Autonomic Nervous System and Cognitive Impairment in Older Patients: Evidence From Long-Term Heart Rate Variability in Real-Life Setting

Changes in cognitive performances and cardiovascular disorders represent a normal phenomenon of the aging process (Di Carlo et al., 2000; Leritz et al., 2011; Harada et al., 2013). Cardiovascular risk factors, such as smoking, obesity, diabetes mellitus, increased cholesterol, and systemic blood pressure (BP) levels, and an inadequate lifestyle may compromise also cerebral blood flow, which in turn can negatively affect cognitive performance (Takeda et al., 2017). Moreover, the same age-related anatomical and functional cardiac changes, including also the autonomic nervous system (ANS), determine cardiac output alteration, causing cerebral blood flow modulation. This variation could interfere with microcirculation and cause cerebral ischemia, particularly in those brain sites that control the different cognitive domains (Chen et al., 2011; Al-Qudah et al., 2015; Takeda et al., 2017). Past and recent evidence (Camm et al., 1996; Shaffer and Ginsberg, 2017; Singh et al., 2018a,b) support the relevance of ANS study by heart rate variability (HRV) assessment as a tool for the noninvasive analysis of cardiovascular autonomic function. HRV, defined as a marker identifying the balance between sympathetic and parasympathetic tone, predicts total mortality, sudden death, cardiovascular disease risk, as well as other morbidities (Sessa et al., 2018). It represents the measure of physiological variation in the interval between consecutive heart sinus beat or in the fluctuations between instantaneous heart rates and provides the importance that the ANS has regarding cardiovascular health and prognosis (Shaffer and Ginsberg, 2017). The parameters describing different HRV characteristics are generated by several algorithms ranging from simple mathematical or geometrical indices to complex nonlinear parameters. However, the interpretation of HRV parameters is complex and, in some cases, controversial (Camm et al., 1996). HRV, certainly, represents the evidence of the control of the ANS on the heart, but at the same time, HRV is very often used to describe the ANS functioning in general. Long-term HRV analysis (24 h) and short-term HRV analysis (5 min) are the two main methods for HRV analysis. The first is the time-domain assessment method, recorded over longer periods, traditionally 24 h, and the second is the frequency-domain assessment method, obtained by spectral analysis performed in shorter segments of 5 min as well as from the averaged over the entire 24-h period (Camm et al., 1996). In particular, a longer recording (24 h) better represents the daily activity, the circadian rhythms, and the response to environmental stimuli (Camm et al., 1996; Nunan et al., 2010; Kuusela, 2013; Sassi et al., 2015; Shaffer and Ginsberg, 2017; Singh et al., 2018a,b). On the contrary, short-term surveys (5 min) are suitable for outpatients, or to perform stimulation tests (Camm et al., 1996; Nunan et al., 2010; Kuusela, 2013; Sassi et al., 2015; Shaffer and Ginsberg, 2017; Singh et al., 2018a,b). Therefore, both methods are not interchangeable and should be used in a complementary way (Camm et al., 1996; Shaffer and Ginsberg, 2017; Singh et al., 2018a,b).

Although there are numerous studies in the literature evaluating the relationship between cognitive performance and cardiovascular risk factors, conversely, the relationship between ANS and cognitive function is less explored in the aged population (Britton et al., 2008; Nicolini et al., 2014). In particular, most previous studies tend to report different HRV measures by assessing them in different ways (during rest and/or during physical activities) and performing HRV in selected populations that do not always produce data consistent with common aging. The key issue is, therefore, the possibility of having effective data in contexts coherent with real life without compromising the effectiveness of the relative results.

Thus, the aim of the current study was to assess the relationship between HRV and cognitive performance in older patients representative of a real-life setting, during daily routine activities.

According to the exclusion criteria, 117 (61 males and 56 females) of the 155 patients recruited for the study were suitable for analysis. Table 1 shows the baseline characteristics of the participants. All patients were aged, slightly overweight, on adequate glycemic and metabolic control, and had an education level mean of 6.2 years (range, 1–18 years). Only 10% of the study participants consume approximately five to eight cigarettes/day. Given the age, the ejection fraction of the left ventricle (LVEF) was found to be normal (53.1 ± 2.1%).

Time and frequency variability values obtained from the 24-h Holter recordings and expressed as median are reported in Table 2. Patients showed longer NN intervals, signifying a prevailing sympathetic functioning. With regard to the frequency-domain measures, LF power was higher than HF power and the LF/HF ratio was resulted >1, thus showing a slight prevalence of the sympathetic system and suggesting a decreased parasympathetic control of heart rate (Table 2). Furthermore, all time-domain measures were higher in males compared to female patients.

By comparing the HRV parameters of males and females, we found significant differences for SDNN and SDANN (117.1 ± 36.7 vs. 97.3 ± 22.4, p < 0.001, and 103.9 ± 32.3 vs. 89.2 ± 21.9, p < 0.005, respectively), signifying an increased sympathetic functioning in male subjects, although males had the HR mean (70 ± 13 vs. 73 ± 9, p < 0.164) and HR mean from 24-h ECG (68 ± 9 vs. 70 ± 8, p < 0.303) lower than the females, perhaps because they were taking β-blockers.

