The effects of medications on circulating levels of cardiac natriuretic peptides

Abstract

Circulating cardiac natriuretic peptide levels are being used increasingly in a range of clinical circumstances. Since it is evident that drugs used in the treatment of cardiovascular disorders can modulate natriuretic peptide levels, we here review the literature documenting these effects. Diuretics, blockers of the renin‐angiotensin system, vasodilator agents, dopamine‐like agonists, amiodarone, and perhaps allopurinol and statins suppress natriuretic peptide levels, most obviously in heart failure. Beta‐blockers stimulate natriuretic peptide concentrations in hypertensive subjects, whereas in heart failure they have little effect or are stimulatory in the short term and inhibitory with sustained therapy. Digitalis compounds and aspirin tend to increase natriuretic peptide levels, and calcium channel blocking agents have varying effects depending on the individual drug and duration of administration. The effects of other drugs are less clear. Additional information is needed regarding the effects of medications along with dissection of the role of altered cardiac secretion versus changes in plasma clearance as explanation for drug‐induced perturbations in natriuretic peptide concentrations. In the meantime, clinicians need to consider the known effects of medications when interpreting plasma levels of the cardiac natriuretic peptides.

• Atrial natriuretic peptide
• B‐type natriuretic peptide
• heart failure
• hypertension

Introduction

The heart secretes the 28 amino acid atrial natriuretic peptide (ANP) and the 32 amino acid B‐type natriuretic peptide (BNP), both of which are biologically active and circulate normally in low picomolar concentrations. They are physiologically important protectors against volume overload and excessive vasoconstriction, acting in concert with other vasodilator/natriuretic/growth‐inhibitory systems to counterbalance contrary systems, most obviously, but not exclusively, the renin‐angiotensin system 1,2. These two bioactive cardiac peptides are cleaved from the carboxy‐terminal end of pro‐hormones (proANP and proBNP) yielding also N‐terminal fragments (98‐amino acid NT‐proANP and 76‐amino acid NT‐proBNP) which have no identified biological activity. Bioactive ANP and BNP are presumed to be secreted in a 1:1 ratio with their respective NT‐pro fragments, but the latter circulate at generally higher concentrations because, presumably again, their elimination rates from the circulation are slower.

As with all circulating hormones, their concentrations in blood reflect rates of secretion and clearance. In regard to secretion rates, stretch of the cardiac chambers (or strictly, transmural pressure) is the dominant stimulus 3,4 although additional factors are capable of modulating secretion. These include the renin‐angiotensin and sympathetic systems (alpha‐1‐agonists are generally seen to be stimulatory and beta‐agonists inhibitory), endothelin‐1, cytokines and growth factors, thyroid hormone, arginine vasopressin, thrombin, glucocorticoids, prostaglandins, and hypoxia/ischaemia 1.

The effect of medications on the secretory rate of natriuretic peptides will depend on change in the sum of the numerous input signals listed above, but dominated by alterations in stretch of the cardiac chambers. For example, angiotensin‐converting enzyme (ACE) inhibitors, when used in the treatment of severe grades of heart failure, reduce stretch of the cardiac chambers while also lowering plasma and cardiac angiotensin II levels and probably reducing sympathetic ‘traffic’ to the heart—all of which should contribute to a decline in the cardiac secretion rate of the natriuretic peptides. The situation is more complex for the beta‐blockers when used in heart failure. In the short term, the beta‐blockers may augment cardiac secretion of the natriuretic peptides concomitant with probable increased stretch of cardiac chambers together with blockade of the inhibitory effect of beta‐agonism, notwithstanding a fall in stimulatory angiotensin II levels 5. Clearance from plasma of biologically active ANP and BNP may be reduced by beta‐blockers 6 thereby contributing to the rise in their circulating levels. By contrast, with chronic beta‐blocker therapy and remodelling of the cardiac chambers (including reduced stretch and a decline in left ventricular mass), cardiac secretion of the cardiac peptides is likely to decline despite the above‐mentioned inhibitory action of these drugs on the plasma clearance rates of ANP and BNP 6. These issues are discussed more fully below.

Clearance of bioactive ANP and BNP from the circulation is accomplished through two pathways, the enzyme neutral endopeptidase 24.11 (neprilysin) and clearance receptors, although binding to their bioactive type A receptors presumably contributes to a small extent. Little is known regarding the clearance mechanisms for the NT‐propeptides and, accordingly, there is a dearth of information on the possibility that medications alter their clearance characteristics.

Circulating levels of the cardiac natriuretic peptides are of scientific interest and of actual, or potential, clinical use in a number of circumstances. First, their levels are of proven benefit in the diagnosis of heart failure for patients presenting with acute dyspnoea. Elevated values are consistent with a diagnosis of heart failure under these circumstances, and normal levels are powerful negative predictors of that diagnosis. Second, their levels in the circulation provide a robust prognostic index, either alone or when combined with additional prognostic factors in patients with established heart failure or post‐myocardial infarction, and also in patients with end‐stage chronic renal failure. Third, their levels hold the potential to function as relatively simple and cheap screening tests for cardiovascular disease, particularly left ventricular dysfunction, in predisposed subjects (diabetics, hypertensives, and the frail elderly, for example) and possibly in the general population. Fourth, guiding the intensity of drug treatment for patients with chronic heart failure according to circulating levels of these peptides has intuitive appeal, the results of the two completed studies are encouraging, and additional, larger studies are underway. Fifth, measurement of cardiac peptide levels might prove clinically useful in a variety of other clinical circumstances including the assessment of severity and progression of valvular heart disease and amyloid heart disease, the early detection of cardiotoxic doses of antimitotic drugs, as one determinant of myocardial status post‐cardiac transplantation, and in guiding the management of patients with end‐stage chronic renal failure.

