Choose Wisely – Avoid Sildenafil for Pulmonary Hypertension in Corrected Left-Heart Valvular Disease (SIOVAC Trial)
“It's these expressions, I never give, That keep me searching for a heart of gold …
And I'm gettin' old.”
An 86 year old woman is admitted with progressive dyspnea and lower extremity edema. 8 months ago she received a trans-catheter aortic valve replacement (TAVR) for severe aortic stenosis, yet her symptoms did not abate. 2 months prior to presentation, a follow up transthoracic echocardiogram was performed which revealed a calculated right ventricular systolic pressure (RVSP) of 75 mmHg; she had tissue Doppler indices consistent with diastolic dysfunction and a normal ejection fraction. With this data, her outpatient cardiologist initiated her on sildenafil and slowly titrated her dose upwards. She had no other changes in her medications and she kept to her low sodium diet strictly; nevertheless, her symptoms accelerated and she was admitted with worsening pulmonary edema and a 5 kg weight gain.
“…not routinely offer pharmacologic treatment with advanced vasoactive agents approved only for the management of pulmonary arterial hypertension to patients with pulmonary hypertension resulting from left heart disease or hypoxemic lung diseases (groups 2 or 3 pulmonary hypertension),” the aforementioned clinical vignette seems to occur with increasing frequency. As the authors note, the most common cause of pulmonary hypertension is left heart disease; further, mitral and aortic abnormalities are an exceedingly common cause of left atrial hypertension.
One rationale behind providing pulmonary vasodilators to patients with left heart disease is that – with time – ‘post-capillary’ hypertension (i.e. congestion from left heart disease) results in arteriolar remodeling and, as a consequence, ‘pre-capillary’ pulmonary hypertension. Indeed, hemodynamic data reveals that treating post-capillary to pre-capillary pulmonary hypertension with phosphodiesterase inhibitors leads to acute hemodynamic improvement in the post-operative state. Nevertheless, there is concern that the pre-capillary resistance that develops in response to high left heart pressure is protective and that relieving this resistance augments left heart preload resulting in clinical deterioration.
What They Did
Two-hundred patients were randomized with roughly 100 receiving placebo and the other half receiving sildenafil 40 mg thrice daily – note that patients with a low BMI, hypotension, or those showing severe hypotension during the vaso-reactivity test initiated 20 mg thrice daily for 2 weeks. Patients were included if they demonstrated a mean pulmonary arterial pressure greater than or equal to 30 mmHg by catheterization within a month prior to randomization, had a successful surgical or percutaneous valvular replacement or repair procedure that resulted in complete correction of left heart valve disease and performed at least 1 year before inclusion. Lastly and importantly, the patients had to be in a stable clinical condition which was defined as having no changes in medication or hospital admissions for heart failure in the previous month.
The patients were mostly elderly women who mostly had either isolated mitral repair or mitral and aortic repair. Atrial fibrillation was very common in both groups as was baseline hypertension and the ejection fraction of the two groups were comparable; 85 in each arm completed the study.
What They Found
This was not a positive trial for sildenafil – essentially every metric favoured placebo in all sub-group analyses. While it was not a large trial making statistical power tenuous, there appeared to be a strong clinical trend for deterioration with the phosphodiesterase inhibitor. There was about a 10% absolute risk increase of heart failure exacerbation compared to placebo; this was statistically significant too. Further, of those who reported feeling ‘worse’ or who had deteriorating functional class – 9 were on sildenafil compared to 2 on placebo. By contrast, of those who reported feeling ‘better’ or who had improved functional class – 27 were on sildenafil compared to 44 on placebo!
There was no clinically or statistically significant difference in 6 minute walk, BNP level or pulmonary artery systolic blood pressure. There was, however, a significant increase in left ventricular volume by MRI in those on sildenafil.
Thoughts & Clinical Implications
There is lots to unpack in this trial, more so than can fit in this brief post. Firstly, there is the divorce between acutely favourable hemodynamic effects and longer-term clinical events. Secondly is the inclusion of patients with clear post-capillary pulmonary hypertension and finally are the purported mechanisms of clinical worsening versus improvement in the diverse group of patients with pulmonary hypertension secondary to left heart disease.
