#NephJC Chat
Tuesday Dec, 6th, 2022 at 9 pm EST (AEST = Dec 7th, 11am)
Wednesday Dec 7th, 2022, at 9 pm IST and 3:30 pm GMT
Lancet. 2022 Nov 4;S0140-6736(22)02034-7. doi: 10.1016/S0140-6736(22)02034-7. Online ahead of print.
Dual endothelin antagonist aprocitentan for resistant hypertension (PRECISION): a multicentre, blinded, randomised, parallel-group, phase 3 trial
Markus P Schlaich, Marc Bellet, Michael A Weber, Parisa Danaietash, George L Bakris, John M Flack, Roland F Dreier, Mouna Sassi-Sayadi, Lloyd P Haskell, Krzysztof Narkiewicz, Ji-Guang Wang, PRECISION investigators.
PMID: 36356632
Introduction
We are all well aware that hypertension is an extremely prevalent condition that causes significant morbidity and mortality. Resistant hypertension, present in roughly 10% of all hypertensive patients (Noubiap, Hear 2019), is as defined, notoriously difficult to treat. Management can be frustrating, especially as there have been no new therapies since the introduction of RAS blockade in the 1970s. Among features that make research difficult in this area are regression to mean, observer bias, varying adherence, and associated high variance which have tripped up many studies in the past (see the NephJC discussion of the denervation trials).
Into this dark arena, comes the PRECISION trial, which brings some light by exploring the first new treatment for resistant hypertension in ages with a tantalizingly novel study design!
Since the discovery of endothelin in the late 1980s (Yanagisawa et al., Nature 1988) there has been much interest in its potential as a therapeutic target in hypertension management. Endothelin-1 (ET-1) is a 21-amino acid peptide that acts as a potent vasoconstrictor by binding to the endothelin A and endothelin B receptors. Several preclinical studies have indeed shown that the effects of ET-1 on renal and vascular cells contribute to the pathogenesis of hypertension (Dhaun et al., Hypertension 2008; Prasad et al, Cardiol Rev 2009) as summarized in this excellent figure (Kohan, Curr Opin Nephrol Hypertens 2010). Despite significant interest in this pathway in these diseases, endothelin receptor antagonists (ERAs) have not yet worked their way into the armamentarium for these conditions.
Figure from Kohan, Curr Opin Nephrol Hypertens 2010
While there is much interest in examining ERAs in the setting of hypertension and CKD, currently, their main clinical use is in the setting of pulmonary arterial hypertension (Kuntz, Lung 2016). Their benefit in pulmonary hypertension is thought to be due to their vasodilatory and anti-hyperplastic effect in the pulmonary vascular bed. We’ve previously explored the endothelin system in the setting of diabetic nephropathy on NephJC, focusing on the RADAR and SONAR trials, which examined albuminuria reduction with atrasentan, a selective endothelin A receptor antagonist. The ERA darusentan, another selective endothelin A receptor antagonist, has been investigated specifically in the setting of resistant hypertension (BP above target despite 3 different classes including a diuretic, ACEI or ARB, and a long-acting CCB), but unfortunately, with mixed efficacy results (Weber et al., Lancet 2009; Bakris et al., Hypertension 2010). Along with conflicting efficacy data, their clinical use has also been limited by significant adverse effects related to fluid retention, which has even contributed to the early termination of a previous RCT due to high heart failure incidence (Mann et al., JASN 2010). Antagonism of endothelin A and B receptors is thought to produce fluid retention by causing vasodilation, which contributes to vasopressin and aldosterone activation, and by causing increased vascular permeability.
The hypothesis that this endothelin pathway may still play an important role in the pathophysiology of resistant hypertension led trial investigators to design the PRECISION study to further investigate. In this novel trial, investigators examine the effects of aprocitentan, a dual endothelin receptor A and B antagonist.
The Study
Methods
PRECISION was a multi-center, blinded, randomized, parallel-group phase 3 study. Participants were enrolled from centers in North America, Europe, Asia and Australia (193 sites in 22 countries). It examined whether the addition of aprocitentan to a 3-agent antihypertensive regimen reduced blood pressure in patients with resistant hypertension after 4 weeks of double-blind treatment. It also sought to determine whether any blood pressure lowering effect was sustained at week 40. The doses selected in this trial (12.5 mg and 25 mg) were based on a phase-2 study suggesting doses in this range maximized efficacy while minimizing rates of fluid retention (Verweij et al, Hypertension 2020).