Whereas elevated circulating levels of the cardiac natriuretic peptides are best known to be associated with heart failure based on systolic and/or diastolic left ventricular dysfunction, interpretation requires realization that their levels increase with age and as renal function deteriorates, are higher on average in females than males, are elevated by tachyarrhythmias, hypertension (especially with concomitant left ventricular hypertrophy and diastolic or systolic dysfunction) and hyperthyroidism, and tend to be lower in the obese and in hypothyroidism.

Apart from these factors, numerous articles attest to the fact that medications, especially those used to treat cardiovascular disorders, can alter levels of the cardiac peptides. Apart from exceptional circumstances, mitral stenosis for example 7,8, levels of the four peptides (ANP, NT‐proANP, BNP and NT‐proBNP) are closely correlated and almost invariably rise or fall in parallel even if their rates of change can differ 9. Accordingly, we will assume in this review that studies reporting drug‐induced perturbations in levels of one peptide are likely to reflect changes in the other three peptides. Nevertheless, it has been observed by a number of workers that the B‐type natriuretic peptides more closely relate to haemodynamic indices and prognosis in cardiac disorders than the two A‐type peptides. Furthermore, in contrast to assays for ANP and NT‐proANP, well validated commercial BNP and NT‐proBNP assays are in wide‐spread, and increasing 10, clinical use. We will tend, therefore, to emphasize B‐type natriuretic peptide responses to medications used to treat cardiovascular disorders whilst acknowledging studies reporting ANP and NT‐proANP levels—which reflect, by and large, the discovery of ANP before BNP and hence the earlier development of assays to measure the A‐type peptides. In this review paper we will restrict our attention largely to studies carried out in man and ignore the abundant studies performed in experimental animals, isolated tissues, cells, and organs. Since drugs which inhibit the enzyme neutral endopeptidase 24.11 along with ACE, the so‐called vasopeptidase inhibitors, are not in clinical use, we will not discuss their reported effects on circulating natriuretic peptide levels.

Key messages

• Interpretation of circulating levels of cardiac natriuretic peptides requires awareness that many medications can alter their concentration through changes in secretion and/or clearance.

• Diuretics, blockers of the renin‐angiotensin system, vasodilator drugs, dopamine‐like agonists, amiodarone, and perhaps allopurinol and statins reduce natriuretic peptide levels, whereas digitalis compounds, aspirin, oestrogens, and some immunosuppressive drugs tend to be stimulatory, and calcium channel blockers have varying effects.

• The possibility that medications other than those noted above also modify plasma natriuretic peptide levels demands further study.

Diuretics

Healthy subjects, essential hypertension

Lijnen et al. reported a substantial (59%) decline in plasma ANP levels in healthy men receiving the chlorthalidone‐like diuretic, xipamide, for 16 weeks, most obvious after 1 week 11. By contrast, patients with essential hypertension were reported by Chalmers et al. to exhibit a rise in plasma ANP levels (from a mean of 92 to 145 pg/mL) over a 4‐week period of monotherapy with hydrochlorothiazide which reduced blood pressure significantly 12. In a study of hypertensive patients on chronic diuretic therapy alone (n = 8) or in combination with other antihypertensive drugs (n = 54), plasma ANP levels were not different from those in untreated hypertensive patients 13.

Heart failure

BNP levels in patients with severe, refractory heart failure fell by more than 50% with intravenous frusemide treatment (with or without concomitant hypertonic saline) over 6 days in association with a natriuresis, and a decline in arterial pressure and heart rate 14. Diuretics combined with nitroprusside reduced ANP, NT‐proANP and BNP levels significantly (by 26%–52%) and promptly (over 1.4 days) in patients with advanced, decompensated heart failure who were removed temporarily from their maintenance ACE inhibitors 15. The effects of diuretics versus nitroprusside in this study are impossible to separate. An apparently contradictory observation by Starklint and colleagues that a single 80‐mg dose of frusemide in patients with apparently stable chronic heart failure was without effect on plasma ANP and BNP levels 16 is not unexpected if, as is likely, daily maintenance diuretic administration elicits little immediate haemodynamic effect under these circumstances.

Addition of aldosterone receptor blockade to conventional medications (including beta‐blockers) for moderate to severe grades of heart failure, results in a fall in circulating natriuretic peptide levels. For example, in patients with New York Heart Association (NYHA) functional class III–IV heart failure associated with a low left ventricular ejection fraction, spironolactone 12.5–50 mg/day reduced plasma BNP levels 23% at 3 and 6 months in the Randomized ALdactone Evaluation Study (RALES) 17. Spironolactone compared to placebo, when added to conventional drug therapy (including beta‐blockers in one‐third of patients), reduced plasma ANP and BNP levels by 39%–55% along with left ventricular mass and volumes over 4 months in 20 patients with mild‐to‐moderate non‐ischaemic heart failure and a reduced left ventricular ejection fraction 18. A similar outcome for patients with moderate degrees of heart failure was reported with 6 months of spironolactone therapy versus placebo by Feola et al. in which BNP values declined by more than 50% 19. For patients with mild heart failure on optimal medical therapy, a statistically significant, if small, fall in BNP was documented over 3 months of spironolactone (12.5–50 mg/day) treatment versus placebo 20. Two studies report no significant effect. The first, by Roonsritong et al., showed that plasma BNP levels were not altered by the addition of spironolactone (25 mg daily) versus placebo for 4 months in elderly patients with mild isolated diastolic left ventricular dysfunction, most of whom were hypertensive 21. In the second study, Kinugawa et al. reported that spironolactone (25 mg/day) had a statistically insignificant suppressive effect on post‐exercise ANP levels in a small cohort (n = 9) of patients with heart failure (NYHA functional class II and III, left ventricular ejection fraction (LVEF) 10%–59%) receiving frusemide and low‐dose maintenance enalapril 22.

