Atrial fibrillation: a left atrial neuro-myopathy?

This editorial refers to ‘Longitudinal association of short-term, metronome-paced heart rate variability and echocardiographically assessed cardiac structure at a 4-year follow-up: results from the prospective, population-based CARLA cohort’ by D. Medenwald et al., on pages 2027–2035. Medenwald et al.1 are to be commended for studying the CARLA (CARdiovascular Living and Ageing in Halle) cohort over 4 years (1436 patients, 50–87 years, supine rest, and paced ventilation). They observed an enlargement of left atrial dimension and a reduction of vagus—related indices of heart rate variability (HRV): reduced standard deviation of normal-to-normal intervals (SDNN), reduced amplitude of high and low frequency power, RMSSND (root mean square of successive normal interval difference) and of SD2 (‘length’ of Poincare’s plot). Attenuation of cardiac vagal tone, more precisely attenuation of phasic vagal action2 is suggested. Given identical ventilation and supine position, only intra-individual differences can yield such changes. With aging, the ability of cardiac vagal activity to adapt to any variation is dampened. Based on their statistical analysis, (i) reduction of cardiac vagal activity evokes changes in left atrial dimension and (ii) changes in left atrial dimension do not lead to modifications of the parasympathetic tone. A causal link may be inferred between reduced cardiac vagal action, left atrial enlargement and atrial fibrillation. However, in a comment to a reviewer, the authors state: « Our data do not support the notion that an… .increase in sympathetic activity or decrease in parasympathetic activity… . « causes » a change; say an increase in LA dimension. In the light of the very high prevalence of hypertension in our cohort, … .the causal chain goes from hypertension via an increase in left atrial dimension to an increase in sympathetic/decreases in parasympathetic tone ». Indeed, causality in medicine does not rely on one statistically significant relationship, but on a set of criteria:3 (i) strength of association, (ii) consistency, (iii) specificity, (iv) temporality (v) biological gradient, (vi) plausibility, (vii) coherence, (viii) experiment, and (ix) analogy. To sum up, one has to provide as many criteria as possible before drawing any causal conclusion. A significant statistical association alone is not sufficient. What does the cardiac vagal system do in a young healthy, resting, supine individual? In the setting of lowered sympathetic and heightened parasympathetic activities, any increase in systolic blood pressure (SBPn) is matched, within the same beat n, by vagal bradycardia: the longer diastolic emptying n triggers a lowered n + 1 SBP, despite an increased stroke volume n + 1. Vagal heart rate reduction drives SBP. The large SDNN (or any index of parasympathetic activity) reflects the ability to buffer blood pressure changes: small beat-by-beat variation in pressure occurring next to resting blood pressure evokes large beat-by-beat changes in RR interval (high sinus variability opposing low pressure variability (Figure 1A vs. 1B in4 and 3B vs. 3D in5). Small pressure changes are linked to large RR intervals variation by a relationship presenting with a high slope sensitivity on a baroreflex graph (RR interval = f(SBP)). This holds as long as the efferent cardiac vagal activity impinging onto the sinus node arrives timely,6 i.e. if heart rate is under 75 bpm. By contrast, upon standing, the heart rate increases within <1–2 s;7 cardiac parasympathetic withdrawal starts very rapidly (<1–2 s) before any increase in vascular/cardiac sympathetic activity (<2–10 s). The consequence of this cardiac vagal withdrawal is to pump more blood into the aorta: upon orthostasis, this matches the sudden drop in blood pressure, overriding diastolic emptying. The sluggishness of the cardiac and vascular sympathetic recruitment within the first 1–2 s and the swiftness of the continuous parasympathetic withdrawal vs. engagement allow to withstand most situations of daily life, except exercise, fight or flight. In these latter circumstances, the sympathetic system is massively involved. Does this apply to elderly individuals not lying in an armchair? There are at least three possibilities. Progressive loss of elastic properties of the arterial vessels may explain why buffering the systolic blood pressure variability becomes more difficult. Cardiac vagal activity decreases with age: phenylephrine infused to achieve 152±19 mm Hg evokes a maximal bradycardia to ≈33 and ≈48 beats per minute (bpm) in the young and elderly volunteer respectively.8 Any hypotension, irrespective how minimal the amplitude of the hypotension is, not matched within the same beat, by parasympathetic withdrawal (<1–2 s) but by a sympathetic activation (>2–10 s). With left ventricular (LV) dysfunction, there is no tachycardia during simulated orthostasis.9 The continuous cardiac parasympathetic withdrawal vs. engagement observed in the young performs poorly in the elderly. The missing bradycardia builds up blood pressure increase, from one beat to the next, as the vasomotor and cardiac sympathetic system runs unabated: in the elderly, repeated systolic blood pressure « surges » cause large BP « lability » (Figure 3B in5) as opposed to small BP variability observed in the young. When this phenomenon recurs over years, vascular damage to the arterial wall occurs.10 The loss of elastic properties of the arterial wall evokes BP lability, hypertension, left atrial enlargement, and ultimately atrial fibrillation.11 Left atrial size, LV wall thickness, and atrial fibrillation are associated.12,13 Aging cardiac vagal motoneurons. Given the difficulty to record cardiac vagal motoneurons in the lower brain stem,14 the central impact of low-pressure stimuli is not documented. Nevertheless, (i) stretching the sinus node in the right atrium decreases high frequency HRV in vagotomized animals following beta-blockade,15 (ii) the early phase of pace-induced reduction in LV ejection co-exists with reduced cardiac vagal activity,16 (iii) a higher cardiac baroreflex slope is associated with a higher incidence of paroxysmal atrial fibrillation,17 and (iv) increased adrenergic drive is followed by vagal predominance immediately before the onset of atrial fibrillation.18 Therefore, ‘the sympathetic…system …promotes atrial and ventricular fibrillation… the parasympathetic…system is protective against ventricular fibrillation but is profibrillatory in the atrium’ in normal and failing hearts.19 The autonomic nervous system is involved, causally or else,3 in the genesis of atrial fibrillation. Medenwald et al.1 have added a brick in the wall: there are physiological15–18 and now statistical/epidemiological1 arguments linking reduced cardiac vagal activity and enlarged left atrium. Modern ECG or echocardiography recorders can easily be fitted with software that can analyse heart rate variability after a 5-min supine rest. Integrative medicine has the potential to lead to precision medicine and some new personalized pharmacological and non-pharmacological therapies20 could emerge to increase cardiac parasympathetic activity. Conflict of Interest: none declared.