Diagnosis and Treatment of Pulmonary Arterial Hypertension Richard Channick, Mda,*, Timothy L

Cardiol Clin 22 (2004) 441–452

Diagnosis and treatment of pulmonary arterial hypertension Richard Channick, MDa,*, Timothy L. Williamson, MDb aDivision of Pulmonary and Critical Care, University of California, San Diego, 9300 Campus Point Drive, La Jolla, CA 92037, USA bDivision of Pulmonary and Critical Medicine Care Medicine, University of Kansas Hospital, 3901 Rainbow Boulevard, Kansas City, KS 66160, USA

Pulmonary arterial hypertension (PAH) is that has employed the term PPH, this article uses characterized by an elevated pulmonary vascular both terms interchangeably. In addition to IPAH, resistance (PVR) that, in the absence of interven- PAH can occur in association with other con- tion, progresses to right ventricular failure and ditions such as connective tissue disease, congen- death. When defined clinically as a mean pulmo- ital heart disease, portal hypertension, HIV, and nary artery pressure greater than or equal to 25 prior ingestion of stimulant drugs such as fenflur- mm Hg at rest or 30 mm Hg with exercise, amines, amphetamines, or cocaine. Of the various idiopathic PAH has been estimated to affect at categories, only PAH is discussed in this article. least 1 to 2 persons per million population [1], The evaluation of pulmonary hypertension, how- although secondary forms of the disease are much ever, should ascertain its proper cause and more common. classification, because many of the treatment The World Health Organization (WHO) clas- options for PAH discussed here are not necessar- sification of pulmonary hypertension (PH) is ily efficacious and may be deleterious in patients reviewed in detail in the article by McLaughlin with pulmonary hypertension secondary to other et al. This classification is based on expert disorders. consensus opinion reached at the WHO Sympo- Although long considered a disease of pulmo- sium on Pulmonary Hypertension in Evian, nary vasoconstriction, PAH should now more France in 1998. It divides PH into five categories, correctly be considered a disease of proliferation. including PAH, pulmonary venous hypertension, Voelkel and Cool further review the pathology of PH associated with respiratory diseases, PH this disorder elsewhere in this issue. Briefly, PAH associated with embolic disease, and PH caused is characterized by increased pulmonary arterial by diseases that directly affect the pulmonary thickness, dilated capillaries, and plexiform le- vasculature, such as sarcoidosis. Experts recently sions that obstruct the vessel lumen [2]. Addition- convened again in Venice, Italy in 2003 [1]. ally, patients are at risk for in situ thrombosis Among the modifications from the Venice meet- secondary to increased platelet aggregation as ing is that pulmonary hypertension without a consequence of decreased nitric oxide and a known association or cause is no longer re- prostacyclin production [3]. ferred to as primary pulmonary hypertension Survival in PAH without intervention is dismal. (PPH), but instead as idiopathic (sporadic) PAH A National Institutes of Health registry followed (IPAH) with a separate category for familial patients with PPH (idiopathic PAH) between 1985 PAH (FPAH). Because of the body of literature and 1988, in an era that preceded currently available therapeutic agents [4]. Survival rates at 1, 3, * Corresponding author. and 5 years were 68%, 48% and 34%, respectively, E-mail address: [email protected] (R. Channick). with a median survival from diagnosis of 2.8 years.

Findings correlated with poorer prognosis in- attention should be paid to excluding chronic cluded WHO functional class III or class IV thromboembolic disease, because a potentially (Box 1), presence of Raynaud’s phenomenon, corrective surgical procedure exists for this disor- increased right atrial pressure, increased pulmo- der [8,9]. nary artery pressure, decreased cardiac index, and The symptoms of PAH are nonspecific. Dysp- decreased diffusion [4]. In another study multivar- nea is nearly universal as the disease progresses, iate analysis suggested pericardial effusion, en- but fatigue, chest pain, syncope, and near syncope larged right atrium, 6-minute walk distance, and are also common [10]. Given the vague nature of mixed venous oxygen saturation as predictors of symptoms, it is not surprising that the time from adverse outcomes [5]. Not all causes of PAH symptom onset to definitive diagnosis averages 2 portend the same survival. Patients with PAH or more years [10]. associated with the scleroderma spectrum of dis- The cardinal findings on physical examination eases, for example, have a worse prognosis than in PAH are an accentuated second heart sound those with idiopathic PAH (PPH) [6]. Pulmonary and a tricuspid regurgitation murmur, but both veno-occlusive disease is associated with perhaps findings may be variably present. As pulmonary the worst survival, with 1-year mortality following hypertension progresses, signs of right-heart fail- diagnosis exceeding 70% [7]. ure may ensue as evidenced by jugular venous distention, a right ventricular heave, lower extremity edema, hepatomegaly, and ascites. During Diagnosis the physical examination particular attention The diagnosis of PAH requires a high index of should also be paid to findings suggestive of suspicion, because presenting symptoms are often diseases known to be associated with PAH, such protean and common to a multitude of cardio- as collagen vascular disorders and chronic liver pulmonary disorders. The goals of the PAH disease, because prognosis and treatment may evaluation are severalfold: to establish the correct vary by underlying cause. Additionally, the pres- cause of dyspnea, to exclude treatable causes of ence or absence of pulmonary flow murmurs PAH, to document the presence and severity of should be noted, because these murmurs are pulmonary hypertension and right ventricular highly suggestive of chronic thromboembolic dysfunction, to recognize the extent of functional disease, pulmonary vasculitis, or pulmonic steno- impairment, and to use this information to help sis, and have not been reported in PAH from formulate a therapeutic regimen. The tools for other causes [9]. accomplishing these tasks are largely those used Pulmonary function tests are often obtained daily by physicians to quantify and qualify other during the course of an evaluation of dyspnea but cardiopulmonary disorders, namely clinical find- have no findings specific for PAH. Diffusion is ings, lung function testing, radiographic studies, often only modestly decreased (69 Æ 25) [10]; echocardiograms, heart catheterization, and exer- more severe diffusion defects may indicate con- cise testing. During the diagnostic process special current parenchymal disease. A pattern of mild

