Is there a heart rate paradox – demystification of the myths?

The introduction of beta blockers in cardiovascular medicine 45 years ago markedly improved the outcome of coronary artery disease, heart failure and hypertension. These drugs have been nicknamed cardio protective. As humans we want simple explanations for complicated matters, but the problem with generalizations are that we often forget that we are dealing with a group of drugs that have different pharmacodynamic profiles and ancillary properties. The clinical targets are also complex and respond nonuniformly to beta blockers. The main target for beta-blocker treatment has been to block a hyperactive sympathetic system in order to reduce myocardial oxygen consumption by a reduction of heart rate, blood pressure and myocardial contractility. Beta receptor blockade may also cause metabolic changes and reduce systemic inflammatory activity. A subclinical hyperactive sympathetic system is a predictor of cardiovascular disease and secondary activation of the sympathetic system in heart failure is markedly aggravating the disease. In epidemiological studies elevated heart rate and blood pressure have proven to be independently associated with cardiovascular mortality and morbidity 1, 2. There is a close association between enhanced sympathetic drive and heart rate in asymptomatic subjects amongst young and hyper dynamic hypertensives, amongst patients with overt coronary artery disease and amongst patients with heart failure 3–5. High resting heart rate also is associated with the metabolic syndrome which is associated with abdominal obesity, insulin resistance, hypertension, atherogenic lipid profile, elevated levels of glucose, renin, noradrenaline, free fatty acids and subclinical inflammatory activity 6, 7. Elevated heart rate is therefore intimately related to cardiovascular risk factors. However, even after adjustments for other cardiovascular risk factors, heart rate remains an independent risk factor for plaque rupture, subsequent myocardial infarction, sudden death, stroke, atrial fibrillation, diabetes, and heart failure irrespective of gender 1, 2, 8, 9. It has therefore been extrapolated from epidemiological studies that heart rate reduction may prevent cardiovascular events. Beta blockers without or with minimal intrinsic sympathomimetic effect have consistently reduced mortality by about 30% in patients with coronary heart disease 10 and about 35% in patients with established systolic heart failure 11. Although beta-blockade effectively reduces heart rate, clinical outcome studies have not tested the implications of a predefined heart rate target. However, observations from single studies and from meta-analysis have demonstrated a direct association between the reduction in heart rate and cardiovascular benefit in ischemic heart disease 12–14 (Fig. 1) and in patients with heart failure caused by systolic dysfunction 15 (Fig. 2). The clinical improvement in heart failure is paralleled by an increase in left ventricular ejection fraction by 5–10% and reduction of atrial natriuretic peptides 16. Similar effects on left ventricular ejection fraction are observed in pacemaker dependent patients with heart failure when pacing rates are reduced from 75 to 55 beats per min 17. Conversely, reversal of beta-blocker induced bradycardia has deleterious effect on ventricular function and survival 18, 19. Effect of beta blockers on heart rate reduction and mortality in patients with ischemic heart disease. Relationship between reduction in heart rate (difference between treatment modalities) and percentage reduction in mortality in large, prospective, double-blind trials with beta blockers. The regression line omits the smallest study (open circle). r = 0.6; P < 0.05. By permission Ref. 13. Effect of beta blockers on heart rate reduction and mortality in patients with heart failure. Relationship between reduction in heart rate (difference between treatment modalities = Δheart rate) and percentage reduction in mortality in large, prospective, double-blind trials with beta blockers in patients with symptomatic heart failure due to systolic dysfunction The regression line : r = 0.59; P = 0.027. By permission Ref. 15. Permission granted by Les Laboratoires Servier, copyright holder of Medicographia. In this journal Messerli argues correctly that in hypertensive patients beta blockers have no incremental benefit beyond the reduction in blood pressure 20. Their meta-analysis do not demonstrate benefit of beta blockers in preventing heart failure over that obtained with other antihypertensive agents 21. This is in accordance with the Blood Pressure Lowering Treatments Collaboration group 22. The seminal Veterans Administration (VA) study was the first to demonstrate clearly that depletion of sympathetic neurotransmitters with reserpin markedly reduced cardiovascular end-points including death and heart failure 23. Messerli and others have clearly demonstrated that the prevention of major cardiovascular events