In summary, what little information is available regarding the effects of diuretics on circulating natriuretic peptide levels in healthy subjects or patients with essential hypertension suggests there may be an initial decline followed by return to, or close to, baseline values. The isolated report that hydrochlorothiazide can stimulate levels of ANP in essential hypertensives 12 needs confirmation. The situation in heart failure is clearer. Diuretics, whether loop or aldosterone receptor blockers, reduce circulating natriuretic peptide levels, more obviously in patients with severe grades of heart failure and evidence of fluid retention than in those with mild cardiac failure. This presumably reflects, to a large extent, a decline in cardiac natriuretic peptide secretion in parallel with reduced pre‐load and stretch of the cardiac chambers. It is not known whether alterations in plasma clearance of the peptides contribute to (or perhaps limit) the fall in peptide levels. For patients who are stabilized on therapy, administration of their daily maintenance dose of diuretic may have little effect on natriuretic peptide levels on that day. Likewise, introduction of diuretics in those with mild degrees of left ventricular dysfunction may show little or no response in natriuretic peptide levels.

Beta‐adrenergic blockers

Of all the drug groups studied, beta‐blockers have received the most attention. The results have often varied between studies, but patterns have, nevertheless, emerged over time.

Normal subjects, hypertension, coronary artery disease

The accumulated evidence is that beta‐blockers, whether selective or non‐selective, with or without intrinsic sympathomimetic activity, with or without α1 blocking activity, whether given as a single dose or chronically, evoke a rise in circulating levels of the natriuretic peptides, more obvious with exercise than at rest (Table I).

Table I. Studies reporting a stimulatory effect of beta‐blocker drugs on circulating levels of cardiac natriuretic peptides in healthy volunteers, or patients with coronary heart disease or hypertension.

Author (reference)	Subjects	n	Beta‐blocker	Duration	Change in:
					ANP	NT‐proANP	BNP	NT‐proBNP
Nakaoka 23	EH	31	any	‘Long term’	↑×2 ^a
	EH	10	atenolol	4 weeks	↑×2 ^b
Thamsborg 24	Normals	6	atenolol	2 hours	↑ R&E ^c
Bouissou 25	Normals	9	propranolol
			pindolol	3 days	↑ R&E ^c
Donckier 26	Normals	6	propranolol	4 days	↑ E 130% ^c
Deray 27	Normals	9	tertatolol
			atenolol	single dose	↑ E×2 ^c
Tsai 28	Normals	6	propranolol	3 days	↑ R&E ^c
Colantonio 29	EH	10	atenolol	7 days	↑ ^b
Kohno 30	EH	7	atenolol		↑ NS ^c
		9	carteolol	1 hour	↑ ^c
Berlin 31	Normals	8	tertatolol	single dose	↑ E 35% ^c
Riley 32	Angina	16	epanolol	4 weeks	↑ E×2 ^c
Elijovich 33	EH	9	atenolol	10 weeks	↑ ^b
Luchner 34	EH	80	various	chronic	↑ ^a		↑ ^a
Papadopoulos 35	EH	23	celiprolol	30 days	↑ 190% ^b
Deary 36	EH	30	bisoprolol	6 weeks			↑×3 ^c
van den Meiracker 37	EH	24	bisoprolol	8 weeks	↑ 93% ^c	↑ 83% ^c	↑ 148% ^c
Marie 38	CAD	55	various	chronic	↑ E ^a		↑ E ^a
Olsen 39	EH	91	atenolol	4 years		↑ ^b		↑ ^b

EH = essential hypertension; CAD = coronary artery disease; R = resting; E = exercise; NS = not statistically significant.

^a Comparator group was untreated controls.

^b Compared with baseline levels, before beta‐blockade, within the same group.

^c Comparison is within the same subjects but without beta‐blocker administration (usually during placebo administration).

A few studies, not shown in Table I, have reported lesser stimulatory effects of beta‐blockade on natriuretic peptide levels. Dahlof and colleagues showed that levels of three peptides, ANP, BNP, and NT‐proBNP, rose slightly, though not statistically significantly, across 36 weeks of atenolol monotherapy in 99 patients with essential hypertension and left ventricular hypertrophy 40. Although there was no time‐matched placebo group in that study, there was a parallel group that received the angiotensin receptor blocker (ARB), losartan, in which levels of the three peptides declined significantly. The difference between atenolol and losartan groups regarding change in the circulating peptide levels was statistically significant for each of the three peptides. Similar results were reported by Davies and colleagues for 17 patients with essential hypertension randomized to atenolol and losartan each for 4 months in a randomized, cross‐over study in which BNP levels were significantly higher on atenolol than losartan 41. Differences between the groups in these two studies probably reflect a minor effect of atenolol to increase natriuretic peptide levels combined with a clearer effect of losartan to reduce them although, in the absence of a placebo‐control group, this remains speculative.

Of the far fewer studies reporting a suppressive action of beta‐blockers on natriuretic peptide levels, two were in subjects receiving beta‐blockers with β2 stimulating or α1 blocking activity. The first, by Omvik and colleagues, demonstrated a 27% decline in resting ANP levels (from 20.5 to 15 pmol/L) at 2 hours after the first dose of carvedilol but no change thereafter in supine, sitting or exercise‐related levels of ANP after 9 months of carvedilol monotherapy 42. In that there was no time‐matched placebo group, however, the initial 27% decline in ANP cannot be attributed confidently to carvedilol. The second study, by Bohlen et al., reported that celiprolol administered to healthy volunteers for 3 weeks reduced resting plasma levels of ANP and attenuated the rise in ANP associated with a glucose load 43. Again, however, interpretation is hindered by the absence of a time‐matched placebo‐control group. Of other studies reporting little or no effect of beta‐blockers, Kantola et al. showed, in 15 healthy volunteers, that a single dose of atenolol, propranolol or pindolol did not alter the response of plasma ANP to exercise compared to placebo, but pindolol enhanced slightly the post‐exercise decline in ANP 44. Legault et al. reported that administration of propranolol 80 mg had no effect on the ANP response to intravenous volume loading in healthy volunteers 45.