Box 1. World Health Organization functional class in pulmonary artery hypertension Class I: Patients in with pulmonary hypertension, but without resulting limitation of physical activity. Ordinary physical activity does not cause undue dyspnea or fatigue, chest pain, or near syncope. Class II: Patients with pulmonary hypertension resulting in slight limitation of physical activity. Patients are comfortable at rest, but ordinary physical activity causes undue dyspnea or fatigue, chest pain, or near syncope. Class III: Patients with pulmonary hypertension resulting in marked limitation of physical activity. Patients are comfortable at rest, but less than ordinary physical activity cause undue dyspnea or fatigue, chest pain, or near syncope. Class IV: Patients with pulmonary hypertension resulting in inability to perform any physical activity without symptoms. Patients manifest signs of right heart failure. Dyspnea and/or fatigue may be present at rest, and discomfort is increased by any physical activity. R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 443 obstructive lung disease has been described in echocardiography was possible in only 44% of the severe PAH [11]. patients, with a correlation coefficient of 0.69. In Radiographic studies may be most useful to 52% of patients studied, the differences between exclude competing explanations for dyspnea and, measurements on right-heart catheterization and if PAH is present, to aid in excluding secondary echocardiography were greater than 10 mm Hg, causes. Chest radiographs may suggest interstitial and 48% of patients classified as having pulmo- or obstructive lung disease and demonstrate nary hypertension on echocardiography were enlarged pulmonary arteries or cardiomegaly. found to be misclassified after right-heart cathe- CT scans may likewise exhibit enlarged pul- terization. Similarly, Cotton et al [16] looked at 78 monary arteries and parenchymal lung disease but liver transplant candidates. Systolic pulmonary may also help detect chronic thromboembolic artery pressures were significantly higher when pulmonary hypertension, pulmonary veno-occlu- measured by echocardiography than when mea- sive disease, pulmonary vascular tumors, or ex- sured by right-heart catheterization, with a corre- trinsic compression of the pulmonary vessels. lation coefficient of only 0.46, and a positive Lung perfusion scanning has a pivotal role in predictive value of 37.5%. Negative predictive the evaluation of PAH, because it may reveal value, however, was nearly 92%. Daniels et al segmental and subsegmental defects consistent present a full discussion of echocardiography in with chronic thromboembolic disease, prompting pulmonary vascular disease elsewhere in this issue. further evaluation by pulmonary angiography. High-probability perfusion scans, although highly Right-heart catheterization suggestive of possible chronic thromboembolic Given its limitations, echocardiography should disease, have also been reported in pulmonary not be used to diagnose PAH definitively. Right- veno-occlusive disease [12]. In contrast, lung heart catheterization remains the standard for perfusion scans in PAH are either normal or determining pulmonary artery pressures and demonstrate a diffuse, mottled appearance. should always be performed before initiation of costly or invasive therapy. In addition to direct Echocardiography quantification of pulmonary hemodynamics (which are correlated with prognosis [4]), right- Echocardiography is a relatively inexpensive, heart catheterization may help exclude congenital noninvasive tool for estimating pulmonary artery heart disease, establish the presence and contri- pressures and detailing cardiovascular anatomy bution of left-sided heart disease, and quantify and function. In addition to providing an estima- vasodilator reserve (discussed further in the sec- tion of peak pulmonary artery pressure, echocar- tion on calcium-channel blockers). The details of diography may provide useful assessments of the right-heart catheterization procedure are dis- chamber and valve morphology, presence of cussed elsewhere in this issue. congenital heart disease, left ventricular systolic and diastolic function, and intracardiac or intra- Functional assessment pulmonary shunting [13]. Pericardial effusions are common (present in more than half of patients Functional impairment in PAH may be iden- with severe PAH), and although they often are not tified subjectively or objectively. Subjectively, the hemodynamically significant, they may carry WHO functional classification detailed in Box 1 prognostic significance, with an association with categorizes patients by reported exercise tolerance. right-heart failure, impaired exercise tolerance, Although this classification system is not perfect, and a poor 1-year survival [14]. Typical findings most PAH practitioners are familiar with its in severe PAH may include tricuspid regurgita- application, and it is widely used in research and tion, right ventricular or atrial enlargement, and therapeutic algorithms. Cardiopulmonary exercise elevated pulmonary artery pressures. Several testing provides an objective assessment of func- studies, however, have shown that the accuracy tional status correlates closely with mortality: of echocardiography is limited when compared a peak VO2max less than 10.4 mL/kg/min indepen- with direct measurement of pulmonary artery dently predicts decreased survival [17]. Cardiopul- pressure by right-heart catheterization. Arcasoy monary exercise testing, however, is expensive, and colleagues [15] examined 374 lung transplant time consuming, difficult to perform for many candidates who underwent both procedures. Esti- patients with advanced disease, and not univer- mation of systolic pulmonary artery pressures by sally available. For these reasons submaximal 444 R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 exercise testing to determine functional status is increase morbidity or mortality. This disease often desirable. The 6-minute walk has been used as afflicts women of childbearing age, and these a reliable, well-validated examination in assessing patients should be strongly counseled to avoid exercise capacity. Simply, patients walk a premea- pregnancy, because the mortality in patients with sured course, at their own pace and without PAH who become pregnant has been reported to encouragement or direction from technicians, and be as high as 30% to 56% [19,20]. Patients are the distance walked is measured. The 6-minute often concerned about the level of activity that walk distance has been shown to be correlated they are allowed to undertake. In general, patients with survival, with a distance less than 332 m should be encouraged to stay active to prevent associated with increased mortality [18]. The test is deconditioning but cautioned to not overexert inexpensive, widely available, not overly burden- themselves to the point of severe dyspnea, chest some to patients or providers, and reproducible. pain, or syncope. Given the adverse effect of Six-minute walk testing has often served as the altitude on the pulmonary vasculature, patients primary outcome endpoint in clinical trials of new with PAH should avoid altitudes higher than therapies for PAH. approximately 4000 feet.