are closely dependent on the extent of blood pressure lowering, independent of drug used – supporting the concept that lower is better at least down to 115/75 mm Hg 22. In the VA study, blood pressure was reduced by combined heart rate reduction and peripheral vasodilatation. Beta blockers reduce blood pressure primarily by reducing heart rate and cardiac output without changing peripheral resistance. Peripheral resistance which actually may increase following upstart of beta-blocker treatment is never reduced below pretreatment levels during long-term follow-up 24. This is in contrast to angiotensin converting enzyme (ACE) inhibitors and calcium antagonists which reduce blood pressure primarily by reducing peripheral resistance 24. Beta receptor blockers and peripheral vasodilators have a complementary effect in hypertension. It is therefore counterintuitive that the heart rate reduction would not benefit in hypertension. Messerli argues against this hypothesis by showing data of a paradoxical inverse relationship between heart rate on beta-blocker treatment at the end of trial and outcome 25. In other words, heart rate reduction is counterproductive in hypertensives. Some caveats are needed in this argument. The authors use heart rates at the end of the studies. This carries a potential flaw in that after the end of the study no more events are recorded. The correct time-point to measure the heart rate would be after 1 month of treatment and record the subsequent end-points. Furthermore, the rates of specific events like nonfatal myocardial infarctions are inversely related to rates of mortality. An event occurring amongst high risk patients are more likely to be fatal and few high risk patients are reported with nonfatal myocardial infarctions. Effective treatment that converts an event from death to a nonfatal myocardial infarction may therefore spuriously improve the relative risk of myocardial infarction at higher heart rates. Higher heart rate reflects high risk patient. The interrelation may therefore invert the relationship between heart rate and relative risk of having a myocardial infarction. This was clearly demonstrated in the timolol trial 12. It is also pertinent to use the difference in heart rates between treatment modalities as an estimator rather than only the heart rate on beta-blocker treatment. By using the heart rate difference between treatment modalities in their analysis the inverted slope between heart rates at the end of trial and relative reduction of mortality becomes less convincing 25, 3, 4). Relative risk of all cause mortality as function of heart rate at the end of the study on beta blockade. Relationship between heart rate in the beta-blocker group at the end of the study and relative risk of mortality. The diameter of the circles represents the weight of each individual trial. The line represents the regression fit with 95% confidence interval for the effect sizes. With permission Ref. 25. Relative risk of all cause mortality as function of heart rate difference between treatment modalities at the end of the study. Redrawn from data in Ref. 15. Relative risk of all cause mortality as a function of heart rate difference at the end of the study between beta-blocker treatment and placebo in the same trials as in Fig. 3. P = not significant. Beta blockers with intrinsic sympathomimetic effects have consistently less protective effect on cardiovascular end-points 13. Most clinical investigations on hypertensives have used the water-soluble atenolol for practical reasons often in combination with a diuretic. There is an unsettled discussion if hydrophilic and lipophilic properties matters for cell permeability in the vasculature and vasomotor centres in the brain and accounts for differences reported with respect to intima media thickness and sudden cardiac death 26. The use of nonselective beta blockers in hypertension has been claimed to be less effective than β-1 selective because of the unopposed β-2 vasodilatation. Even atenolol which is the most frequently used comparator in hypertension trials is only moderately β-1 selective and blocks 25% of β-2 receptors 27. A possible criticism against meta-anlyses of hypertensives is that there is no discrimination between different categories of hypertension. The young to middle-aged hyper dynamic patients with elevated sympathetic activity and the elderly patients with ‘pipe-stem’ noncompliant arteries and reduced renin and sympathetic activity behaves differently to treatment 27. As pointed out by Messerli beta blockade in hypertensive patients causes a paradoxical increase in central systolic blood pressure during induced bradycardia. The compensatory increase of stroke volume causes early reflection of the pressure waves from the stiff arteries which increase central blood pressure, suggested to increase the loading conditions of the coronary vessels and the myocardium 28. The treatment target in hypertensive patients is peripheral blood pressure. In beta-blocker trials heart