In summary, the vast majority of studies report that beta‐blockers increase circulating levels of the cardiac natriuretic peptides in healthy subjects and patients with essential hypertension or coronary artery disease. What might explain this stimulatory effect of beta‐blockers? Luchner and Schunkert, reviewing data from studies in animals and man, concluded that two mechanisms are likely be involved, one increasing cardiac production of the peptides, the other inhibiting their plasma clearance 6. First, withdrawal of beta‐receptor‐mediated inhibition of cardiac natriuretic peptide release from the heart would leave stimulatory, alpha‐receptor effects of cardiac sympathetic innervation, unopposed 6,46. In addition, Olsen et al. 39 surmised, as have others 31, that beta‐blockers might augment cardiac production of the peptides through a bradycardia‐associated increase in diameter of the atria and ventricles which thereby heightens cardiac wall stress. Second, a beta‐blocker‐related decrease in extracardiac clearance of ANP and BNP through down‐regulation of natriuretic peptide clearance receptors 47,48 would, other factors being equal, increase circulating ANP and BNP levels and, in the presence of increased cardiac secretion, exaggerate the rise in their circulating concentrations. The latter mechanism cannot be invoked to explain beta‐blocker‐induced increases in NT‐proANP and NT‐proBNP, however, since these two peptides are almost certainly not cleared by natriuretic peptide clearance receptors.

Heart failure

Early studies reported that beta‐blockers, when added to conventional maintenance therapy, stimulated an increase in circulating levels of the natriuretic peptides in the short term. For example, Sanderson et al. 49 showed that oral metoprolol 6.25 mg increased ANP and BNP levels at 5 hours and 24 hours (although the vasodilating beta‐blocker, celiprolol, suppressed both ANP and BNP, significantly so at some time points). The RESOLVD investigators demonstrated that controlled‐release metoprolol compared with placebo increased both NT‐proANP and BNP levels at 24 weeks in patients with NYHA functional class II–IV heart failure and a left ventricular ejection fraction (LVEF) <40% 50. Davis and colleagues 5 later reported that metoprolol increased significantly all four cardiac peptides (ANP, NT‐proANP, BNP and NT‐proBNP) at 6 weeks in a randomized, balanced, controlled, parallel‐group design (Figure 1). This paper, which investigated also the effects of metoprolol on secretion and clearance rates of ANP and BNP, will be discussed further below.

Figure 1. Plasma cardiac natriuretic peptide levels(mean±SEM) in 16 patients with congestive heart failure before (pre‐treatment) and 6 weeks after the introduction of metoprolol (closed symbols) or unchanged treatment (open symbols). From reference 5 with permission.

In contrast to the above observations, other reports, more numerous than those already quoted, are that natriuretic peptide levels decline when beta‐blockers are added to conventional pharmacotherapy for heart failure. It is noteworthy that in most, if not all 51, of these latter studies, the effects of beta‐blockers were assessed over a period of months 52–64, rather than days or weeks. The cumulative results suggest, therefore, that beta‐blockers often, though not invariably, have a biphasic effect on circulating natriuretic peptide concentrations, increasing their levels or evoking little change over days and weeks, but decreasing them in the longer term. The study by Fung et al. 60, in which peptide levels were measured before, then serially after introduction of beta‐blockade, illustrates well this biphasic effect. In that study, 49 patients with chronic heart failure, mean LVEF 26%, receiving maintenance diuretic, ACE inhibitor and digoxin therapy, were randomized to metoprolol or carvedilol therapy. Up to 4 weeks, ANP and NT‐proBNP levels were unaltered from baseline values, but by 12 weeks both peptides had fallen significantly and were lower again at 52 weeks (Figure 2). Formal statistical analysis of results from all reported studies show a trend for a log‐linear relationship between duration of beta‐blocker treatment and percentage change in natriuretic peptide levels from baseline, most obvious for ANP (Figure 3) and NT‐proBNP. Statistical significance (P<0.05), however, was not achieved for any of the four cardiac natriuretic peptides.

Figure 2. Mean plasma levels of NT‐proBNP (N‐BNP) in 49 patients with chronic heart failure given either metoprolol (Meto) or carvedilol (Carv) for 52 weeks. From reference 60 with permission.

Figure 3. Percentage change from baseline(taken as 100%) in mean plasma levels of ANP with the introduction of beta‐blocker therapy in patients with heart failure according to duration of treatment with the beta‐blocker (range 5 hrs–20 months). Data were taken from all available studies reported in the literature. Although the percentage change in ANP appeared to relate to duration of therapy, the log‐linear association did not reach conventional statistical significance (P = 0.061).

It seems evident (and unsurprising) that changes in natriuretic peptide levels relate predominantly to the haemodynamic response to beta‐blocker therapy. Thus, for example, Fujimara and colleagues reported that plasma ANP and BNP levels fell in patients who ‘responded’ to 3–5 months of treatment with carvedilol or metoprolol and, furthermore, correlated with concomitant changes in LVEF 65. Kawai et al. similarly noted positive correlations between changes in haemodynamics and BNP levels across 6 months of carvedilol therapy 53 as did Hara and colleagues during metoprolol administration for 6 months 52. Numerous other authors have reported congruent haemodynamic and natriuretic hormone changes with beta‐blocker therapy in patients with chronic heart failure.

One factor which, seemingly, determines the magnitude of the natriuretic peptide response is whether beta‐blocker therapy is initiated before or, more conventionally, after ACE inhibition. Sliwa et al. 61 demonstrated that the effect of carvedilol to reduce plasma NT‐proBNP levels at 6 and 12 months in patients with newly diagnosed idiopathic cardiomyopathy (NYHA functional class II–IV and LVEF <40%) was much greater when the beta‐blocker was started before the ACE inhibitor perindopril, compared to when perindopril treatment was initiated first (Figure 4).