Treatment Supportive therapies Despite limited evidence of their value, a num- In the past 2 decades an exponential growth of ber of nonspecific supportive therapies are fre- interest and research in PAH has engendered quently used in PAH, especially as right-heart several therapeutic options that were not available failure ensues. The decision to initiate any of these as recently as 5 years ago. The ideal goals of treatments must be made on an individualized treatment are to improve symptoms and survival basis. using the least invasive therapies possible. Given the expanding therapeutic options in PAH, it is Oxygen. Hypoxemia should be recognized and important to understand which patients are ap- treated, both at rest and with exercise. When given propriate for which therapy. Used inappropri- acutely, 100% oxygen has been demonstrated to ately, many of the available PAH therapies have reduce mean pulmonary artery pressure and PVR the potential for adversely affecting patient out- modestly and to increase the cardiac index [21]. come, so the dictum of primum non nocere must be Few data exist regarding the long-term benefits of paramount. For example, using PAH medications oxygen in this disorder, but extrapolating the in patients with even modest degrees of left positive effect of oxygen therapy on survival in ventricular systolic or diastolic dysfunction can patients with chronic obstructive pulmonary dis- worsen clinical status. It is recommended that order [22], it seems reasonable that hypoxemia right-heart catheterization be performed before should be recognized and treated at rest and with initiating therapy with most agents specific for the exercise. Additionally, nocturnal oxygen desatu- treatment of PAH, because right-heart catheteri- ration is common in PAH [23], even in the absence zation is a powerful resource to determine con- of concurrent sleep-disordered breathing, and tributions of left-sided heart disease and also adequate oxygenation at night should be assured provides indirect evidence of right ventricular as well. compensation, knowledge that is useful in select- ing the appropriate therapeutic regimen. Pa- Wafarin. Warfarin therapy is standard in patients tients with WHO functional class III and with IPAH (PPH) and is often recommended in class IV symptoms should be considered for other forms of PAH based on two retrospective referral to a center with specific focus and ex- studies demonstrating a modest survival benefit. pertise in the management of pulmonary vascular PAH may be complicated by in situ thrombosis, disorders. and low-level anticoagulation with warfarin is often used in an effort to help prevent this process. General treatments Two studies have established a beneficial effect on Lifestyle modification mortality [24,25]; one of these studies also ex- Treatment should be directed toward both hibited improvement in quality-of-life indices [24]. pharmacologic interventions and lifestyle modifi- Patients with chronic indwelling intravenous cath- cation, with avoidance of behaviors that might eters and right-to-left intracardiac shunting may R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 445 be especially strong candidates for chronic anti- their response in depression of total pulmonary coagulation. resistance: those with a depression of less than 20% were categorized as nonresponding, those Digoxin. Digoxin is often added to the regimen of with a depression of 20% to 50% as moderately patients with right-heart failure. In 17 patients responding, and those with a depression above examined in the catheterization laboratory, di- 50% as highly responding. Two-year survival rates goxin was shown to improve cardiac output for these groups were 38%, 47%, and 62%, res- somewhat and to decrease norepinephrine levels pectively (P < 0.05). [26]. Long-term data on its efficacy in PAH are Use of calcium-channel blockers for PAH lacking, however. should be reserved for the small subset of patients Diuretics. Diuretics are also commonly used in who exhibit near normalization of pulmonary PAH to decrease intravascular volume and to artery pressures during vasodilator challenge with improve dyspnea. It is postulated that improve- a preserved cardiac output. Long-term clinical ments in right ventricular geometry may decrease response should be closely monitored, and treat- septal intrusion into the left ventricle, improving ment should be escalated if deterioration occurs. left-ventricular function as well. It thus seems that calcium-channel blockers are rarely beneficial and should be used only in patients with a significant acute vasodilator Specific therapies response and never in nonresponders. Calcium- channel blockers with significant negative inotro- In properly selected patients the therapies pic effect, such as verapamil, should be avoided. discussed here have been shown to modify the Treatment with calcium-channel blockade should course of PAH significantly and to improve be avoided in patients who do not respond at symptoms, hemodynamics, and survival. vasodilator challenge, because serious adverse events have been reported in these patients [29]. Calcium-channel blockade Systemic hypotension in the setting of a fixed right Acute vasodilator challenge testing during ventricular cardiac output may lead to decreased right-heart catheterization may identify the small right coronary perfusion, with potential to cause percentage of patients with PAH who have disease syncope, chest pain, or even death. characterized primarily by vasoconstriction. Ap- propriate testing agents include inhaled nitric Treatment sequence oxide, adenosine, and prostacyclin. The definition Currently, the WHO functional class of pa- of a vasodilator responder has varied, but using tients largely guides therapy. Supportive therapy is a definition of a decrease in PVR of greater than used as indicated. In patients with functional class 20%, Rich and colleagues [25] found 26% of I and class II symptoms, specific therapy is not patients to be responders. These patients were indicated, because the cost and risk of therapy in treated for up to 5 years with high doses of these patients has not been justified. Patients who calcium-channel blockers (nifedipine with an av- are WHO functional class III should be considered erage daily dose of 172 mg and diltiazem with an for oral therapy, and patients whose disease average daily dose of 720 mg). At 5 years 94% of progresses on oral therapy and patients with responding patients were alive, compared with advanced class IV symptoms should be considered 55% of patients who were not responders. A for intervention with prostacyclin. Unfortunately, recent retrospective study by Sitbon et al [27] sug- cost and insurance coverage can also influence the gests that only half of acute vasodilator respond- selection of available agents: the cost of bosentan