rate reduction and the negative inotropic effect are decisive for the effect on blood pressure because they have no or very little effect on peripheral resistance 24. Nature has devised a nice model which demonstrates the functionality of heart rate reduction. The diving seal can survive for prolonged periods under water by an instantaneous reduction of heart rate from 100 to 5–7 beats per min, demonstrating the importance of the concept of oxygen sparing effects obtained by inducing bradycardia 29. When submerged the whole peripheral circulation is virtually shut off markedly reducing the buffering capacity of the periphery. However, reflected pressure waves from the central aorta are effectively dampened by a huge saccular expansion of the ascending aorta that acts like a ‘Windkässel’ in these animals which effectively avoids blood pressure changes during the dive. Messerli demonstrates an important dichotomous difference in the beta-blocker effect on stroke. A significant prevention of strokes were observed amongst patients younger than 60 years and an excess in patients older than 60, when compared to other antihypertensive agents 25. This is commensurate with the different mechanisms for hypertension in young and middle-aged compared to the elderly patients. Hypertension in young and middle-aged patients are more likely caused by a hyperactive sympathetic condition whilst peripheral resistance increase with ageing and in elderly patients hypertension is characterized by being primarily systolic due to a noncompliant aorta and by reduced sympathetic tone and no elevation of heart rate. Hypertensive elderly may therefore be less likely to benefit from beta blockade compared with young and middle-aged hypertensives unless combined with a vasodilator. Elevated heart rate during cardiac and noncardiac surgery is associated with increased cardiovascular events and occurrence of atrial fibrillation. The Perioperative Ischemic Evaluation Study (POISE) trial demonstrated that perioperative beta-blocker treatment prevented atrial fibrillation and reduced major coronary end-points by 16% driven by nonfatal myocardial infarction, but at the expense of higher all cause mortality and stroke 30. Messerli warns about the association between perioperative beta receptor blockade and stroke unless the patients have ischemic heart disease and heart failure. The flaw of the POISE trial was to start beta-blocker treatment at full dose 2 h before the operation. This is similar to the adverse effect observed in the Clopidogrel and Metoprolol in Myocardial Infarction Trial (COMMIT) trial when beta-blocker treatment was started during the evolving infarction 31. In the perioperative setting the ‘first dose’ hypotension and symptomatic bradycardia are risk factors. The lesson we learnt in heart failure studies was that due time should be given for up titration and for adjustment to the beta-blocker intervention. POISE illustrates nicely the problems which may arise when beta blockade is initiated at the time of operation. In contrast the Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiographic Stress Group (DECREASE) trial started bisoprolol treatment in high risk patients 7 days before the operation and subsequently observed a reduction of cardiac events by 34% 32. Above the age of 80 years 4–5% of all individuals have atrial fibrillation. Hypertension and a high heart rate are strongly associated with new onset of atrial fibrillation 9, 33. Beta-blocker treatment can prevent or delay the onset of atrial fibrillation in hypertensives 9 and can prevent recurrence of atrial fibrillation after cardio version 34, but as pointed out by Messerli beta blockers are not antiarrhythmic agents per se once atrial fibrillation is present. However, emerging information suggest that rate control compared to rhythm control is associated with better prognosis supporting the concept of heart rate reduction may benefit patients with atrial fibrillation 35. Messerli advocates against the use of beta blockers because they may cause diabetes at the risk of increasing mortality and stroke compared to nondiuretic antihypertensive agents 36. Especially the combination of beta blocker and diuretic treatments may presage new onset diabetes. The metabolic changes of beta blockers are primarily associated with β-2 receptor blockade, whilst highly β-1 selective agents or agents containing α-1 blocking properties will essentially be free of metabolic changes 37. Their role in preventing cardiovascular events in diabetic patients with ischemic heart disease is as yet unrivalled 38, 39. In the future, we can hope for β-1 selective beta blockers combined with β-3 receptor agonists which will increase fatty acid oxidation and enhance insulin sensitivity. The important questions raised by Messerli et al. reminds us that even after 45 years with beta blockers we still have things to settle with respect to their clinical role, but it is too early to write of beta blockers. No conflict of interest was declared.