Spironolactone reduces brachial pulse wave velocity and PIIINP levels in hypertensive diabetic patients


Justine Davies, Division of Medicine and Therapeutics, Ninewells Hospital and Medical School, Dundee, UK. Email: j.i.davies@dundee.ac.uk



To assess whether spironolactone has beneficial effects on blood pressure (BP), N-terminal propeptide of type III procollagen (PIIINP) and pulse wave velocity (PWV) in hypertensive, type II diabetics.


Ten patients with type II diabetes and hypertension were enrolled in a randomized, double-blind crossover study comparing 4 months’ treatment with spironolactone and placebo with a 4-week washout phase. BP, PIIINP and carotid-radial PWV were measured at the end of each treatment phase.


Compared with placebo, spironolactone reduced systolic BP by 15.6 ± 46.1 mmHg (P = 0.005, 95% CI 2.7–28.5 mmHg), PIIINP by 0.6 ± 0.3 µg l−1 (P = 0.04, 95% CI 0.02–1.1 µg l−1) and PWV by 0.6 ± 0.2 m s−1 (P = 0.008, 95% CI 0.18–1.02 m s−1).


Spironolactone is effective at reducing systolic BP and brachial artery stiffness as indicated by PWV. It also reduces PIIINP in type II diabetic patients with hypertension.


The incidence of cardiovascular disease among type II diabetic patients remains unacceptably high despite improved control of glucose, cholesterol and blood pressure. Additional approaches are therefore required to reduce the incidence of cardiovascular events in this population.

There has been much recent interest in renin-angiotensin-aldosterone system (RAAS) blockade in diabetes, where both direct angiotensis II (AII) receptor blockade and angiotensin converting enzyme (ACE) inhibition have been shown to have beneficial effects [1, 2]. Aldosterone and AII are known to share a plethora of deleterious cardiovascular effects [3]. It follows that blocking aldosterone receptors may produce similar beneficial effects to withdrawal of AII. This has been the case in heart failure, where after ACE inhibition, aldosterone blockade resulted in further significant reductions in cardiovascular morbidity and mortality [4, 5]. This beneficial effect in heart failure is probably, in large part due to improving myocardial fibrosis and endothelial function [3].

An increase in vascular stiffness as assessed by an increase in aortic pulse wave velocity (PWV) is a recognized prognostic marker in patients with diabetes [6]. Furthermore a positive correlation has been seen between vascular stiffness and aldosterone levels in patients with heart failure [7]. It is possible that aldosterone blockade may improve vascular stiffness and offer prognostic benefits in diabetic patients [8]. However, two recent sets of data suggest that aldosterone blockade may only be useful in patients with hypertension. Firstly, the beneficial effects on mortality in the EPHESUS study was limited to those with hypertension [5]. Secondly, we were recently surprised to discover that spironolactone worsened endothelial function in normotensive diabetics, though this effect was not significant in those patients whose blood pressure (BP) fell with spironolactone [9].

We therefore postulated that spironolactone treatment in a population of type II diabetic patients with hypertension might have a beneficial effect on the marker of vascular stiffness, PWV and on N-terminal propeptide of type III procollagen (PIIINP) (a marker of cardiovascular fibrosis) as well as reducing BP.


Ten patients with type II diabetes and hypertension (Systolic BP >140 mmHg or diastolic >85 mmHg, corresponding to our local guidelines for treatment at the time of the study) were enrolled. Patients were excluded if they had heart failure, potassium above 5.0 m m l−1, creatinine greater than 200 µmol l−1, hepatic failure or sensitivity to spironolactone. The local ethics committee approved the protocol, and all patients gave written, informed consent.

Patients were randomized to receive either 50 mg spironolactone or placebo in a double-blind, cross-over fashion. Spironolactone treatment was initiated at 25 mg day−1 (one tablet). After 2 weeks, if potassium and renal function were within the limits defined above for entry into the study, spironolactone dose was increased to 50 mg day−1 (two tablets). Urea and electrolytes were measured every 2–4 weeks. If values exceeded the limits for entry into the study the spironolactone dose was reduced. There was a 4-week washout period between treatments. All other medications remained unchanged for the duration of the study.

Study parameters were measured at baseline, and after 4 months treatment with placebo and spironolactone. A leader effect was excluded. Results were then analysed by anova.