ers will benefit long term. In this study only 6% of therapy can exceed US $25,000 per year, and the all patients did well functionally when treated costs of epoprostenol and treprostinil therapy can with calcium-channel blockers alone. These pa- each exceed US $100,000 per year. Although tients were those who achieved nearly normal interest is growing in testing combination therapy, pulmonary arterial pressures (mean, <37 mm Hg) limited data are currently available. and PVR (<5 Wood units) with adequate cardiac output during the vasodilator test. Similarly, Endothelin receptor antagonists Raffy et al [28] examined 91 consecutive patients Endothelin is a smooth muscle mitogen and undergoing vasodilator challenge with epoproste- vasoconstrictive agent. Its production is increased nol. Patients were classified by the magnitude of in pulmonary hypertension [30] and is strongly 446 R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 correlated with pulmonary hemodynamics and encompassing variables such as cardiac index, 6-minute walk data [31]. Effects are mediated mean pulmonary artery pressure, and mean right through endothelin-A and endothelin-B receptors atrial pressure. Actual survival in patients re- that are variably distributed throughout the lung. ceiving bosentan at 1, 2, and 3 years was 96%, Both receptors may contribute to pulmonary 89% and 86%, respectively, compared with pre- artery smooth muscle cell proliferation and vas- dicted survivals of 69%, 57%, and 48%, respec- cular remodeling [32]. Blockade of these receptors tively, at the same time intervals. has generated significant interest as a therapeutic Bosentan (Actelion Pharmaceuticals Ltd., target. Allschwil, Switzerland, 2003) is a class X drug in Bosentan (Tracleer) is a nonselective endothe- pregnancy, because animal studies have identified lin receptor antagonist that has proven efficacy it as a teratogen. Female patients of childbearing in two double-blinded, multicenter, placebo- age should be tested for pregnancy before starting controlled trials. It is currently the only oral treatment with bosentan and should use two forms medication approved in the United States for of contraception, including a barrier method, PAH. In a pilot study, Channick and colleagues because bosentan may affect the efficacy of oral [33] randomly assigned 32 patients with idiopathic contraceptive. Patients taking bosentan should not PAH or PAH associated with the scleroderma receive concurrent glyburide because an increased spectrum of disease to treatment with bosentan risk for elevations in transaminases has been (62.5 mg twice daily for 1 month followed by associated with this combination, and they should 125 mg twice daily) or placebo in a study lasting not receive concurrent cyclosporine because mark- 12 weeks. Six-minute walk distance improved by edly elevated bosentan levels have been reported 70 m in the treatment group but declined by 6 m with coadministration of both drugs. in the placebo group. Cardiac index improved, The Food and Drug Administration (FDA) and PVR decreased, with associated improve- approved bosentan for use in WHO functional ments in Borg Dyspnea Index scores and func- class III or class IV patients. Efficacy of this drug tional class. Three patients in the placebo group may not be seen for 2 to 3 months after initiation withdrew from the study because of worsening of therapy, and it should not be used as sole clinical status. A larger follow-up study therapy for patients with advanced class IV (BREATHE-1, the Bosentan Randomized Trial disease who need a more immediate response. of Endothelin Antagonist Therapy) enrolled 213 Bosetan is, however, the only currently approved WHO functional class III or class IV patients with oral therapy for PAH and should be considered idiopathic PAH or PAH associated with connec- first-line therapy for patients with WHO func- tive tissue disease and followed them for 16 weeks tional class III symptoms. [34]. Six-minute walk distances again improved, as did Borg Dyspnea Index scores and WHO func- Prostacyclins tional class. Although the preliminary study Prostacyclin (PGI2) is produced by the vascular exhibited adverse events that did not differ from endothelium and has important vasodilatory, placebo [33], BREATHE-1 found that some antiplatelet aggregation, and antiproliferative patients suffered a significant elevation of hepatic effects. The production of prostacyclin synthase, transaminases [34]. These elevations resolved in all the enzyme necessary for prostacyclin synthesis, is patients with dose reduction or discontinuation of decreased in patients with PAH [3], as is the bosentan, but the finding reflects the need for at excretion of stable metabolites of PGI2 [36]. least monthly monitoring of liver function tests in Replacement of prostacyclin, therefore, is patients who receive this drug. a reasonable therapeutic option from a pathophys- When bosentan was introduced as a treatment iologic standpoint. The agent with which there is for PAH, there was concern that delaying therapy the most experience is epoprostenol (Flolan). with prostacyclin would adversely affect survival. Epoprostenol is a synthetic salt of prostacyclin Recent survival data with bosentan offer some that has an extremely short half-life, 3 to 5 reassurance that this is not the case. McLaughlin minuets, on average. Because of its short half-life, and colleagues [35] reported long-term follow-up it must be given by continuous infusion through of 169 patients with PPH who had been enrolled an indwelling venous catheter. In addition to the in the initial bosentan trials. Observed survival previously mentioned effects of endogenous pros- was compared with predicted survival using an tacyclin, chronic epoprostenol administration also equation based on initial NIH registry data improves right ventricular function and geometry R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 447

