Heart rate variability (HRV) reflects cardiac autonomic function, which decreases with age, especially in frail patients and those with a poor functional status.1 For elderly outpatients undergoing cardiac rehabilitation (CR), exercise tests, when feasible, provide significant information regarding functional capacity.2 However, there is little evidence on the possible relationship between HRV in older adults with cardiovascular diseases (CVD) and functional capacity estimated using an exercise test.3 Therefore, the aim of this study was to verify whether and to what extent HRV is correlated with the functional capacity of older adults by evaluating the association between HRV parameters and physical performance during exercise tests in elderly patients undergoing CR.

METHODS

This was an observational study of older patients consecutively enrolled in the Cardiac Rehabilitation Program of the Italian National Research Centre on Aging, Fermo, Marche, Italy. The ethics committee of the institution approved the study protocol (430DGEN/19). The inclusion criteria were age >=65 yr and cardiology referral to undergo CR. Active smokers, patients with permanent atrial fibrillation or cardiac pacemakers, patients who are unable to exercise, and patients for whom exercise is contraindicated were excluded from the study.

At enrollment, each patient underwent a comprehensive cardiac examination, including standard transthoracic echocardiogram with a Vivid7 (GE), 24h-Holter electrocardiogram (ECG) with H12+ digital (Hillrom-Mortara), and exercise test on a treadmill (Jaeger-Vyntus) according to the Bruce protocol. The indices of functional capacity measured were exercise duration (sec) and metabolic equivalents (METs) achieved at the peak of exercise.4

The following HRV time-domain indices were obtained from the 24h-Holter ECG of each patient: SD of normal-to-normal (NN) intervals (SDNN); SD of the averaged NN intervals (SDANN); mean of the SD of all NN intervals (SDNNi); root mean square of successive NN interval differences (RMSSD); and percentage of successive NN intervals >50 msec (pNN50). In the frequency domain analysis, low frequency (LF) (range: 0.04-0.15 Hz) and high frequency (HF) (range: 0.15-0.40 Hz) were calculated using fast-Fourier transformation in 5-min segments for the entire duration of the recording. The LF/HF ratio (LF/HF) was calculated as well.

STATISTICAL ANALYSIS

On the basis of a previous study, we estimated that, with a power of 0.8 and a type I error probability of .05 (two-sided P), >=29 patients should be enrolled.3 Data were analyzed using SPSS Statistics for Windows version 20 (IBM). In the descriptive analysis, categorical variables were expressed as n (%), whereas continuous variables were expressed as mean +/- SD or median (IQR). Associations between HRV measures and physical performance parameters were examined using the R-Pearson or Kendall tau-b analysis, as appropriate. A multivariable stepwise model was constructed to test any independent relationship between the duration of the exercise test as the outcome variable and other covariates. We included the following covariates in the model: SDNN, LF/HF, age, body mass index, diabetes, left ventricular ejection fraction, renin-angiotensin system-acting agents, and [beta]-blockers. A two-tailed P < .05 value was considered statistically significant.

RESULTS

Forty-one older patients with a median age of 72 (68, 75) yr, predominantly male (95%), were included in this study. Thirty-six (88%) participants had ischemic heart disease, 34 (83%) had dyslipidemia, 33 (80%) had hypertension, and 10 (24%) had diabetes. The median body mass index was 27.2 (24.9, 28.7) kg/m². Thirty (73%) patients were taking renin-angiotensin system-acting agents, whereas 35 (85%) were taking [beta]-blockers. The median ejection fraction on echocardiogram was 60 (55, 65) %. The mean duration of the exercise test was 448.4 +/- 22.8 sec, whereas the median METs at exercise peak was 10.2 (7, 13.5).

For the time-domain HRV parameters, the mean SDNN was 134 +/- 54 msec, whereas the mean SDANN was 99 +/- 31 msec. The median SDNNi was 58 (43, 98) msec, the median pNN50 was 10.6 (4.6, 20.1) %, and the median RMSSD was 49 (32, 116) msec. For the frequency domain parameters, the median LF was 414 (247.5, 2378.5) msec2, the median HF was 578 (252, 3394) msec2, and median LF/HF was 0.88 (0.6, 1.24). The SDNN, SDANN, and SDNNi were significantly associated with exercise duration (R: 0.567, P < .001; R: 0.489, P = .001; Kendall tau: 0.276, P = .011, respectively) and METs (Kendall tau: 0.353, P = .004; Kendall tau: 0.347, P = .004; Kendall tau: 0.324, P = .007, respectively).

The scatterplot in the Figure depicts the association between SDNN and the functional capacity parameters. The LF was the only frequency domain index associated with functional capacity expressed as METs (Kendall tau: 0.250, P = .039). In the stepwise multivariable model, SDNN was the main determinant of exercise duration (r²: 0.321, P < .001), followed by age (r² variation: 0.080, P = .030).

Figure. Association between SDNN and functional capacity measures. Significant positive associations between SDNN and exercise test duration (circles; continuous trend line) and between SDNN and METs achieved at the peak of the exercise test (triangles; dashed trend line) are evident. Abbreviations: METs, metabolic equivalents; SDNN, SD of normal-to-normal intervals.

DISCUSSION

This study showed that in elderly patients with CVD, time-domain HRV indices reflecting sympathetic activity are associated with functional capacity measures. It is reasonable that SDNN had the strongest relationship with performance in the exercise test because it has been proven that it can reflect the responses of the autonomic nervous system to changes in workload.5 Even the association between LF and METs seems logical because LF mainly mirrors heart rate fluctuations secondary to baroreceptor activity, which plays a key role in moderating increase in blood pressure during exercise.6 The finding that pNN50 and RMSSD were unrelated to functional capacity could be attributed to the effects of respiration and sleep since these indices reveal mostly parasympathetic nervous system activity.7 Interestingly, LF/HF was not associated with physical performance. In fact, although widely used, LF/HF seems to be an inappropriate index for describing "sympatho-vagal balance."8

This study was limited by its small sample size, cross-sectional design, and a low number of female participants. However, it is relevant in that it highlights that in older outpatients undergoing CR, better HRV values are associated with better physical performance, probably because they reflect healthier aging.9

Significant benefits of CR, including reduction of all-causes mortality, are demonstrated in aged patients.10 In addition to providing prognostic information on CVD risk, routine evaluation of HRV using 24h-Holter ECG may predict the physical performance of older persons undergoing CR. However, larger studies are needed to clarify the relationships between HRV and functional capacity in elderly patients with CVD.

ACKNOWLEDGMENT

The authors thank Anna Libertazzi, RN, and Michela Coco, RN, for their nursing support. The authors thank Editage (http://www.editage.com) for English language editing.

REFERENCES