Participants that passed that initial screening were switched on to a standardized background antihypertensive regimen that included a single-pill combination of amlodipine (5 or 10 mg), hydrochlorothiazide (25 mg) and valsartan (160 mg) for 4 weeks. There was a subsequent single-blind run-in period (4 weeks) that added placebo to the standardized background regimen to help exclude placebo responders. It’s important to note that the standardized background regimen was continued throughout the entire remainder of the study duration after initiation. Notably, participants who were on beta blocker therapy at the time of screening were continued on this therapy throughout the study period.
Study Population
Inclusion criteria
Adults >18 years old
Uncontrolled BP despite at least 3 antihypertensives (all different classes) within 1 year of screening
Treatment with at least 3 antihypertensives for at least 4 weeks prior to the screening visit
Sitting SBP (average of 3 automated office BP measurements) ≥140 mmHg at the initial screening visit
Mean trough sitting SBP ≥ 140 mmHg after being switched to the standardized background antihypertensive regimen
Women of childbearing age included if:
Negative pregnancy test at screening and baseline
Willing to take pregnancy tests during and after the study
Willing to take birth control from screening up to at least 30 days after randomized treatment discontinuation
Randomization inclusion criteria (end of run-in period)
Stable dose of the standardized background antihypertensive regimen for at least 1 week prior to the end of the run-in period
Mean trough sitting SBP ≥ 140 mmHg using office BP measurements
≥ 80% compliance (tablet counting) to placebo and the standardized background antihypertensive therapy
Exclusion criteria
Participants with poor treatment adherence
Secondary causes of HTN except for OSA
Severe hypertension (2 separate measurements of sitting SBP≥180 mmHg and/or sitting DBP≥110 mmHg)
Pregnant or lactating participants
Unstable cardiac disease at screening or past history of the same as judged by the investigator (i.e. uncontrolled symptomatic arrhythmia, atrial fibrillation, CHF NYHA stage II with MS or AS, CHF NYHA stage III or IV)
Severe kidney disease (CKD-EPI eGFR < 15 ml/min/1.73m2)
T1DM
Concurrent treatment with medications affecting BP such as > 80mg/day of furosemide, cyclosporine and tacrolimus, tyrosine kinase inhibitors, stimulants, sympathomimetics, systemic corticosteroids, abiraterone, or regular use of PDEi
Laboratory criteria
ALT or AST > 3 times the upper limit of normal range, or severe hepatic impairment.
Hemoglobin < 100 g/L
NT-proBNP ≥ 500 pg/mL
Intervention
There were 3 parts to the study design:
Double-blind randomized treatment 1:1:1 with aprocitentan 12.5 mg, 25 mg or placebo (4 weeks)
Single-blind active treatment with aprocitentan 25 mg (32 weeks)
Double-blind withdrawal with re-randomization 1:1 to aprocitentan 25 mg or placebo (12 weeks)
30-day safety follow-up period after the last study treatment dose
Supplementary Figure S1 from Schlaich et al, Lancet 2022
In addition, intake of standard baseline therapy (i.e. single pill combination of valsartan, amlodipine, hydrochlorothiazide) was monitored by assessing both participant urine via liquid chromatography with tandem mass spectrometric to detect valsartan, and by direct observed treatment intake, performed before start of the ambulatory BP monitoring (ABPM).
Trial Endpoints
Primary endpoint
Change in the mean trough sitting SBP measured by office BP measurements from baseline to week 4 (Part 1).
Secondary endpoints
The key secondary outcome was change in the mean trough sitting SBP measured by office BP measurements from withdrawal baseline (week 36) to week 40 (Part 3).
Other secondary endpoints included:
Change in mean trough sitting diastolic blood pressure from baseline to week 4 and from withdrawal baseline (week 36) to week 40
Change in 24 h SBP and DBP using ambulatory blood pressure monitoring (ABPM) from baseline to week 4 and from withdrawal baseline (week 36) to week 40
Change from withdrawal baseline (week 36) to weeks 38, 44, and 48 in mean trough sitting SBP, DBP and urine ACR
BP Measurements
The primary and key secondary endpoint is office BP - and this was measured by unattended automated oscillometric BP - similar to SPRINT, but using a different device (Microlife Watch BP). This automates a resting of 5 minutes and provides an average of multiple readings (design paper mentions 5, Lancet paper mentions 3 readings). In addition, ABPM was a secondary outcome - and a key aspect is that direct observed therapy was performed prior to the ABPM (similar to the SPYRAL denervation trials).