Blood pressure was taken as the average of three readings, measured with the patients sitting and rested.

Carotid to radial PWV was performed by a trained operator as per manufacturers instructions after a period of 15 min supine bed rest using the Sphygmocor™ system (Scanned Medical, Gloucestershire, UK). The foot of the pulse pressure wave was ascertained using the intersecting tangent method. If necessary, measurements were repeated until a reading of good quality was obtained.

Samples of aldosterone and renin were taken and ana-lysed using radioimmunoassay (RIA) kits supplied by Diasorin Ltd (Diasorin Ltd, Wokingham, UK). PIIINP was analysed by RIA kit (Quidel Ltd, Oxford, UK). Inter assay coefficients of variation were 3.9%, 8% and 4.9%, respectively, for renin, aldosterone and PIIINP.


Patient characteristics are presented in table 1 and outcome parameters are presented in table 2.

Table 1.  Patient characteristics
Age, y66 ± 7.3
Known coronary artery disease3
Known hypertension8
Cholesterol (mM l−1)5.01 ± 0.82
HbA1C (mM l−1)7.57 ± 1.24
ACEI/AII blockade (n)2
Beta-blockade (n)7
diuretics (n) 1
Ca channel blockers (n) 5
Nitrates (n) 1
Statins (n) 4
Aspirin (n) 5
Metformin (n)4
Sulphonylureas (n)4
Rosiglitazone (n)1
Insulin (n)0
Table 2.  Outcome variables after treatment with spironolactone and placebo
  • *

    Indicates significant difference between spironolactone and placebo at the P < 0.05 level.

Systolic BP (mmHg) 145 ± 15 153 ± 22137 ± 21*
Diastolic BP (mmHg) 78 ± 9 79 ± 11 75 ± 7
MAP (mmHg) 100 ± 3103 ± 3 96 ± 3*
Heart rate (bpm) 63 ± 1.6 63 ± 1.6 62 ± 1.6
PWV (m s−1)8.85 ± 1.11 8.7 ± 1.0 8.1 ± 1.1*
PIIINP (ug l−1)3.41 ± 0.9 3.5 ± 1.0 2.0 ± 0.6*
Aldosterone (pg ml−1) 329 ± 191400 ± 223509 ± 355
Renin (ng ml−1 h−1) 0.6 ± 0.7 2.2 ± 3.6 7.1 ± 10.7*
Potassium (m m l−1) 4.2 ± 0.8 4.2 ± 0.8 4.4 ± 0.8*

All patients had aldosterone levels within normal limits. All but two patients had normal renin levels. All patients completed the study.

Seven patients were treated with 50 mg spironolactone daily, two with 25 mg daily and one with 25 mg of spironolactone on alternate days. Rising potassium prevented dose increments in the three patients who failed to achieve a dose of 50 mg day−1.

Renin levels significantly increased on spironolactone by 4.94 ± 2.05 ng ml−1 h−1 (P < 0.05, 95% CI 0.6–9.2 ng ml−1 h−1); aldosterone levels did not significantly increase with spironolactone (P = 0.3).

Systolic BP was significantly lower on spironolactone compared with placebo by 15.6 ± 6.1 mmHg (P = 0.005, 95% CI 2.7–28.5 mmHg). Diastolic BP was also reduced by spironolactone treatment by 3.6 ± 3 mmHg, but this was not significant (P = 0.24, 95% CI −2.65 to 9.08 mmHg)). Spironolactone significantly reduced PWV by 0.6 ± 0.2 m s−1 (P = 0.008, 95% CI 0.18–1.02) compared with placebo. The improvement in PWV on spironolactone remained significant even after change in systolic BP was controlled for in a multivariate analysis (mean difference of 0.51 ± 0.24 m s−1, P < 0.05, 95% CI 0.09–1 m s−1). There was also no correlation between change in BP and change in PWV with spironolactone compared with placebo (r = 0.13, P = 0.73). However when mean arterial pressure (MAP) was controlled for in a multiple regression analysis the decrease in PWV with spironolactone was only of marginal significance (mean difference 0.45 ± 2.25 m s−1, P = 0.05, 95% CI 0.005–0.9 m s−1). The results were similar when MAP2 was used (mean difference of 0.42 m s−1, P = 0.06).