Figure 4. Impact on plasma levels of NT‐proBNP of initiating treatment with carvedilol (BB) before perindopril (ACEI) versus perindopril before carvedilol in patients with newly diagnosed heart failure (n = 38 and 40, respectively). In each case, the alternative drug was added after 6 months of treatment with the initial drug. *P<0.0005 versus baseline data; †P<0.01 versus change from baseline in the ACEI‐first group. From reference 61 with permission.

Finally, the magnitude of haemodynamic dysfunction and level of natriuretic peptides at baseline may be, again unsurprisingly, another determinant of natriuretic peptide response. In this regard, Yoshizawa and colleagues showed that BNP levels declined with 16 or more weeks of carvedilol or metoprolol therapy in patients with heart failure (NYHA functional class II–IV, LVEF <40%, and taking diuretics, ACE inhibitors or ARBs, and digoxin in most cases) who had the highest baseline BNP levels but increased in those who were in the lowest quartile for BNP at baseline 63.

What factors determine these beta‐blocker‐induced biphasic effects on circulating natriuretic peptide levels in heart failure? Davis et al. showed that the introduction of metoprolol increased the secretion rate of ANP, NT‐proANP, BNP, and NT‐proBNP at 6 weeks (when intracardiac pressures were likely to have risen) while also reducing the clearance rate of bioactive ANP and BNP from the circulation 5. As discussed already, longer‐term beta‐blocker therapy, whilst possibly still reducing the clearance rate of ANP and BNP, would be expected to inhibit cardiac secretion of the four peptides as cardiac remodelling and improved myocardial functioning develops—notwithstanding withdrawal of any direct beta‐adrenergic inhibitory action on natriuretic peptide secretion from the heart.

End‐stage renal failure

Hara and colleagues 66 showed a marked fall in plasma ANP and BNP levels across 4 months of treatment with metoprolol in 14 patients with dilated left ventricles on chronic haemodialysis (Figure 5). End‐systolic and end‐diastolic left ventricular dimensions decreased over the same time interval and, presumably, accounted for the fall in ANP and BNP levels. If, as in heart failure, beta‐blockade reduced the clearance rate of the two bioactive peptides from the circulation, the decline in cardiac secretion of ANP and BNP may be greater than was reflected by the fall in their plasma levels.

Figure 5. Plasma levels of norepinephrine and natriuretic peptides before and after 4 months of treatment with metoprolol in 14 patients on chronic haemodialysis. Changes in norepinephrine were not statistically significant. From reference66 with permission.

Angiotensin‐converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs)

Essential hypertension

Of the reports in the literature, most document a fall in circulating natriuretic peptide levels during chronic ACE inhibitor or ARB therapy in patients with essential hypertension 39,40,67–75. In almost all studies, blood sampling was carried out with the patients supine or sitting, but Januszewicz et al. documented also that the exercise‐induced rise in ANP was attenuated by 6 months of captopril therapy 69. Similarly, Dessi‐Fulgheri and colleagues showed that 4 weeks of benazapril therapy not only reduced ANP to 60% of baseline values but also blunted the ANP response to a 1‐litre saline infusion 67. That both A‐type and B‐type natriuretic peptides respond similarly to these drugs is shown in the study of Dahlof where ANP, BNP, and NT‐proBNP levels all declined across 36 weeks of losartan treatment 40 and in the report by Olsen et al. of a decline in both NT‐proANP and NT‐proBNP during losartan administration 39. The combination of an ACE inhibitor and ARB was shown by Anan and colleagues to have a greater BNP‐lowering effect than either drug alone 75. That study, along with many of the others quoted above, involved patients with left ventricular hypertrophy amongst whom natriuretic peptide levels are commonly elevated. Some workers have recorded a statistically significant positive association between the treatment‐related declines in left ventricular mass and plasma natriuretic peptide levels 72,74.

In contrast to the above observations, a small number of studies have documented no change in circulating natriuretic peptide levels in association with ACE inhibitor or ARB therapy 12,13,33,37,41. Of these, only the study of Davies et al. was in patients with left ventricular hypertrophy, and in this small cohort (n = 17) plasma levels of BNP did fall (from 17.9±9 (SD) to 10.9±7.4 pg/mL), albeit not statistically significantly, across 4 months of treatment with losartan 41.

Chronic heart failure

These drugs, which inhibit the formation or actions of angiotensin II, reduce circulating levels of the natriuretic peptides in patients with heart failure secondary to reduced systolic function of the left ventricle 61,76–91. This is so after the first dose of drug 76,78, at least for patients with severe grades of heart failure and disturbed central haemodynamics, and also in the longer term with maintenance therapy 61,79,81–83,85,87. The fall in the A‐type natriuretic peptides with the first dose of ACE inhibitor is faster than for the B‐type peptides 78, a phenomenon seen also with acute vasodilator administration in heart failure (vidae infra). Since the suppressive effect of these drugs on natriuretic peptide levels is dose‐dependent 85,86,92, it is no particular surprise that two studies by Kasama et al., in which a low dose of enalapril (5 mg daily), given for 6 months to patients with NYHA functional class II or III heart failure, altered neither haemodynamic indices nor plasma BNP levels whereas comparator drugs (perindopril 2 mg daily and valsartan 80 mg/day) improved haemodynamics and reduced BNP levels 90,91. This 5‐mg daily dose of enalapril is well below that used commonly in clinical practice and is considerably lower than the mean dose (18.4 mg/day) shown in the CONSENSUS study to increase longevity in patients with NYHA functional class IV heart failure 93.