Statistical Analysis
The study used a mixed model with factors for treatment group, visit, and treatment by visit interaction, along with covariates for baseline SBP and the interaction between baseline and visit. Prespecified subgroup analyses were also performed.
The authors also performed sensitivity analyses to assess the impact of deviations from the missing-at-random assumption of the mixed models and of premature discontinuation of double-blind treatment or adding a diuretic or antihypertensive rescue medication. Supportive analyses were also done to examine the impact of deviating from protocol and substitution rules.
Funding
The trial was funded by Idorsia Pharmaceuticals Ltd, which is the company that manufactures aprocitentan. They had input into the design, data analysis, manuscript writing, and decision to publish the findings. Four of the authors are Idorsia employees, and the analysis was done by one of them.
Results
Study Population
The study was conducted between June 18, 2018, and April 25, 2022. There were 1965 patients screened, of which 730 patients were randomized. Of these 730 randomized individuals, 704, 613 and 577 completed parts 1, 2, and 3 of the study, respectively. Participant flow through the study is summarized in Figure 1 from the paper.
Figure 1 from Schlaich et al, Lancet 2022
Of interest to us, the design paper (Danaietash et al, J Clin HT 2022) also describes some more nuggets from the first part of the study. Without such an elaborate screening and run-in phase, many of these pseudo-resistant hypertension participants would have been randomized and created noise in the outcomes.
Figure 2 from Danaietash et al, J Clin HT 2022 showing the details of screen failures and the reasons.
The mean age at the time of screening was about 62 years across all groups. There were 434 (59%) male participants and 605 (83%) were white. There were 82 (11%) Black patients and they represented 37% (78 of 211) of the participants from the USA. Five hundred sixteen (71%) patients were on maximal dosing of the standardized background antihypertensive regimen (10 mg/25 mg/160 mg) and 423 (58%) patients continued beta blocker therapy. Most participants were on ≥4 antihypertensives at the time of enrollment (n = 460, 63%). Diabetes was present in 395 (54%) of patients, 143 (20%) participants had a baseline history of CHF, and a minority of participants had OSA (n=103, 14%). Interestingly, across all groups the degree of hypertension was much less at baseline using ABPM (~137/83 mmHg).
Table 1 from Schlaich et al, Lancet 2022
Primary outcome
The least square mean change in SBP between baseline and 4 weeks in part 1 of the study was –15.3 (0.9) mmHg for aprocitentan 12.5 mg, –15.2 (0.9) mmHg for aprocitentan 25 mg, and –11.5 (0.9) mm Hg for placebo. This corresponded to a difference compared to placebo of –3.8 (1.3) mmHg (p=0.0042) and –3.7 (1.3) mmHg (p=0.0046) for aprocitentan 12.5 mg and 25 mg respectively.
Figure 2A from Schlaich et al, Lancet 2022
Diastolic blood pressure also decreased with both aprocitentan doses when compared to placebo (–3.9 mmHg with the 12.5 mg dose and –4.5 mmHg with the 25 mg dose).
Figure 2B from Schlaich et al, Lancet 2022
The sensitivity and supporting analyses supported the main findings of the study. The authors specifically highlight that adding or uptitrating diuretics or using rescue antihypertensives did not affect the primary and key secondary results.
There was no significant difference in the primary outcome with the subgroup analyses based on sex, baseline weight, BMI, race, region, diabetes status, beta blocker use, or protocol subgroups (summarized in Supplementary Figure S3 below). The authors hypothesize that aprocitentan was favored in particular subgroups on the basis of age, proteinuria and baseline kidney function (age ≥ 75 years group at the 25 mg dose, uACR 300 mg/g sub-group at both doses, and eGFR 15 – <60 sub-group at the 25 mg dose).
Supplementary Figure S3 from Schlaich et al, Lancet 2022
Secondary outcomes
In terms of the key secondary outcome of change in the sitting SBP in part 3 of the study from withdrawal baseline (week 36) to week 40, there was a significant increase with placebo compared to aprocitentan (5.8 mmHg, p < 0.0001) that was maintained up until week 48. There was a similarly significant increase in DBP with placebo compared to aprocitentan (5.2 mmHg, p < 0.0001).
ABPM at the end of part 1 performed similarly, with decrease in the 24h SBP (-4.2 mmHg with the 12.5 mg dose; -5.9 mmHg with the 25 mg dose) and DBP (-4.3 mmHg with 12.5 mg dose; -5.8 mmHg with the 25 mg dose) with aprocitentan treatment. Notably, blood pressure reduction was most pronounced with nocturnal measurements (Figure 3). Likewise, ABPM in part 3 of the study after 4 weeks of withdrawal (comparing weeks 36 and 40) there was a significant increase in SBP (6.5 mmHg) and DBP (6.8 mmHg) in the placebo group compared to the 25 mg aprocitentan group.