PIIINP levels were significantly lower on spironolactone compared with placebo, by 0.56 ± 0.26 µg l−1 (P = 0.04, 95% CI 0.02–1.1 µg l−1). This change was no longer significant after systolic BP and MAP were controlled for (P = 0.2 for both).

When potassium was controlled for in a multiple regression analysis the decreases in systolic BP and PIIINP seen with spironolactone were no longer significant (P = 0.1 and 0.2, respectively). However, the decrease in PWV still achieved significance (P = 0.05).


We have found that, in a hypertensive diabetic population 4 months’ treatment with spironolactone significantly reduced systolic BP, PWV and PIIINP. The effect of spironolactone on PWV was independent of its effect on BP. We controlled for MAP and MAP2 as these parameters are likely to have more impact on distensibility and PWV than systolic BP. The effect of spironolactone on PWV after controlling for MAP was still significant, but in this small number of patients significance was lost after MAP2 was controlled for. This does, however, suggest that MAP may have contributed to our results.

Spironolactone is a well recognized antihypertensive agent, and may be especially effective in patients in whom aldosterone levels are high [10]. Our study is the first to show that in a normo-aldosteronaemic diabetic population with hypertension spironolactone is effective at reducing BP and this fall in BP could, in part be due to the small increase in potassium seen. But, interpreting the effect of potassium on BP here is difficult as spironolactone's effects are to decrease BP and increase potassium. Therefore, it cannot be said with any certainty that the increase in potassium led to the decrease in BP. The findings of this study are at variance with the findings of our previous randomized, placebo-controlled study where spironolactone (average dose 47.5 mg) worsened the prognostic marker of endothelial function [9]. Although PIIINP and PWV were not assessed in our earlier study, as endothelial function is thought to be the first step in atherosclerosis it follows that worsening in endothelial function would be associated with increased vascular stiffness. However in our previous study spironolactone did not produce a fall in BP (BP was 128/72 mmHg (20/9) during placebo and 128/71 mmHg (27/9) during spironolactone), whereas in this current study spironolactone significantly reduced BP. In our first study when the subgroup of patients who had a fall in BP were analysed separately the effects of spironolactone were not significant. This fall in BP was limited to patients who were not on ACE inhibitors and it should be noted that in this present study the majority of patients were also not on ACE inhibitors. It may be that this decrease in BP protects against the potential adverse effects of spironolactone in these patients.

In this current type II diabetic population spironolactone resulted in a significant reduction in vascular stiffness, as assessed by PWV. Spironolactone also reduced PIIINP, a marker of cardiovascular fibrosis. In heart failure, where PIIINP has been linked to survival outcomes, spironolactone similarly results in a reduction in PIIINP [4]. However, whilst heart failure is known to be associated with increased activity of the RAAS, diabetes is usually associated with normal activity of this system [8]. Indeed our population of diabetic patients had renin and aldosterone levels that were within normal limits. Our findings of a reduction in PIIINP following treatment with spironolactone suggests that even ‘normal’ levels of aldosterone may amplify cardiovascular stiffness in hypertensive diabetics. However, there was no correlation between change in aldosterone levels and PIIINP or PWV in this small number of patients.

We cannot be certain that the beneficial effects on PWV are due to aldosterone blockade rather than to its effects on blood pressure per se, especially as significance was just lost after controlling for MAP2. However, changes in PWV were still significant after controlling for systolic BP and MAP. This is not out of character with other studies that have shown that decrease in BP does not always result in an increase in vessel compliance. Furthermore, it is possible to improve vascular compliance without decreasing blood pressure [11, 12]. It is possible that brachial PWV was improved in our study as a result of the blockade of aldosterones effects on vascular fibrosis [13].

A potential criticism of this study is that we assessed PWV in the brachial artery which is not subject to atherosclerosis. All of the studies demonstrating a prognostic role for PWV have done so by measuring carotid-femoral PWV by alternative methods. However, although the brachial artery is not subject to the atherosclerotic process when endothelial function is assessed in the brachial artery this correlates with that in the coronary circulation, a common site of atherosclerosis. Also endothelial function in the brachial artery is known to be a prognostic marker. It seems reasonable to conclude that despite the lack of frank atherosclerosis, the brachial artery suffers from similar physiological alterations leading to atheroma as other arteries. Although we acknowledge that PWV in the brachial artery may not provide as much information as that assessed in the aorta.

In addition we did not measure plasma viscosity and therefore are unaware whether any change may have affected our results.