Underestimation of cardiac vagal control in regular exercisers by 24-hour heart rate variability recordings

Abstract

Objective

To examine whether ceiling effects at long inter beat intervals (IBIs)cause an underestimation of cardiac vagal control in regular exercisers by time and frequency-domain measures of respiratory sinus arrhythmia (RSA).

Methods

24-hour ECG and respiration recordings were performed in 26 regularly exercising subjects, actively engaged in aerobic training for the past year, and enrolled in supervised training in the six weeks pre-study, and in 26 age- and sex-matched non-exercisers. Sleep and waking levels of cardiac vagal control were estimated by RSA obtained through the peak–valley method, by the standard deviation of the IBIs, the root mean square of successive IBIs, and the high frequency IBI spectral power.

Results

In 11 of the exercisers the IBI–RSA relationship was characterized by a quadratic relationship. This reflected a ceiling effect at very long IBI values attained by regular exercisers, particularly during the nighttime recording. Irrespective of this ceiling effect, RSA as well as other heart rate variability (HRV) measures was still significantly larger in the exercisers with a quadratic IBI–RSA relationship than in non-exercisers or exercisers with a linear IBI–RSA relationship.

Conclusions

We conclude that a subgroup of regular exercisers is characterized by a low heart rate paired to high levels of cardiac vagal control. In these exercisers, vagal control is underestimated from HRV measures in ambulatory recordings. Inspection of the IBI–RSA relationship should be routinely added when HRV measures are used to index cardiac vagal control.

Introduction

Regular vigorous exercise is associated with many favorable physiological adaptations, including a lower resting heart rate. Dual blockade studies point to a lower intrinsic heart rate as the most replicated source of this resting bradycardia in exercisers (Lewis et al., 1980, Bouchard et al., 1988, Bouchard et al., 1999, Katona et al., 1982, Smith et al., 1989, Goldberger et al., 2001, Kingwell et al., 1992, Uusitalo et al., 1996) and this is supported by findings in animals (Lin and Horvath, 1972, Negrao et al., 1992). In addition to lowered intrinsic heart rate, exercisers have long been hypothesized to possess a stronger vagal control over the heart rhythm (Ekblom et al., 1973). Animal studies and studies in cardiac patients generally support an effect of exercise on cardiac vagal control (Gutin et al., 2005, Goldsmith et al., 2000, Billman, 2002, Mueller, 2007, Goldsmith et al., 2000) but in healthy human subjects, the evidence for an exercise-induced shift in vagal control is more controversial (Goedhart et al., 2008).

Cardiac vagal control is most often quantified by time- or frequency domain indices of heart rate variability (HRV) in the respiratory frequency range, also called respiratory sinus arrhythmia (RSA) (Berntson et al., 1997, Martinmaki et al., 2006, Nunan et al., 2010, Task Force of the European Society of Cardiology the North American Society of Pacing, 1996). A number of cross-sectional studies reported higher RSA in regular exercisers (Buchheit et al., 2005, Martinmaki et al., 2006, Dixon et al., 1992, Goldsmith et al., 1992, Shin et al., 1997) but not all studies support this difference (Hatfield et al., 1998, Goedhart et al., 2008) and some even report the opposite finding of lower RSA in exercisers compared to non-exercisers (Sacknoff et al., 1994). Notably, various randomized controlled training studies that assigned untrained subjects to a non-exercise control manipulation or a standardized exercise training program have failed to find a specific training-induced increase in RSA (Loimaala et al., 2000, Uusitalo et al., 2004, Boutcher and Stein, 1995, de Geus et al., 1996).

Here we hypothesize that a specific methodological problem in assessing RSA in well-trained exercisers may have led to an underestimation of the beneficial effects of regular exercise on cardiac vagal control in previous studies, both cross-sectional and longitudinal. We note that in a meta-analysis of training studies by Sandercock et al. (2005) a significant training-induced increase in RSA was seen during short-term laboratory recordings at rest, but not in 24-hour ambulatory recordings. Furthermore, some of the ambulatory studies suggested that training effects on HRV may be confined to the daytime but absent in the whole recording or nighttime levels (Schuit et al., 1999, de Geus et al., 1990). An obvious difference between (laboratory) daytime and ambulatory nighttime recordings is the absolute level of heart rate attained at night. As indicated as early as 1993 by Malik and Camm (1993) RSA may not be a reliable index of cardiac vagal control in subjects with a low heart rate.

RSA is formally defined as the difference between the shortest inter beat interval (IBI) during inspiration and the longest IBI during expiration (Katona and Jih, 1975, Grossman, 1983) and the main physiological rationale to use it as an index of cardiac vagal control is that neural vagal activity is selectively inhibited during inspiration but not during expiration. However, at very high vagal activity, a ceiling effect may prevent lengthening of the IBI during expiration more than during inspiration (Malik and Camm, 1993). High vagal activity causes a large occupancy of the available muscarinergic receptors on the sinoatrial (SA) node, and at this level of saturation any further increases in acetylcholine may no longer linearly increase the IBI as it would at low to moderate levels of cardiac vagal activity (illustrated in Fig. 1).

This ceiling effect is expected to cause a quadratic relationship between IBI and RSA. A quadratic shape of the IBI–RSA relationship has indeed been found in studies manipulating vagal activity by phenylephrine and nitroprusside infusion (Goldberger et al., 2001). The quadratic shape of the IBI–RSA relationship will cause an underestimation of cardiac vagal control by RSA in subjects with low resting heart rates.

Because regular exercisers often have lower resting heart rates than non-exercisers we hypothesize 1) that exercisers more often have a quadratic IBI–RSA curve, 2) that RSA underestimates cardiac vagal control in these exercisers, and 3) that this underestimation is aggravated during 24-hour ambulatory monitoring that includes evening and night time recordings when vagal activity is high.