Endocrine and haemodynamic changes in resistant hypertension, and blood pressure responses to spironolactone or amiloride: the PATHWAY-2 mechanisms substudies

Abstract

Background: In the PATHWAY-2 study of resistant hypertension, spironolactone reduced blood pressure substantially more than conventional antihypertensive drugs. We did three substudies to assess the mechanisms underlying this superiority and the pathogenesis of resistant hypertension.

Methods: PATHWAY-2 was a randomised, double-blind crossover trial done at 14 UK primary and secondary care sites in 314 patients with resistant hypertension. Patients were given 12 weeks of once daily treatment with each of placebo, spironolactone 25-50 mg, bisoprolol 5-10 mg, and doxazosin 4-8 mg and the change in home systolic blood pressure was assessed as the primary outcome. In our three substudies, we assessed plasma aldosterone, renin, and aldosterone-to-renin ratio (ARR) as predictors of home systolic blood pressure, and estimated prevalence of primary aldosteronism (substudy 1); assessed the effects of each drug in terms of thoracic fluid index, cardiac index, stroke index, and systemic vascular resistance at seven sites with haemodynamic monitoring facilities (substudy 2); and assessed the effect of amiloride 10-20 mg once daily on clinic systolic blood pressure during an optional 6-12 week open-label runout phase (substudy 3). The PATHWAY-2 trial is registered with EudraCT, number 2008-007149-30, and ClinicalTrials.gov, number NCT02369081.

Findings: Of the 314 patients in PATHWAY-2, 269 participated in one or more of the three substudies: 126 in substudy 1, 226 in substudy 2, and 146 in substudy 3. Home systolic blood pressure reduction by spironolactone was predicted by ARR (r²=0·13, p<0·0001) and plasma renin (r²=0·11, p=0·00024). 42 patients had low renin concentrations (predefined as the lowest tertile of plasma renin), of which 31 had a plasma aldosterone concentration greater than the mean value for all 126 patients (250 pmol/L). Thus, 31 (25% [95% CI 17-33]) of 126 patients were deemed to have inappropriately high aldosterone concentrations. Thoracic fluid content was reduced by 6·8% from baseline (95% CI 4·0 to 8·8; p<0·0001) with spironolactone, but not other treatments. Amiloride (10 mg once daily) reduced clinic systolic blood pressure by 20·4 mm Hg (95% CI 18·3-22·5), compared with a reduction of 18·3 mm Hg (16·2-20·5) with spironolactone (25 mg once daily). No serious adverse events were recorded, and adverse symptoms were not systematically recorded after the end of the double-blind treatment. Mean plasma potassium concentrations increased from 4·02 mmol/L (95% CI 3·95-4·08) on placebo to 4·50 (4·44-4·57) on amiloride (p<0·0001).

Interpretation: Our results suggest that resistant hypertension is commonly a salt-retaining state, most likely due to inappropriate aldosterone secretion. Mineralocorticoid receptor blockade by spironolactone overcomes the salt retention and resistance of hypertension to treatment. Amiloride seems to be as effective an antihypertensive as spironolactone, offering a substitute treatment for resistant hypertension.

Funding: British Heart Foundation and UK National Institute for Health Research.

Figure 1

PATHWAY 2 study design Stages in the crossover trial at which measurements for the three mechanistic substudies were taken. Green shows substudy 1, blue shows substudy 2, and pink shows substudy 3. During the 12-week drug cycles, the lower dose was given for the first 6 weeks, then the higher dose was given for the second 6 weeks. No washout period was used between cycles. The figure shows one of the 24 possible passages through the drug cycles, the order of which was randomly assigned within blocks of 24 patients. A+C+D=an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, a calcium channel blocker, and a diuretic.

Figure 2

Participant numbers and measurements in the PATHWAY-2 mechanisms substudies

Figure 3

Correlations of plasma aldosterone, renin, and ARR, with blood pressure response to spironolactone averaged (mean) across the 6-week and 12-week visits of each treatment cycle (A) Relation between baseline plasma renin, aldosterone, and the ARR and the home systolic blood pressure response to spironolactone. (B) Best-fit relation between plasma aldosterone and renin concentrations at baseline. Regression equations for change in systolic blood pressure (y): y=(−25·20)+6·86 × (log₁₀renin), r²=0·116 (proportion of variance accounted for by the model); y=8·92–9·85 × (log₁₀aldosterone), r²=0·034; and y=(−8·87)–6·87 × (log₁₀ARR), r²=0·138. Regression equation for aldosterone vs renin: log₁₀aldosterone=2·60–0·279 × (log₁₀renin) + 0·081 × (log₁₀renin)², r²=0·043. ARR=aldosterone-to-renin ratio.

Figure 4

Change in haemodynamic parameters from baseline after 12 weeks treatment with spironolactone, doxazosin, bisoprolol, and placebo Stroke index, cardiac index, vascular resistance index, and thoracic fluid index were measured at baseline and the end of each double-blind treatment cycle. Least squares means adjusted for gender, height, weight, smoking history, baseline systolic blood pressure, and the baseline measurement of the outcome, from a mixed model allowing for correlations between repeat measurements in each patient. Coloured bars show least squares mean values and black bars show 95% CIs.

Figure 5

Effect of placebo, amiloride, spironolactone, doxazosin, and bisoprolol on clinic blood pressure after 6 weeks of treatment Both amiloride and spironolactone reduced systolic and diastolic blood pressure (unadjusted means) versus placebo (p<0·0001) and versus both doxazosin or bisoprolol (p<0·0001). The horizontal line at 140 mm Hg shows target clinic systolic blood pressure. Coloured bars show mean values and black bars show 95% CIs.

Figure 6

Correlation between change in systolic blood pressure in patients treated with amiloride and in those treated with spironolactone after 6 weeks of treatment Dotted lines show corresponding points on y axis (for amiloride) and x axis (for spironolactone) for mean responses (red) and zero responses (blue).