. Data generated from the PHC database (previously referenced) indicates that increased serum creatinine was predictive of worse functional class, decreased DLCO, worse exercise tolerance, and in those with a serum creatinine of >1.4 there was a HR for death of 2.54 times the rate of death in those with creatinine <1.0 mg/dl (P<0.0001).25 Worse renal function is a more useful predictor of increased mortality in those patients where RAP as a measure of hemodynamic function is relatively well preserved. So as patients' hemodynamics deteriorate, the discriminatory capacity of elevations in serum creatinine becomes less useful in predicting mortality.
Another simple serum analysis, measurement of serum albumin, was evaluated as a predictor of death or lung transplantation in a relatively small single-center study of 84 patients. This concluded that a higher serum albumin was associated with improved survival.26
In broad terms, patients with connective tissue disease in general and scleroderma in particular are recognized to have a much worse prognosis with PAH than the IPAH patient population, and as such warrant careful follow up and earlier referral.
Patients with congenital heart disease-associated PAH as a population have a better survival rate than IPAH and connective tissue disease patients. However, to date, no data have been published that would suggest that congenital heart disease patients with equivalent disease severity (as assessed by functional class, 6MW, DLCO reduction, RAP elevation, reduced CI, hypocapnia, etc) fare any better than IPAH patients. There is some suggestion that preservation of vasoreactivity27 (defined as a 20% drop in TPR in response to inhaled nitric oxide) in this patient population may predict a better outcome. However, the single study indicating this did not control for severity of PAP elevation or for baseline TPR.
Patients with pulmonary veno-occlusive disease (PVOD) or pulmonary capillary hemangiomatosis, which is thought to be a manifestation of progressive PVOD, have a particularly poor outcome. When PVOD is suspected based on BAL evidence of hemosiderin-laden granules in alveolar macrophages, and/or interstitial changes on high resolution CT scan, therapeutic options are substantially limited. The use of prostacyclin analogs is generally considered to be contraindicated and potentially life threatening, and little data other than anecdotal exist regarding the response to other PAH-specific therapies such as endothelin receptor antagonists, and phosphodiesterase 5 inhibitors. Suspicion of these diagnoses should generally precipitate immediate transplant referral.
As indicated previously, though substantial changes have occurred in survival in PAH patients in the current treatment era, substantial risk remains. The risks identified are presented in Table 1. To date, no weighting of the relative impact of these predictors of survival has been published to try to simplify risk assessment of these complex patients when being considered for transplantation. It is hoped that data from the REVEAL Registry will permit such an approach to relative risks of these predictors of mortality. For those patients judged to be deteriorating on therapy, bilateral sequential lung transplant or, occasionally, heart-lung transplant is the preferred option.
Since the implementation of the OPTN LAS, which bases organ allocation on disease severity, there has been an increase in the number of transplants being undertaken, and a substantial decrease in the overall time on the waiting list for all patients awaiting lung transplantation. However, an elegant review of the United Network for Organ Sharing database comparing frequency of death or transplant by diagnosis in the pre-LAS and post-LAS eras clearly shows that the pulmonary hypertensive patient population, though always less likely to be transplanted than any other group, are now substantially disadvantaged. Every other major diagnosis (cystic fibrosis, chronic obstructive pulmonary disease, pulmonary fibrosis) has a much greater frequency of transplant, while PH patients experience a much greater likelihood of death while waiting on the transplant list (Figure 9).28
Cumulative incidence curves comparing transplantation and death on the waiting list between diagnoses before and after implementation of the Lung Allocation Score. Idiopathic pulmonary arterial hypertension (solid line). Idiopathic pulmonary fibrosis (dashed line). Chronic obstructive pulmonary disease (gray line). Cystic fibrosis (dotted line). Reprinted with permission of the American Thoracic Society. Copyright American Thoracic Society. Chan K, 2009, Idiopathic pulmonary arterial hypertension and equity of donor lung allocation in the era of the lung allocation score: are we there yet? Am J Resp Crit Care Med, 180, pp 385-387.
This discrepancy reflects the fact that at the time of implementation of the current LAS, risk factors for mortality for most of the parenchymal lung diseases were recognized and included in the scoring system. In contrast, risk factors for mortality in PAH were ill defined or not validated. The LAS therefore includes factors that frequently do not accurately reflect the severity of illness of PAH patients: lung function, exercise capacity, and the need for oxygen or ventilatory support.
Baseline data from the REVEAL Registry have now been published.29 This huge database of some 3500 sequentially enrolled patients from 54 US centers is now being analyzed and the risk factors detailed previously, as well as others, have been evaluated as survival predictors in PAH. Preliminary data have been presented at national meetings proposing new appropriately weighted risk factors to be included in the LAS, which may result in a more equitable allocation of donor lungs to this disenfranchised patient population.