Treatment for pulmonary hypertension in Australia: too far too fast?

Pulmonary arterial hypertension (PAH) is the result of increased pulmonary resistance due to vasoconstriction, proliferation and then fibrosis in the lung microcirculation.1 Although all are common components of PAH, anti-proliferative and anti-fibrotic therapies have not been successful clinically so far. Instead, treatment to vasodilate the pulmonary microcirculation with nitrates, endothelin receptor antagonists such as bosentan, prostanoids (such as iloprost, epoprostenol and treprostinil) and phosphodiesterase-5 inhibitors such as sildenafil have been developed with apparent benefit to a number of patients with PAH. However, the extent of benefit and the population likely to derive benefit is difficult to quantify as double-blind randomised clinical trial data or similar with clinically relevant end points have not been forthcoming, despite the huge uptake of the medications in Australia over the last 5 years.2 Thus papers published in this issue of the Internal Medicine Journal are welcome and provide reassurance on toxicity but still lack the clarity to help clinicians estimate the comparative extent of benefit. There are several outstanding questions needing to be answered. First, whether PAH that is idiopathic or secondary to a variety of aetiologies are different disease entities which respond differently to medications, and whether progressive stages of disease respond differently. Second, whether the value put on the surrogate markers of disease status in PAH is appropriate and third, whether the improvement in these surrogate measures is sufficiently clinically meaningful to commit hospital and community resources. Finally, has the time arrived for randomized controlled studies appropriately powered and of sufficient duration with clinically relevant end points to be conducted? Vasodilator treatment for PAH (idiopathic or connective tissue disease related) was licensed in Australia in 2002 on the basis of placebo-controlled, randomized trial data demonstrating that patients receiving bosentan over a 16-week period of follow up had a mean improvement of 36 m on their 6-min walk test (baseline value of 330 m), as well as improved dyspnoea indices (mean treatment effect of 0.6 using a scale of 1–10) and increased time to clinical worsening.3,4 The patients involved in the two studies were a well-defined population (mean age 48.7 years ± 15.8, 71% primary and 23% secondary PAH). Translation of outcomes from this well-defined study cohort to other patient groups with various other causes and stages of PAH rapidly followed. Practically, this was driven by a consumer need for treatment for a disease that had few medical options and which often manifests in a relatively young population. Practitioners and the pharmaceutical industry have been active in the post-marketing phase, undertaking further small studies and often with no randomization or controls, making comparison with a similar group of patients difficult. Furthermore, the sponsoring of active treatment within centres of expertise in managing patients with PAH was promoted despite the uncertain clinical data at that time. An example of this activity is the dissemination of documents containing select data from a patient registry to local clinicians purporting to demonstrate a benefit of a PAH drug in both idiopathic and scleroderma-associated PAH.5 These particular data were a post-hoc analysis of registry data from patients with PAH of many aetiologies and showed a non-statistical difference between patients with idiopathic versus scleroderma-related PAH continuing and/or starting bosentan. Standard statistical features were not presented, nor was there presentation of any adjustments made for the different baseline characteristic of patients (e.g. disease stage, severity) from the 15 centres participating in the special access scheme and contributing database. Moreover, these data did little to address the most important aspect, which concern quality of life and survival estimates. In a worse-case interpretation, both of these diseases may respond similarly to this PAH medication, but inferiorly compared to placebo, and with cost. Documents such as these (often without confidence intervals or sensitivity analyses) showing small annual differences over an uncertain time frame in patients receiving a new medication as opposed to matched historical controls, controls that received treatment in another healthcare system in an earlier time frame prior to specialist pulmonary hypertension clinics, are not uncommon at large medical conferences. What worries us is that these are used as supporting evidence to list these medications on Hospital Formularies or the Pharmaceutical Benefits Scheme. Information such as that from a Cochrane analysis of 11 randomised controlled trials (RCT) with 1457 PAH patients showing no mortality benefit, despite marginal clinical benefit in some of the surrogate markers,6 or a 12-month double-blind RCT of bosentan which did not show any difference in the primary efficacy end point (a change in the 6-min walk distance from baseline up to month 12) or secondary end points including time to death, worsening results of pulmonary function or the occurrence of clinically important worsening,7 are not provided by clinicians to accompany the request, in our experience. There have been additional meta-analyses to the Cochrane data, including all population groups and all therapies. Galie et al. showed marginal mortality benefit (absolute risk reduction [ARR] 1.6%) with confidence interval just statistically significant (confidence interval 0.39–1.00; P= 0.048).8 A 2010 meta-analysis reviewing all 21 comparative studies (predominantly using historical controls) including all PAH therapies did demonstrate a significant reduction in mortality (relative risk [RR] 43%), and small, clinically insignificant change in surrogates.9 However the mortality benefit was driven by intravenous prostanoid treatment and was not actually observed for endothelin receptor antagonists or phosphodiesterase-5 inhibitors. This is an important finding, and supports the study outcomes of the observational Alfred study reported10 suggesting that intravenous epoprostenol, despite lack of evidence of superiority may be of some short-term benefit in stabilizing unwell functional class IV patients, albeit at an estimate of tens of thousands of dollars per year per inpatient. However, it should be noted that all treatment arms in the above-quoted meta-analyses are compared against historical controls. These may have been too sick to have treatment, or included from a previous era before supportive therapies such as warfarin and calcium channel blockers were consistently available. Further, the mean study length was 14 weeks in the Ryerson meta-analysis, and 16 weeks in a further analysis11 which is too short to define long-term benefits. Interestingly, the few studies that have examined outcomes beyond 16 weeks failed to demonstrate lasting benefits even though the short-term results were encouraging. Additionally, many patients who do get initial benefit of these agents deteriorate over time. This may infer that the microcirculation is becoming fibrotic and treatment should be ceased, although it is possible also that as different pathophysiological pathways are involved in this process that a second agent may be of benefit. It does not necessarily follow that continuing on the failed drug with an additional agent would provide any benefit over the second therapy alone, an apparent developing practice in this illness and for which evidence of benefit needs to be demonstrated. These current practices then are of great concern not only for policy makers, regulatory authorities and pricing agencies, but for patients also. The opportunity cost of supplying a drug that is not likely to work, when there may be others that do needs to be considered. Worse, patients may suffer side effects without any benefit. There is a further question regarding the value we are putting on surrogate markers of disease status in PAH. Many studies of medications have used surrogate end points to drive reimbursement hurdles. Caution, however, has been flagged regarding whether or not these improve quality of life or survival.12 In the PAH setting, we need to know how 6MWT translates into morbidity and mortality benefit and whether this translation to clinical outcome differs in different conditions.13 In the Macchio meta-analysis, for example, it was highlighted that the mechanism by which mortality was reduced was unrelated to a specific drug class or dose, haemodynamics or the change in 6MWT (the primary end point in most trials). This strongly suggests neither surrogates nor dose are helpful; our current clinical trials are not measuring an end point that can help us make an informed choice regarding therapy.11 These points are also pertinent to other publications14,15 in this issue of the Journal regarding pulmonary emboli (PE). This is because despite the awareness of idiopathic PAH, PE remain the most studied and ‘treatable’ of all causes of PHT, if diagnosed. However, they are unlikely to respond to the aforementioned medications; it is thus imperative this cause of PAH is considered and antithrombotic medications instituted, if appropriate. We suggest measuring morbidity and mortality directly. We think surrogates could be helpful in certain circumstances; however, there also needs to be appropriate validation and comparison with hard clinical end points. One is still left wondering why when we have three retrospective publications from three large centres in Australia10,16,17 why a prospective RCT with clinically relevant end points has not yet been completed. We feel the time has come to question the clinically meaningful benefit of these medications in PAH, and recommend National Health and Medical Research Committee standard evidence-based guidelines be developed to guide funders of drugs. Whilst we acknowledge the importance and support of industry in managing patients with this condition, guidelines need to be developed by specialists at arm's length from the funding that comes from the pharmaceutical industry. If surrogate outcomes are used, then these need to be clinically relevant (such as rates of hospitalization and overall survival) and validated so that we can predict likelihood of response to therapy. We also need trials with longer controlled periods (at least 1 year) of follow up. Studies such as the SERAPHIN18 study which are now underway are encouraging. PAH remains a progressive, debilitating and fatal disease, so there is some urgency to undertake appropriate trial designs to produce definitive results about long-term morbidity and mortality in various patient groups affected by PAH to guide appropriate therapy. The authors would like to thank Mrs Breeann Silvester, School of Medicine, University of Queensland, Princess Alexandra Hospital, Brisbane for administrative assistance.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.