Does addition of an ARB to maintenance ACE inhibitor therapy for heart failure alter natriuretic peptide levels? Latini and colleagues showed that the ARB, valsartan, reduced plasma BNP levels in patients in the Valsartan Heart Failure Trial (Val‐HeFT) whether or not they were on maintenance ACE inhibitor and/or beta‐blocker therapy 82. Of course it is probable that the outcome depends upon the maintenance dose of ACE inhibitor since, as mentioned already, there is a dose‐response relationship between ACE inhibitors and natriuretic peptide levels.

With regard to possible mechanisms to explain the above observations, Crozier et al. demonstrated, in the earliest of these studies, that the first dose of an ACE inhibitor, ramipril, induced parallel falls in plasma ANP, plasma angiotensin II, arterial pressure, pulmonary artery diastolic pressure and right atrial pressure 76. Longer‐term therapy with an ACE inhibitor or ARB likewise induced concordant changes in haemodynamic indices (including left ventricular volumes) and both ANP and BNP levels in the RESOLVD study 94. These and subsequent data point to the probability that treatment‐induced changes in transmural pressure in the cardiac chambers, perhaps contributed to also by the fall in levels (or actions of) angiotensin II 95,96, a known stimulator of natriuretic peptide synthesis, accounted for the decline in circulating natriuretic peptide levels. A prompt decline in plasma angiotensin II concentrations and perhaps withdrawal of a high level of sympathetic ‘traffic’ to the heart with initiation of ACE inhibitor therapy might account for the temporal dissociation between an early fall in plasma ANP and BNP levels and a later reduction in left ventricular diastolic diameter reported by Yoshimura et al. 85.

Whether ACE inhibition slows the rate of clearance of ANP and BNP from the circulation is open to dispute. If so, the decline in cardiac secretion of the biologically active natriuretic peptides might be greater than is reflected by the fall in their plasma levels during ACE inhibitor therapy.

Calcium channel blockers

The effect of calcium channel blockers on natriuretic peptide levels seems to depend on the individual drug, its duration of administration and possibly, the disorder being treated.

Verapamil

Verapamil given orally or intravenously as a single dose has little or no effect on resting ANP levels in normal volunteers or in patients with essential hypertension or coronary heart disease 97–99 although it reportedly impairs the rise in ANP induced by rapid right ventricular pacing 97 or by an intravenous saline load 100. With longer‐term administration, verapamil has generally been found to stimulate plasma ANP levels in healthy volunteers and patients with essential hypertension 101–104.

Amlodipine

Amlodipine, in contrast to verapamil, has been reported, from well designed studies, to suppress circulating levels of ANP, NT‐proANP, and BNP over 2–6 weeks in essential hypertensives 105,106 and over 2 and 26 weeks in patients with heart failure due to non‐ischaemic cardiomyopathy 107. Additional studies have shown either no effect of amlodipine when added to ACE inhibitor therapy, or a statistically insignificant suppressive action on plasma ANP levels in essential hypertensives 108,109.

Nifedipine

Nifedipine given acutely stimulates a rise in ANP levels in normal subjects and essential hypertensives 98,99,110. No effect on ANP levels was seen by Colantonio et al. in ten patients with essential hypertension receiving nifedipine over 7 days 29. When given for 14 days to healthy volunteers or patients with Raynaud's phenomenon, nifedipine suppressed the rise in ANP induced by a cold‐pressor test 111. Monotherapy for essential hypertension with nifedipine for 4–6 months reduced ANP concentrations by approximately 40% 70,112. It appears, therefore, that nifedipine stimulates natriuretic peptide levels in the short term (possibly through activation of the sympathetic nervous system) but with chronic administration, at least in essential hypertension, it has a suppressive action.

Diltiazem

Diltiazem 60 mg, given orally, increased plasma ANP levels by approximately 80% at 90 and 120 minutes in seven healthy volunteers 98. Four weeks of diltiazem therapy, 180 mg daily, altered neither resting nor exercise‐induced increments in plasma ANP in 16 patients with angina pectoris 32. In patients with myocardial infarction and reperfusion therapy, diltiazem compared to placebo reduced plasma BNP levels over 4 weeks 113.

Others

Felodipine, like nifedipine, increased ANP levels over 2 and 24 hours in patients with essential hypertension 99. Again, as with nifedipine, felodipine tended to suppress ANP levels with longer‐term administration in patients with essential hypertension 114. It also attenuated the rise in ANP levels over 3 and 12 months in patients with heart failure 115.

Nitrendipine 20 mg daily reduced ANP levels in plasma by 16% over 6 weeks in patients with hypertension and type 1 diabetes mellitus 116.

Chronic intake of calcium channel antagonists, alone (n = 17) or in combination (n = 28) for hypertension was associated with higher ANP levels than in untreated patients 13. Interpretation of these data is hampered by pooling of data from patients taking different agents within the group of calcium channel blockers.

In summary, it is difficult to perceive a uniform pattern of response of the natriuretic peptides to calcium channel antagonists. Interpretation of much available data is compromised by the absence of time‐matched placebo administration. It appears, nevertheless, that chronic administration of verapamil stimulates natriuretic peptide levels whereas amlodipine, nifedipine, felodipine, nitrendipine, and perhaps diltiazem, have a suppressive action.

Vasodilators

In patients with heart failure, especially those in whom central haemodynamics are severely perturbed, oral or intravenous vasodilators (nitroglycerine, nitroprusside, and nicorandil) reduce plasma natriuretic peptide levels 9,15,117–121. Plasma ANP and NT‐proANP concentrations fall promptly and in parallel with pulmonary artery wedge and right atrial pressures but, according to Larsen and colleagues, the B‐type peptides respond less briskly, paralleling the slower decline in systemic vascular resistance 9. Nitroprusside, which reduced arterial pressure levels to baseline in healthy volunteers receiving noradrenaline infusions, reduced also ANP to baseline values 122.