[37] and improves the net balance between pul- ment algorithm described later, it is reserved for monary endothelin-1 clearance and release [38]. WHO functional class IV patients and (rarely) for Epoprostenol has been investigated in a variety WHO functional class III patients who are not of patient populations. In 81 patients with idio- candidates for less-invasive therapy. pathic PAH (PPH) and WHO functional class III Because the instability of epoprostenol requires and class IV symptoms followed for a 12-week placement of an indwelling catheter for continu- period, epoprostenol improved 6-minute walk ous infusion, with the subsequent risk of compli- distances from 315 to 362 m, whereas patients in cations, there is much interest in providing the control group deteriorated from 270 to 204 m prostacyclin therapy through alternate mecha- over the study period [39]. Quality-of-life indices nisms of delivery. Treprostinil (Remodulin) is were improved, as were pulmonary hemodynam- a prostaglandin analogue that is stable at room ics, albeit modestly. There were eight deaths, all in temperature (unlike epoprostenol, which must be the placebo group. kept cool) and has a much longer half-life than The response to epoprostenol of patients with epoprostenol that permits subcutaneous adminis- PAH associated with the scleroderma spectrum of tration of this drug. Treprostinil is delivered by disease has also been examined in a randomized, continuous subcutaneous infusion in a manner controlled fashion in 111 patients, also over a similar to continuous insulin administration. The 12-week period [40]. Exercise capacity, dyspnea delivery pump is roughly the size of a pager and scores, and functional class improved in the can be worn inconspicuously by the patient. treatment group. There was no demonstrated Two studies have demonstrated the safety and difference in survival, although the study was efficacy of treprostinil. Initial pilot studies com- not powered to do so. pared intravenous treprostinil with intravenous A recent Cochrane review synthesized the epoprostenol, subcutaneous treprostinil with in- results of seven randomized, controlled trials travenous treprostinil, and subcutaneous trepros- using intravenous prostacyclin or one of its tinil with placebo [47]. Intravenous treprostinil, analogues and found that over a 12-week period subcutaneous treprostinil, and epoprostenol pro- prostaglandins seem to improve exercise capacity, duced similar reductions in PVR (approximately functional class, and several cardiopulmonary 20%). When compared with placebo, subcutane- hemodynamic variables [41]. In addition to these ous treprostinil produced statistically significant beneficial effects, epoprostenol has been shown to improvement in 6-minute walk distance (increase affect survival positively in studies involving both of 37 Æ 17 m versus a decline of 6 Æ 28 m in the adults and children [42–44]. placebo arm) and trends toward improved cardiac Epoprostenol, however, has a number of un- index and reduced PVR. favorable characteristics that limit more wide- A larger multicenter, randomized trial focused spread applicability. Its short half-life requires on 470 patients with idiopathic PAH, PAH continuous intravenous administration, necessi- associated with connective tissue disease, and tating indwelling central venous access and the PAH associated with congenital systemic to pul- associated risks of catheter-related sepsis, celluli- monary shunts [48]. Patients with WHO class II, tis, thrombosis, pneumothorax, catheter dislodge- class III, or class IV symptoms were enrolled. ment, and hemorrhage [39,40]. Additionally, After 12 weeks there was a 16-m difference in adverse effects related to the drug itself are median 6-minute walk distances between treat- common and include jaw pain, diarrhea, flushing, ment and placebo groups, and dyspnea indices headaches, nausea, vomiting, and anorexia improved in treated patients. Eighty-five percent [39,40]. Caution should be exercised when consid- of patients, however, reported significant pain at ering epoprostenol administration for subsets the infusion site, and 8% of patients withdrew of PAH patients who have a significant fixed from the study because of pain. Other typical downstream obstruction such as pulmonary veno- adverse effects common to prostacyclin analogues, occlusive disease or pulmonary capillary heman- such as diarrhea, jaw pain, flushing, and lower giomatosis. Case reports in both disorders report extremity pain, were also reported. pulmonary edema that complicates epoprostenol In the United States, the FDA has approved delivery [45,46]. treprostinil for patients with PAH characterized Despite its significant drawbacks, epoprostenol by WHO class II, class III, and class IV symptoms. is potentially life saving and is among the quickest Improvements in 6-minute walk distances, how- acting of agents currently available. In the treat- ever, do not reach the magnitude demonstrated by 448 R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 epoprostenol, and to the authors’ knowledge, no series, or preliminary studies but have not yet significant improvement in mortality has been been examined in a large-scale randomized trial demonstrated with treprostinil. Site pain is a sig- and have not received FDA approval. Two nificant problem for patients receiving this drug endothelin receptor antagonists fall into this and limits its utility in the management of PAH. category. Sitaxsentan is an endothelin-A receptor Although approved for most classes of symptoms, antagonist investigated by Barst and colleagues treprostinil should primarily be reserved for pa- [53] in a pilot study involving patients with tients with class IV symptoms who are not idiopathic PAH or PAH associated with collagen candidates for continuous intravenous infusion vascular disease or congenital systemic to pulmo- of epoprostenol and for patients with class III nary shunts. Six-minute walk distances improved symptoms for whom treatment with currently from 466 Æ 132 m to 515 Æ 141 m (P = 0.006), available oral therapies is not appropriate. and there were modest improvements in mean Two other prostacyclin analogues have been positive airway pressure and PVR. There were, investigated and used outside the United States, however, two cases of acute hepatitis, one of but neither is currently available in the United which was fatal. A larger, randomized trial is States. Iloprost is a nebulized prostaglandin with ongoing at this time. Other experimental endo- demonstrated beneficial effects on functional thelin antagonists, such as ambrisentan, are also class, 6-minute walk distance, and hemodynamics in the preliminary phases of investigation. [49–51]. A mild cough, minor headache, jaw pain, Sildenafil (Viagra) is a phosphodiesterase-5 and nausea were reported adverse effects in these inhibitor that is gaining increasing attention for studies. Because of its short half-life, iloprost must the treatment of pulmonary vascular disease. be administered six to nine times per day. Sildenafil has potential vasodilatory properties Beraprost is an oral prostacyclin analogue that mediated through cyclic GMP and cyclic AMP is approved in Japan for PAH patients. The largest systems. Several small studies, short-term hemo- trial examined 130 WHO functional class II and dynamic studies, and case reports have suggested class III patients and showed a difference in 6- beneficial results in pulmonary hypertension of minute walk distance of 25.1 m between treatment various origins, with improvement in hemody- and placebo groups, improved Borg dyspnea in- namics, exercise capacity, functional class, and dex, and no change in hemodynamics or function- dyspnea indices [54–58]. Sildenafil has also been al class [52]. Adverse effects included headache, used in combination with prostacyclins in small flushing, jaw pain, diarrhea, leg pain, and nausea. case series and in hemodynamic studies, also with Table 1 summarizes the drugs used in treating encouraging results [59–61]. In a small open-label PAH. study, 16 patients with pulmonary fibrosis and pulmonary hypertension were randomly assigned Experimental therapies to intravenous epoprostenol or sildenafil [62]. A number of therapies for PAH have shown Both epoprostenol and sildenafil were found to promise in anecdotal case reports, small case reduce PVR, but epoprostenol in this setting