Figure 3 from Schlaich et al, Lancet 2022
Change in Proteinuria
Treatment with aprocitentan also yielded remarkable reduction in proteinuria, with reduction of -28% and -31% with the 12.5 mg and 25 mg doses respectively in part 1 of the study. The decrease in proteinuria was maintained in part 2, and there was a significant increase in part 3 of the study after 4 weeks of withdrawal in the placebo group compared to the treatment group.
Safety
The drug was generally well tolerated. The main side effect was fluid retention or edema, which was reported more frequently with aprocitentan in a dose-dependent manner (9.1%, 18.4% and 2.1% in the 12.5 mg, 25 mg and placebo groups respectively in part 1). Of the participants who reported fluid retention, 103 (49%) patients managed this with the addition of a diuretic. A total of 7 patients discontinued the treatment due to edema or fluid retention, and this side effect was more commonly reported among patients with stage 3-4 CKD. There was no mention of participant’s sodium intake or any dietary restrictions.
There were also a total of 11 patients who required hospitalization for heart failure, of which only 1 patient was in the placebo group at the time of admission. While all 11 of these patients had a history of diabetes, only 5 of them had a pre-existing history of heart failure. Two patients ultimately discontinued study treatment due to heart failure.
Of the 11 reported treatment-emergent deaths (of which 5 were related to CVD), none of these were thought to be related to the study treatment. There was no observed hepatotoxity and there was a significant decrease in hemoglobin (-8.0 g/L, -8.5 g/L, -0.4 g/L in the 12.5 mg, 25 mg and placebo groups respectively) associated with aprocitentan. Adverse events are summarized in Table 2 below.
Table 2 from Schlaich et al, Lancet 2022
Discussion
The PRECISION trial found significant improvement in blood pressure control in the setting of resistant hypertension with the addition of the dual ERA aprocitentan. The improvement in both SBP and DBP was sustained over a 48 week period, and the authors demonstrated reversibility in the withdrawal (part 3) phase of the study.
Although a consistent improvement in blood pressure was demonstrated in all parts of the study with the use of aprocitentan, the actual magnitude of the improvement was relatively modest (in the range of -3.7 to -5.9 mmHg over a 4 week period), and was notably less than previously demonstrated in the phase 2 study (Verweij et al, Hypertension 2020). Comparatively, the PATHWAY-2 study demonstrated an average decrease in SBP with spironolactone of -8.7 mmHg compared to placebo after 12 weeks of treatment (Williams et al., Lancet 2015) in the setting of resistant hypertension. Indeed, the magnitude of BP reduction achieved in the aprocitentan arm of the present study was more comparable to the doxazosin and bisoprolol arms in PATHWAY-2 trial. However, we should be cautious about comparisons of the numbers between these two trials given the many differences between the two trials. Apart from the use of office and ABPM in PRECISION (versus home BP in PATHWAY-2), the meticulous lengths to ensure adherence (single pill combination, long screening and run-in periods) and the long follow up (PATHWAY-2 was 12 weeks of therapy in 4 cycles of crossovers), the most important aspect is that the resistant hypertension population is enriched with primary aldosteronism. Thus the effect of spironolactone being greater is quite plausible - though we should remain cautious given this was not a head:head comparison between spironolactone and aprocitentan.
In addition, fluid retention was the most commonly reported side effect, as we have come to expect with this class of medications. Although about half of the participants reporting this side effect were effectively managed with the addition of further diuretic therapy, it is hard to discount the 6 participants who were hospitalized for heart failure in the absence of a baseline history of CHF. In addition, managing fluid retention with another diuretic, for a total of 5 antihypertensives (including a thiazide at baseline), does significantly contribute to the overall pill burden in these patients, and is perhaps not quite feasible. Flozins have also arrived since the design of this study, and much of the patient population included in PRECISION would likely benefit from these for other indications (diabetes, CKD, CHF). SGLT2 inhibitors also have a modest BP lowering effect, and would decrease risk of heart failure, rather than the increased risk observed with aprocitentan (and lower BP in resistant hypertension, see Tran et al, Am J Hyp 2020). It would appear that the use of ERAs requires very close volume status monitoring, particularly in patients with a history of heart failure, which may prove to be difficult to coordinate in day-to-day clinical practice.