Alpha‐adrenergic blockers

Of the few studies addressing the effects of alpha‐adrenergic blocking drugs on natriuretic peptide levels in healthy volunteers or hypertensive patients, most report little or no change in ANP and BNP either at rest or with exercise following single‐dose or sustained (4–6 weeks) therapy 106,123–125. Kohno et al, however, demonstrated a decline in plasma ANP levels at rest and with exercise in 26 elderly hypertensive patients after acute prazosin administration 30. A similar outcome was observed in a study of eight healthy volunteers who, after a single dose of prazosin 1 mg, exhibited a lesser response in plasma ANP to exercise than with placebo although the difference was not statistically significant 31. One contrary report, that terazosin therapy over 30 days stimulated plasma ANP levels slightly (by 16.4%) in patients with hypertension and prostatic hypertrophy, is not readily interpreted in the absence of a time‐matched control group 126. This might be especially pertinent since maintenance medications were withdrawn 2 weeks prior to the study and this alone may have resulted in a ‘moving baseline’ for ANP levels.

In summary, the sparse literature available suggests that alpha‐adrenergic blockade has little effect or an inhibitory action on natriuretic peptide levels in man. An inhibitory effect might be anticipated from in vitro and animal studies that demonstrate that alpha‐adrenergic agonists stimulate ANP and BNP secretion whereas blockade of alpha‐adrenergic receptors is inhibitory 6.

Positive inotropic agents

Digitalis

Of the three papers reporting effects of digitalis compounds on ANP levels in patients with heart failure, all demonstrated a stimulatory effect over 40 minutes–6 days 127–129 and this was so also for BNP in the report by Tsutamoto 127. Interestingly, two of these reports showed a clear dissociation between digitalis‐induced changes in pulmonary capillary wedge pressure and in ANP and BNP levels 127,129: whereas wedge pressures fell, natriuretic peptide levels rose. These data suggest that the natriuretic peptide‐stimulatory action of digitalis compounds, demonstrated ex vivo130,131, must be powerful in order to overcome a decline in cardiac chamber stretch. The possibility that digitalis compounds might act to reduce the plasma clearance rate of ANP and BNP, thereby contributing to the rise in their circulating levels, however, cannot be discounted. Whether the observed increase in ANP and BNP levels contributes to the beneficial therapeutic effects of digitalis compounds, as has been suggested 129, remains to be determined.

Dopamine‐like agents

Dobutamine and dopamine agonists, administered over minutes to 2 days, reportedly reduce circulating ANP and BNP levels in patients with severe grades of heart failure 132–135. Treatment with ibopamine, an orally active dopamine agonist, for 24 weeks likewise suppressed plasma ANP levels, this time in patients with mild to moderate heart failure 79. An exception to this general rule was reported by Avgeropoulou et al. who showed no change in plasma BNP following a 24‐hour infusion of dobutamine in patients with advanced, decompensated heart failure 136. These workers, however, measured BNP levels only at baseline (pre‐infusion) and 1 and 4 days post‐infusion, not during the infusion itself.

For patients not in heart failure, usually undergoing investigations for coronary artery disease, intravenous dobutamine had little or a minor stimulatory effect on natriuretic peptide levels 137–139. Dopamine infusion in premature infants had no effect on ANP levels 140.

Phosphodiesterase III inhibitors

Milrinone reduced ANP levels along with cardiac filling pressures over 2 hours in seven patients with congestive heart failure 141. The same drug, given intermittently by intravenous infusion over 2 years to a patient with chronic pulmonary thromboembolism, was associated with a substantial (40%–50%) decline in both ANP and BNP levels and improved clinical status, although no decline in mean pulmonary artery pressure was observed 142.

Levosimendan

This drug, which has both positive inotropic (calcium sensitization of cardiac myofilaments) and vasodilator actions, was shown to suppress plasma BNP levels significantly and substantially for up to 5 days following 24 hours of intravenous infusion in patients with decompensated heart failure 136,143. Serial 24‐hour infusions every 3 weeks likewise reduced plasma NT‐proBNP levels while simultaneously reducing left ventricular volume indices in 17 patients with decompensated chronic heart failure 144.

Xamoterol

Compared with placebo, this beta‐1‐partial agonist drug, given over 10 days, was without effect on plasma levels of ANP in patients with or without heart failure who had suffered acute myocardial infarction 145.

Miscellaneous

Amiodarone

Two studies of patients with heart failure demonstrated a substantial decline in plasma levels of natriuretic peptides during chronic amiodarone therapy in a dose of 100–200 mg daily. The first 146 was in 13 patients intolerant of beta‐blockers but otherwise receiving diuretics (all), ACE inhibitors (n = 10) and digitalis (n = 10). Compared with controls (n = 9), those receiving amiodarone showed a steady fall in BNP levels to less than 50% baseline values at 3 months. The second study, by Shiga and colleagues, was in patients with chronic heart failure (NYHA functional class II–IV) and arrhythmias who were receiving diuretics, ACE inhibitors or ARBs and beta‐blockers with or without digoxin 147. Here again, those receiving amiodarone (n = 46) compared to placebo (n = 21 patients who, instead of amiodarone, received an implantable cardioverter defibrillator) reduced plasma BNP levels steadily over 6 months (from a mean of 287 to 180 pg/mL).

Central alpha‐2 agonists

Clonidine injection increased ANP levels in healthy control women and normotensive obese women 148. This fits well with an abundant literature from animal studies which also, in general, describes a stimulatory action of clonidine on natriuretic peptide levels. In contrast, Spinetti et al. reported a 50% reduction in plasma ANP levels and suppression of the ANP response to saline infusion after 3 months of clonidine administration in five post‐menopausal women 149. Interpretation of these directionally opposite observations is difficult given the different patient cohorts and time frames involved.