Table 1 Drugs used in treating pulmonary arterial hypertension FDA Available in Drug Route Approval United States Notes Epoprostenol Intravenous (continuous) Yes—WHO Yes Infection, catheter class III/IV complications, rebound Treprostinil Subcutaneous (continuous) Yes—WHO Yes Infusion site pain class II/III/IV Iloprost Inhaled (6–9 times/d) No No Frequent inhalation Beraprost By mouth (4 times/d) No No Bosentan By mouth (2 tunes/d) Yes Yes Hepatic toxicity Sitaxsentan By mouth (1 time/d) No No Experimental Ambrisentan By mouth (1 time/d) No No Experimental Sildenafil By mouth (3 times/d) Not for PAH Yes Off-label/experimental use Abbreviations: FDA, Food and Drug Administration; PAH, pulmonary arterial hypertension. R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 449 increased ventilation/perfusion mismatch and ad- ventricular failure despite maximal medical ther- versely affected arterial oxygenation. Sildenafil, in apy, as a bridge to transplantation if maximal contrast, maintained ventilation/perfusion match- medical therapy has failed, and as a palliative ing and raised arterial oxygenation. In original procedure when no other option exists [1].It studies in patients with erectile dysfunction, the should be attempted only in institutions with drug has been shown to be relatively safe, with experience in its use. Contraindications include headache, flushing, and dyspepsia the most com- patients with impending death, receiving maxi- monly reported adverse effects [63]. Visual dismal cardiopulmonary support, or with preproce- turbances were also reported in this same series. dure systemic oxygen saturations of less than Phase III trials are currently ongoing with silde- 90% [1]. nafil in the setting of PAH. For the time being, its PAH is a declining indication for lung and use should be considered investigational. heart-lung transplantation compared with other A myriad of pathophysiologic defects in PAH indications for transplantation (constituting 4.5% are potential therapeutic targets, and it is conceiv- of procedures over a 7-year period [70]), probably able that combination therapy acting on two or because of advances in the medical management more targets could improve outcomes in this of the disease. In the United States, most trans- disease. Ongoing studies are scrutinizing this plant centers perform bilateral single-lung trans- hypothesis. plantations for PAH, although a small percentage A number of novel therapies have been sug- of centers perform single-lung transplantation for gested in PAH in animal models or small case this disorder. Heart-lung transplantation is re- reports. They include aerosolized adrenomedullin served for patients with PAH secondary to [64], oral dehydroepiandrosterone (DHEA) [65], congenital systemic-to-pulmonary shunts that simvastatin [66], and vasoactive intestinal peptide cannot be surgically corrected by isolated lung [67]. All are in the most preliminary stages of transplantation. Heart-lung transplantation in investigation and warrant further study before PAH is more frequently employed in Europe widespread clinical application. [70]. Early mortality is higher in patients with PAH than that in patients with other indications for transplantation, but overall survival approx- Surgical options in pulmonary hypertension imates 83%, 73%, 57%, 45%, and 23% at 3 Chronic thromboembolic pulmonary hyper- months, 1 year, 3 years, 5 years, and 10 years, tension should be differentiated from pulmonary respectively [70]. hypertension of other causes, because a potentially Timing of lung transplantation is problematic. curative operative procedure exists for this entity. There is currently, in the United States, no Chronic thromboembolic pulmonary hyperten- priority for severity of patient illness, and priority sion and the surgery to treat it, namely pulmonary is strictly related to time accrued while waiting for thromboendarterectomy, are beyond the scope of transplantation. Wait times average approximate- this article and are reviewed elsewhere in this ly 17 months but can be 36 months or longer [71]. issue. Heart-lung transplant lists may have even longer Atrial septostomy was first described in 1983 wait times. Therefore, if a physician waits to refer [68] and is an option in patients with severe PAH a patient with severe PAH until the patient is in who do not respond to maximal medical thera- extremis, it is unlikely that the patient will survive pies. Sandoval and colleagues [69] reported their to transplantation. Although the decision to refer experiences with the procedure in a series of 15 for lung transplantation must be individualized to such patients who underwent graded balloon each patient, in general consideration for referral dilation atrial septostomy. In the 14 patients should be given to the patient who has ‘‘symp- who survived the procedure, right ventricular tomatic, progressive disease which, despite medi- end diastolic procedure was immediately de- cal or surgical treatment, leaves the patient in creased, and cardiac output was improved, NYHA III or NYHA IV’’ [72]. From a practical although at the expense of systemic oxygen standpoint, because the clinical course of pa- desaturation secondary to the increased right to tients with PAH can be quite variable, the authors left shunting. Functional class and exercise ca- refer patients for lung transplant evaluation if pacity were improved as well. Atrial septostomy medical management necessitates the initiation should be considered in patients with severe of intravenous or subcutaneous prostaglandin PAH and recurrent syncope or progressive right administration. 450 R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452