One must be exceptionally careful when using pulmonary vascular resistance [PVR] or the trans-pulmonary gradient [TPG] as markers of pulmonary arterial resistance. Simply, the reason is that both the PVR and TPG will rise when the pulmonary arterial systolic [PAs] pressure rises. With passive congestion from the left heart, there is parallel recruitment and dilation of the pulmonary vascular tree. Paradoxically, this decreases classical resistance, but the congested vessels have decreased compliance; that is, the vessels become stiffer. When blood is ejected against stiff vessels, the systolic pressure rises disproportionately which will mathematically increase the PVR and TPG.
One, albeit imperfect, way to avoid this problem is to use the pulmonary arterial diastolic [PAd] pulmonary artery occlusion pressure [Ppao – i.e. the ‘wedge pressure’] difference. In health, the PAd – Ppao gradient should be less than 5 mmHg, and certainly less than 7 mmHg. Thus, a value less than 7 mmHg suggests a normal pulmonary arterial ‘resistance.’
Patients in this trial certainly had Group II pulmonary hypertension. The average Ppao [i.e. ‘wedge pressure’] in both groups on initiation was 22-23 mmHg. Mathematically, both groups had elevated PVR and TPG [3.1-3.4 Wood Units and 15-16 mmHg, respectively]. Nevertheless, both groups had a normal PAd – Ppao gradient [2-3 mmHg] at baseline. Per the reasoning above, the PVR and TPG elevations do not reflect an increase in resistance, rather, they reflect decreased vascular compliance.
With acute vasoreactivity testing [patients were given 100 mg sildenafil sublingually], the PAd – Ppao fell minimally, but the PVR and TPG fell more prominently. While it may be tempting to think that sildenafil lowered vascular resistance based on this data, it seems more likely that sildenafil simply improved pulmonary vascular compliance, either directly and/or by ‘decongesting’ the vascular volume of the lungs. That the cardiac index remained stable or increased slightly with sublingual sildenafil and Ppao fell somewhat, there may have been mild improvement of left ventricular output – potentially by left ventricular afterload reduction.
There is old data – in healthy individuals – demonstrating that increased Ppao results in an acute – perhaps auto-regulatory – reduction in pulmonary vascular compliance; feasibly, this is an adaptive response to regulate pulmonary vascular flow – just as other organs regulate their flow. Might we be destroying this protective mechanism with pharmaceuticals? It brings to mind a thoughtful essay written by Scott Aberegg about the ‘Normalization Fallacy.’ Do abnormal values need to be automatically and thoughtlessly corrected?
Critics of SIOVAC may argue that these patients with elevated Ppao and low PAd – Ppao gradient should not have been studied and their worsened outcomes and enlarged left ventricular volume at 6 months were predictable. These critics may point to papers where sildenafil exerted positive hemodynamic effects such as this 2014 paper by Jiang et al. in post-cardiac surgery patients.
Like SIOVAC, the study by Jiang and colleagues demonstrated that sildenafil improved PVR, but red flags are raised when the authors note a limitation of their study was that they didn’t report the PAd – Ppao gradient. Why not? They had all the data available – why not report this? They did present PA systolic pressure, mean pulmonary artery pressure and the Ppao. Fortunately, you’re in luck; using the power of mathematics, I calculated the PA diastolic pressure from their data and placed it in a table below along with the PAd – Ppao gradient
Table 1: Note that I rounded all data to whole numbers. Rows shaded in blue are calculated - PA diastolic was calculated from reported mean PA pressure and PA systolic pressure. Please note an error in this table - the baseline PAs for the sildenafil group is 49 mmHg, not 42 as listed.
This data notably reveals that the reduction in PVR was driven almost entirely by the fall in PA systolic pressure. The diastolic pressure dropped as well, but it tracked the fall in the Ppao. These patients appeared to have a fixed increase ‘true’ resistance [i.e. the PAd – Ppao was above 7 mmHg in all], but this was not affected by sildenafil. This trial also revealed an improvement in left ventricular stroke volume and stroke work suggesting that some of the benefit was simple LV afterload reduction. If the left ventricle improves, it will shift blood volume out of the pulmonary vascular tree and ‘decongest’ the stiff blood vessels. This is in addition to any direct vascular relaxing properties sildenafil carries.
Thus, the effects of sildenafil on the pulmonary vascular tree are complex – there may be some reduction in true resistance, but sildenafil may also improve vascular compliance and shift blood volume out of the pulmonary bed. The latter two effects will decrease the PVR and TPG by reducing the PA systolic pressure; but don’t let that fool you. Demand the PAd – Ppao gradient and – much more importantly – demand long-term clinical follow-up!