Aspirin

In a study of 42 survivors of acute myocardial infarction who required an ACE inhibitor because of heart failure or a reduced LVEF (<40%) during their hospital stay, 27 were randomized to aspirin (150 mg daily) and 15 to warfarin. There were no changes in plasma ANP levels across 3 months in either group 150. In contrast, two other studies reported aspirin to have a stimulatory effect on natriuretic peptide levels. The first, by Jug et al., showed that aspirin 100 mg/day increased plasma NT‐proBNP slightly but significantly over 8 weeks in 18 patients with stable chronic heart failure (NYHA functional class II or III and LVEF <40%) and ischaemic heart disease who were taking ACE inhibitors 151. The second study, by Meune and colleagues, also in patients (n = 19) with stable chronic heart failure, reduced LVEF, and taking an ACE inhibitor, demonstrated a rise in plasma BNP levels (from a mean of 107 to 144 pg/mL) during treatment with aspirin 325 mg/day for 14 days 152. A parallel group of 17 patients receiving clopidogrel exhibited no change in BNP.

Heparin

Of 102 patients with chronic heart failure due to dilated cardiomyopathy and receiving conventional ACE inhibitor, beta‐blocker, and diuretic therapy, those randomized to low‐molecular‐weight heparin daily for 12 months (n = 52) exhibited a significant decline in plasma NT‐proBNP levels (and left ventricular diameters) at 6 and 12 months compared with those (n = 50) who did not receive heparin 153. Confirmation of this observation is needed since the study was neither blinded nor placebo‐controlled.

Histamine H₂ receptor blockade

Kim and colleagues reported from a retrospective study and from a randomized, open‐label protocol, that the histamine H₂ receptor blocker famotidine, given to patients with congestive heart failure and gastro‐oesophageal reflux disease, induced a decline in plasma BNP levels in parallel, apparently, with a fall in left ventricular diameters and arterial pressure 154. Most likely the fall in BNP was secondary to improved cardiac function as manifest by the reduction in left ventricular diameters rather than to any direct effect of histamine blockade on BNP secretion since, at least in rabbits, histamine has been shown to inhibit (rather than stimulate) secretion of natriuretic peptides by atrial myocytes via H₂ receptors 155.

Non‐steroidal anti‐inflammatory drugs

Indomethacin given acutely, by mouth or intravenously, failed to alter circulating ANP levels in healthy subjects, patients with chronic glomerulonephritis or patients on chronic intermittent haemodialysis 156–158. In ten patients with essential hypertension, however, indomethacin administration for 1 week increased mean plasma ANP levels from 16 to 21.5 pg/mL 159.

Allopurinol

Gavin and Struthers showed, in a randomized, parallel trial of 50 patients with heart failure (NYHA functional class II or III) associated with reduced systolic function and receiving conventional medications, that allopurinol, 300 mg/day, reduced plasma BNP levels slightly but significantly compared with placebo over the 3‐month study 160. The authors suggested that the fall in BNP levels might have reflected allopurinol‐induced improvements in endothelial and cardiac function.

Statins

Strey and colleagues reported that treatment with atorvastatin 40 mg daily for 6 weeks resulted in slightly, though statistically significantly, lower levels of plasma ANP levels compared to placebo (42 versus 51 pmol/L) in a double‐blind, placebo‐controlled, cross‐over study of 23 patients with chronic heart failure of non‐ischaemic aetiology and normocholesterolaemia 161.

Oestrogens

Oestrogen replacement therapy activates the cardiac natriuretic peptide system 162–164, perhaps via renal sodium retention 164 or stimulation of the renin‐angiotensin system 165.

Immunosuppressive drugs

Notwithstanding the fact that the immunosuppressive drug, doxorubicin, is capable of suppressing ANP and BNP gene expression in cultured myocytes in vitro166, this and similar cardiotoxic agents stimulate circulating levels of the cardiac natriuretic peptides in those who have, or are developing, cardiac failure 167–171.

Overview

As noted in the introduction, natriuretic peptide levels and particularly the B‐type peptides, are being used increasingly in clinical practice under a variety of circumstances. The effects of drug therapy on circulating levels of natriuretic peptides need to be taken into account when interpreting such levels. Of the drugs used to treat cardiovascular disorders, the diuretics reduce elevated levels of natriuretic peptides in heart failure, but their effects in patients with hypertension and in healthy volunteers are less clear. Beta‐blocking drugs stimulate natriuretic peptide levels in normal subjects and in patients with hypertension or coronary artery disease. For patients with cardiac failure, the effect of beta‐blockers is more complex. In the short term they tend to increase peptide levels whereas in the longer term they are inhibitory, at least when clinical and haemodynamic improvement results. Drugs which block the renin‐angiotensin system reduce natriuretic peptide levels, most obviously when pre‐treatment levels are elevated as in severe grades of heart failure but also in essential hypertension, especially when there is left ventricular hypertrophy. Different calcium channel blockers appear to have different effects. Thus, verapamil has little or no effect in the short term but often stimulates natriuretic peptide levels with chronic administration. Amlodipine tends to suppress, or alter little, natriuretic peptide levels whereas nifedipine and felodipine are stimulatory in the short term and inhibitory with chronic administration. Vasodilator drugs, when used in the management of heart failure, appear to lower natriuretic peptides in parallel with haemodynamic responses. Alpha‐adrenergic receptor blocking drugs suppress, or have little effect. Whereas digoxin stimulates natriuretic peptide levels even, apparently, when haemodynamic indices improve, dopamine‐like agents reduce their circulating concentration congruent with improvements in central haemodynamics. Amiodarone, allopurinol, and atorvastatin have been shown to reduce natriuretic peptide levels in patients with heart failure whilst aspirin has a minimal or stimulatory effect. Histamine H₂ receptor blockade and low‐molecular‐weight heparin have each been reported to reduce B‐type natriuretic peptides in heart failure. In the few studies completed, opposite effects of clonidine on ANP have been reported, and indomethacin increased slightly, or was without effect on, ANP.

Acknowledgements

We are grateful to Barbara Griffin for secretarial assistance.