Summary [8] Daily PO, Dembitzsky WP, Iversen S, et al. Current early results of pulmonary thromboendarterectomy The diagnosis of pulmonary hypertension re- for chronic pulmonary embolism. Eur J Cardio- quires a high index of suspicion and careful thorac Surg 1990;4:117. attention to assessing the severity and classifica- [9] Fedullo PF, Auger WR, Kerr KM, et al. Chronic tion of disease. Proper evaluation and under- thromboembolic pulmonary hypertension. N Engl J standing of determinants of severity in PAH are Med 2001;345:1465. necessary to guide appropriate therapy. There are [10] Rich S, Dantzker DR, Ayres SM, et al. Primary now highly effective therapies for PAH that have pulmonary hypertension. A national prospective study. Ann Intern Med 1987;107:216. meaningfully improved the outcome for patients, [11] Meyer FJ, Ewert R, Hoeper MM, et al. Peripheral and it likely that that the future is even brighter, airway obstruction in primary pulmonary hyper- with development of combination regimens and tension. Thorax 2002;57:473. additional therapies that attack specific targets [12] Bailey CL, Channick R, Auger W, et al. ‘‘High within the pulmonary vasculature and right ven- probability’’ perfusion lung scans in pulmonary tricle. Despite the increasing availability of oral venoocclusive disease. Am J Respir Crit Care Med therapies for PAH, consideration should be given 2000;162:1974. to referral of patients with functional class III and [13] McGoon MD. The assessment of pulmonary IV symptoms to specialized PAH centers that may hypertension. Clin Chest Med 2001;22:493. have additional experience in the timing and [14] Hinderliter AL, Willis PWt, Long W, et al. Frequency and prognostic significance of pericar- nature of escalation of therapy when needed. dial effusion in primary pulmonary hypertension. Referral also may help concentrate a fairly rare PPH Study Group. Am J Cardiol 1999;84:481. population of patients to facilitate research that, [15] Arcasoy SM, Christie JD, Ferrari VA, et al. one hopes, may lead to continued advances in the Echocardiographic assessment of pulmonary hy- management of this disease. Although survival pertension in patients with advanced lung disease. remains limited, the authors are encouraged by Am J Respir Crit Care Med 2003;167:735. the interest and commitment to this disease that [16] Cotton CL, Gandhi S, Vaitkus PT, et al. Role of has engendered continued progress in the de- echocardiography in detecting portopulmonary velopment of an armamentarium to fight it. hypertension in liver transplant candidates. Liver Transpl 2002;8:1051. [17] Wensel R, Opitz CF, Anker SD, et al. Assessment References of survival in patients with primary pulmonary hypertension: importance of cardiopulmonary ex- [1] Simonneau G, Galie` N, Rubin L, Langleben D, ercise testing. Circulation 2002;106:319. Seeger W, Domenighetti G, et al. Clinical classifi- [18] Miyamoto S, Nagaya N, Satoh T, et al. Clinical cation of pulmonary hypertension. J Am Coll correlates and prognostic significance of six-minute Cardiol 2004;43:55–125. walk test in patients with primary pulmonary [2] Meyrick B. The pathology of pulmonary artery hypertension. Comparison with cardiopulmonary hypertension. Clin Chest Med 2001;22:393. exercise testing. Am J Respir Crit Care Med 2000; [3] Tuder RM, Cool CD, Geraci MW, et al. Prostacy- 161:487. clin synthase expression is decreased in lungs from [19] McCaffrey RN, Dunn LJ. Primary pulmonary patients with severe pulmonary hypertension. Am J hypertension in pregnancy. Obstet Gynecol Surv Respir Crit Care Med 1999;159:1925. 1964;19:567. [4] D’Alonzo GE, Barst RJ, Ayres SM, et al. Survival [20] Weiss BM, Zemp L, Seifert B, et al. Outcome of in patients with primary pulmonary hypertension. pulmonary vascular disease in pregnancy: a system- Results from a national prospective registry. Ann atic overview from 1978 through 1996. J Am Coll Intern Med 1991;115:343. Cardiol 1998;31:1650. [5] Raymond RJ, Hinderliter AL, Willis PW, et al. [21] Roberts DH, Lepore JJ, Maroo A, et al. Oxygen Echocardiographic predictors of adverse outcomes therapy improves cardiac index and pulmonary in primary pulmonary hypertension. J Am Coll vascular resistance in patients with pulmonary Cardiol 2002;39:1214. hypertension. Chest 2001;120:1547. [6] Kawut SM, Taichman DB, Archer-Chicko CL, et [22] Group NOTT. Continuous or nocturnal oxygen al. Hemodynamics and survival in patients with therapy in hypoxemic chronic obstructive lung pulmonary arterial hypertension related to systemic disease: a clinical trial. Ann Intern Med 1980;93: sclerosis. Chest 2003;123:344. 391. [7] Holcomb BW Jr, Loyd JE, Ely EW, et al. [23] Rafanan AL, Golish JA, Dinner DS, et al. Pulmonary veno-occlusive disease: a case series Nocturnal hypoxemia is common in primary and new observations. Chest 2000;118:1671. pulmonary hypertension. Chest 2001;120:894. R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452 451

[24] Frank H, Mlczoch J, Huber K, et al. The effect of structure and function in primary pulmonary anticoagulant therapy in primary and anorectic hypertension. Primary Pulmonary Hypertension drug-induced pulmonary hypertension. Chest 1997; Study Group. Circulation 1997;95:1479. 112:714. [38] Langleben D, Barst RJ, Badesch D, et al. Contin- [25] Rich S, Kaufmann E, Levy PS. The effect of high uous infusion of epoprostenol improves the net doses of calcium-channel blockers on survival in balance between pulmonary endothelin-1 clearance primary pulmonary hypertension. N Engl J Med and release in primary pulmonary hypertension. 1992;327:76. Circulation 1999;99:3266. [26] Rich S, Seidlitz M, Dodin E, et al. The short-term [39] Barst RJ, Rubin LJ, Long WA, et al. A comparison effects of digoxin in patients with right ventricular of continuous intravenous epoprostenol (prostacy- dysfunction from pulmonary hypertension. Chest clin) with conventional therapy for primary pulmo- 1998;114:787. nary hypertension. The Primary Pulmonary [27] Sitbon O, Humbert M, Jais X, Ioos V, Hamid AM, Hypertension Study Group. N Engl J Med 1996; Parent F, et al. Acute vasodilator responsiveness 334:296. and long-term response to calcium-channel blockers [40] Badesch DB, Tapson VF, McGoon MD, et al. in different forms of pulmonary artery hypertension Continuous intravenous epoprostenol for pulmo- [abstract A86]. In: Programs and abstracts of the nary hypertension due to the scleroderma spectrum 100th International Conference of the American of disease. A randomized, controlled trial. Ann Thoracic Society. Orlando, Florida: 2004. p. A210. Intern Med 2000;132:425. [28] Raffy O, Azarian R, Brenot F, et al. Clinical [41] Paramothayan NS, Lasserson TJ, Wells AU, et al. significance of the pulmonary vasodilator response Prostacyclin for pulmonary hypertension. Cochrane during short-term infusion of prostacyclin in primary Database Syst Rev CD002994, 2003. pulmonary hypertension. Circulation 1996;93:484. [42] Barst RJ, Maislin G, Fishman AP. Vasodilator [29] Sitbon O, Humbert M, Jagot JL, et al. Inhaled therapy for primary pulmonary hypertension in nitric oxide as a screening agent for safely children. Circulation 1999;99:1197. identifying responders to oral calcium-channel [43] Kuhn KP, Byrne DW, Arbogast PG, et al. Out- blockers in primary pulmonary hypertension. Eur come in 91 consecutive patients with pulmonary Respir J 1998;12:265. arterial hypertension receiving epoprostenol. Am J [30] Stewart DJ, Levy