Moreover, despite going to commendable lengths to exclude placebo responders prior to randomization with the implementation of the run-in period, it’s worth noting that there was a significant decrease in blood pressure (~ 11 mmHg) in the placebo arm of the study. Is this due to regression to mean? That would be unlikely given the long screening and run-in periods with multiple measurements which excluded those with pseudo-resistant hypertension. Is there a true ‘placebo’ effect or something along the Hawthorne effect - driven by better adherence or other behaviors that may lower BP (in the belief that one is receiving the active drug). Adherence could be one explanation why the ABPM numbers overall are lower than the office BP (since direct observed therapy was done prior to ABPM).
In the present study, the PRECISION trial authors put forth aprocitentan as an important alternative option for resistant hypertension that does not share spironolactone’s risk of hyperkalemia. However, it is difficult to preferentially reach for ERAs such as aprocitentan over spironolactone and other fourth line agents for resistant hypertension given the relative cost, and risk of fluid retention, in this highly comorbid patient population with a high prevalence of CKD, diabetes, obesity and ischemic heart disease. While we do not have available cost data regarding aprocitentan, published data around the use of bosentan in the setting of pulmonary hypertension reveals the generic form’s cost is prohibitively high at $45/day which corresponds to $16 398/year (CADTH 2015).
So, if we could take cost out of the equation, who should be considering using this drug for? Generally speaking, its long half life of 44 hours is certainly attractive for blood pressure control. However, even spironolactone is quite long acting - and unlike aprocitentan, it lowers risk of badness from heart failure rather than increasing it. As mentioned before, resistant hypertension is enriched with primary aldosteronism - and there is little doubt that mineralocorticoid antagonists would be preferable in that subgroup. As the authors suggest, patients with a history of hyperkalemia (particularly those refractory to potassium binding therapies), who may be thus intolerant to potassium-sparing diuretics, could also possibly stand to benefit from aprocitentan for resistant hypertension.
The results of the PRECISION trial suggest aprocitentan might be a reasonable therapeutic in patients who also have a history of proteinuric CKD who might stand to benefit from its antiproteinuric effects. Again, in this population, flozins have established themselves as being deserving candidates with clinical endpoint data. Lastly, aprocitentan could also be an interesting option for patients with a concomitant history of pulmonary hypertension who could potentially benefit from multiple indications for ERAs.
Strengths
This was an exceptionally designed randomized, blinded, controlled, parallel-group study and authors should truly be applauded for their thoughtful study design and execution. The authors systematically attempted to exclude participants with pseudo-resistant hypertension using both a screening period on standardized background therapy and run-in period with the addition of placebo to exclude placebo-responders. They also had stringent monitoring for treatment adherence, limiting compliance bias in this study.
Limitations
It is important to take note of the pharmaceutical company’s significant role in most aspects of this trial, from design, all the way to publication. BP is a surrogate outcome, but we have discussed previously why a hard endpoint trial in this population will be prohibitive (N = 150,000, $3 billion).
What questions remain?
Following the PATHWAY-2 study (Williams et al., Lancet 2015), spironolactone emerged as the guideline recommended preferred fourth line agent for resistant hypertension. Particularly in light of aprocitentan’s side effect of fluid retention and relative cost, future head-to-head comparisons of spironolactone and aprocitentan are needed to help determine whether the latter should be offered as an alternative treatment option in the setting of resistant hypertension. In addition, whether SGLT2-inhibitors may also play a role in the setting of resistant hypertension, and how it compares to ERAs also remains unknown.
ERAs as a class are not going anywhere though (Heerspink et al, CJASN 2022). Despite the RADAR/SONAR setbacks, atrasentan is still in development, and so is Sparsentan (a dual ERA + ARB). Another one is being trialed in combination with a Flozin (ZENITH).
Conclusions
The PRECISION trial has demonstrated that the addition of aprocitentan to a standardized antihypertensive regimen in the setting of resistant hypertension does produce a significant decrease in the SBP and DBP. Whether the decrease demonstrated in this robust study will translate into clinically meaningful outcomes remains to be seen. While this is an exciting novel therapeutic, its side effect profile and relatively modest performance in the resistant hypertension arena compared to existing contenders make us balk at incorporating it into our toolkit at this time.
Summary prepared by Susan Thanabalasingam,
Queen’s University
Kingston, Canada
NSMC Intern 2022
Reviewed by Swapnil Hiremath, Raad Chowdhury, Jamie Willows