RD, Cernacek P, et al. Increased Respir Crit Care Med 2003;167:580. plasma endothelin-1 in pulmonary hypertension: [44] Sitbon O, Humbert M, Nunes H, et al. Long-term marker or mediator of disease? Ann Intern Med intravenous epoprostenol infusion in primary pul- 1991;114:464. monary hypertension: prognostic factors and sur- [31] Rubens C, Ewert R, Halank M, et al. Big vival. J Am Coll Cardiol 2002;40:780. endothelin-1 and endothelin-1 plasma levels are [45] Humbert M, Maitre S, Capron F, et al. Pulmonary correlated with the severity of primary pulmonary edema complicating continuous intravenous pros- hypertension. Chest 2001;120:1562. tacyclin in pulmonary capillary hemangiomatosis. [32] Davie N, Haleen SJ, Upton PD, et al. ET(A) and Am J Respir Crit Care Med 1998;157:1681. ET(B) receptors modulate the proliferation of [46] Palmer SM, Robinson LJ, Wang A, et al. Massive human pulmonary artery smooth muscle cells. Am pulmonary edema and death after prostacyclin J Respir Crit Care Med 2002;165:398. infusion in a patient with pulmonary veno-occlusive [33] Channick RN, Simonneau G, Sitbon O, et al. disease. Chest 1998;113:237. Effects of the dual endothelin-receptor antagonist [47] McLaughlin VV, Gaine SP, Barst RJ, et al. Efficacy bosentan in patients with pulmonary hypertension: and safety of treprostinil: an epoprostenol analog a randomised placebo-controlled study. Lancet for primary pulmonary hypertension. J Cardiovasc 2001;358:1119. Pharmacol 2003;41:293. [34] Rubin LJ, Badesch DB, Barst RJ, et al. Bosentan [48] Simonneau G, Barst RJ, Galie N, et al. Continuous therapy for pulmonary arterial hypertension. N Engl subcutaneous infusion of treprostinil, a prostacyclin J Med 2002;346:896. analogue, in patients with pulmonary arterial [35] McLaughlin V, Sitbon O, Rubin LJ, et al. The effect hypertension: a double-blind, randomized, place- of first line bosentan on survival of patients bo-controlled trial. Am J Respir Crit Care Med with primary pulmonary hypertension. Abstract, 2002;165:800. American Thoracic Society, 2003. [49] Hoeper MM, Schwarze M, Ehlerding S, et al. Long- [36] Christman BW, McPherson CD, Newman JH, et al. term treatment of primary pulmonary hypertension An imbalance between the excretion of thrombox- with aerosolized iloprost, a prostacyclin analogue. ane and prostacyclin metabolites in pulmonary N Engl J Med 2000;342:1866. hypertension. N Engl J Med 1992;327:70. [50] Olschewski H, Ghofrani HA, Schmehl T, et al. [37] Hinderliter AL, Willis PWT, Barst RJ, et al. Effects Inhaled iloprost to treat severe pulmonary hyper- of long-term infusion of prostacyclin (epoprostenol) tension. An uncontrolled trial. German PPH Study on echocardiographic measures of right ventricular Group. Ann Intern Med 2000;132:435. 452 R. Channick, T.L. Williamson / Cardiol Clin 22 (2004) 441–452

[51] Olschewski H, Simonneau G, Galie N, et al. [62] Ghofrani HA, Wiedemann R, Rose F, et al. Inhaled iloprost for severe pulmonary hypertension. Sildenafil for treatment of lung fibrosis and N Engl J Med 2002;347:322. pulmonary hypertension: a randomised controlled [52] Galie N, Humbert M, Vachiery JL, et al. Effects of trial. Lancet 2002;360:895. beraprost sodium, an oral prostacyclin analogue, in [63] Goldstein I, Lue TF, Padma-Nathan H, et al. Oral patients with pulmonary arterial hypertension: a sildenafil in the treatment of erectile dysfunction. randomized, double-blind, placebo-controlled trial. Sildenafil study group. N Engl J Med 1998;338:1397. J Am Coll Cardiol 2002;39:1496. [64] Kandler MA, Von Der Hardt K, Mahfoud S, et al. [53] Barst RJ, Rich S, Widlitz A, et al. Clinical efficacy Pilot intervention: aerosolized adrenomedullin re- of sitaxsentan, an endothelin-A receptor antagonist, duces pulmonary hypertension. J Pharmacol Exp in patients with pulmonary arterial hypertension: Ther 2003;306:1021. open-label pilot study. Chest 121:1860;2002. [65] Bonnet S, Dumas-de-La-Roque E, Begueret H, et [54] Bharani A, Mathew V, Sahu A, et al. The efficacy al. Dehydroepiandrosterone (DHEA) prevents and and tolerability of sildenafil in patients with reverses chronic hypoxic pulmonary hypertension. moderate-to-severe pulmonary hypertension. Indian Proc Natl Acad Sci U S A 2003;100:9488. Heart J 2003;55:55. [66] Girgis RE, Li D, Zhan X, et al. Attenuation of [55] Ghofrani HA, Schermuly RT, Rose F, et al. chronic hypoxic pulmonary hypertension by sim- Sildenafil for long-term treatment of nonoperable vastatin. Am J Physiol Heart Circ Physiol 2003;285: chronic thromboembolic pulmonary hypertension. H938. Am J Respir Crit Care Med 2003;167:1139. [67] Petkov V, Mosgoeller W, Ziesche R, et al. [56] Michelakis E, Tymchak W, Lien D, et al. Oral Vasoactive intestinal peptide as a new drug for sildenafil is an effective and specific pulmonary treatment of primary pulmonary hypertension. J vasodilator in patients with pulmonary arterial Clin Invest 2003;111:1339. hypertension: comparison with inhaled nitric oxide. [68] Rich S, Lam W. Atrial septostomy as palliative Circulation 2002;105:2398. therapy for refractory primary pulmonary hyper- [57] Molina J, Lucero E, Luluaga S, et al. Systemic tension. Am J Cardiol 1983;51:1560. lupus erythematosus-associated pulmonary hyper- [69] Sandoval J, Gaspar J, Pulido T, et al. Graded tension: good outcome following sildenafil therapy. balloon dilation atrial septostomy in severe primary Lupus 2003;12:321. pulmonary hypertension. A therapeutic alternative [58] Sastry BK, Narasimhan C, Reddy NK, et al. A for patients nonresponsive to vasodilator treatment. study of clinical efficacy of sildenafil in patients with J Am Coll Cardiol 1998;32:297. primary pulmonary hypertension. Indian Heart J [70] Trulock E, Edwards L, Taylor D, et al. The registry 2002;54:410. of the International Society for Heart and Lung [59] Ghofrani HA, Rose F, Schermuly RT, et al. Oral Transplantation: Twentieth Official Adult Lung sildenafil as long-term adjunct therapy to inhaled and Heart-lung Transplant Report–2003. J Heart iloprost in severe pulmonary arterial hypertension. Lung Transplant 2003;22:625. J Am Coll Cardiol 2003;42:158. [71] Pielsticker EJ, Martinez FJ, Rubenfire M. Lung and [60] Ghofrani HA, Wiedemann R, Rose F, et al. heart-lung transplant practice patterns in pulmo- Combination therapy with oral sildenafil and in- nary hypertension centers. J Heart Lung Transplant haled iloprost for severe pulmonary hypertension. 2001;20:1297. Ann Intern Med 2002;136:515. [72] Physicians American Society for Transplant Phys- [61] Stiebellehner L, Petkov V, Vonbank K, et al. Long- iciano, American Thoracic Society, European Re- term treatment with oral sildenafil in addition to spiratory Society, et al. International guidelines for continuous IV epoprostenol in patients with pulmo- the selection of lung transplant candidates. Am J nary arterial hypertension. Chest 2003;123:1293. Respir Crit Care Med 1998;158:335.