Regarding global cognitive assessment (Table 3), the MMSE and MoCA scores show a mild cognitive decline. All cognition data were adjusted for the average number of educational years (6.2 ± 3.8 years). In particular, 69 (58.9%) and 81 (69.3%) patients show normal cognitive performances, while 48 (41.1%) and 36 (30.7%) patients show a mild cognitive decline, as revealed by both the MMSE and MoCA test, respectively. No significant cognitive profile differences were observed between males and females (Table 4).

The most frequently reported comorbidities included hypertension (54.9%), chronic kidney disease (28.4%), and type 2 diabetes mellitus (28.3%), all on adequate therapeutic control. Almost one-fifth of the patients (23.4%) were treated with β-blockers.

The aim of our study was to assess the association between HRV, as a measure of ANS functioning, and cognitive performance in elderly patients during daily activity, thus representing a real-life setting of the older population. The results suggest a connection between HRV and global cognitive performance, with a stronger association between the sympathetic parameters and better cognitive performance. In contrast, no correlation between the indices of parasympathetic modulation and cognitive performance was found. Our findings suggest that only increased sympathetic functioning, but not reduced parasympathetic modulation, implies a good contribution of sympathetic autonomic regulation to the overall cognitive function. Activation of the autonomic system has been reported to occur during cognitive assessment (Britton et al., 2008; Nicolini et al., 2014; Luque-Casado et al., 2016). Neuroimaging studies have already shown prefrontal cortex activation during cognitive functioning and autonomic system activation (Lane et al., 2009). Other studies have also evaluated ANS modulation using a combination of cardiovascular reflex tests in patients with Alzheimer’s disease (AD) and/or mild cognitive impairment (MCI), but contrasting data were reported. Depressed (de Vilhena Toledo and Junqueira, 2008), increased (Algotsson et al., 1995), and unchanged (Struhal et al., 2016) parasympathetic functions and increased sympathetic functions (de Vilhena Toledo and Junqueira, 2008) were observed. Moreover, only a few studies investigated autonomic function using the HRV. Zulli et al. (2005) evaluated the effect of HRV in subjects affected by AD and MCI using 24-h ECG monitoring. This latest study showed that both the LF and HF power spectral mean values were significantly lower in patients with AD compared to patients with MCI. In a cross-sectional study performed in 5-min electrocardiographic recordings focusing on changes from baseline to stimulation of sympathovagal balance, Nicolini et al. (2014) showed a reduced autonomic function in MCI patients. Unfortunately, the authors did not demonstrate the feasible occurrence of autonomic dysfunction in basal condition. da Silva et al. (2018), using a meta-analysis approach, showed a greater degree of autonomic dysfunction in patients affected by dementia. In contrast, lower levels of resting parasympathetic function have been related to poorer attention and executive function tasks during both stressful and non-stressful situations (Hansen et al., 2009). Unfortunately, there are some pitfalls in all the studies. First, multiple protocols used a meta-analysis approach, while others studied HRV by analyzing short-term measurements based on 5-min recordings, which does not correspond to a complete view of daily life. As reported by Shaffer and Ginsberg (2017), the 24-h HRV recordings achieve greater predictive power than short-term measurements. Therefore, our study design provided 24-h ECG recordings focusing on time-domain indices and on frequency-domain measurements. Although the biological interpretation of these indices is complex and debated, and many other mechanisms may influence these indices during long-term HRV analysis, our findings showed an increased sympathetic tone as expressed by high levels of LF, the LF/HF ratio, SDNN, and SDANN and as confirmed by the positive and significant correlation between the time-domain indices and the frequency-domain parameters with better cognitive performance. It is also true that the LF/HF ratio has long been considered as an index of sympathovagal balance, but this viewpoint has also been strongly criticized (Billman, 2013) because the physiological bases are not clear (Laborde et al., 2017). However, our data showed both high levels of LF and the LF/HF ratio, but, in particular, showed the significant correlation between the MMSE and MoCA cognitive tests and SDNN, representing this latter parameter as the gold standard for a more accurate evaluation of total circadian HRV (Singh et al., 2018a).

Besides the structural and functional changes in the ANS, the age-related brain functional declines also involve changes in receptor numbers and function. In fact, the number of intracerebral nicotinic receptors remarkably decreased in the elderly (Billman, 2013). This reduction can be compounded by the degeneration of cholinergic neurons in the basal forebrain, as occurs in AD (Gannon et al., 2015).

So far, the involvement of the cholinergic system in learning and memory processes is controversial. A possible role of receptors of the locus coeruleus (LC) in the pathogenesis of AD is supported by preclinical and clinical data showing that LC degeneration plays a significant role in AD pathogenesis (Mravec et al., 2014). The use of β-blockers also seems to support such hypothesis. The β-adrenergic receptors are involved in the enhancement of memory after emotional experiences (Gliebus and Lippa, 2007), and adrenergic signaling plays a role in the retrieval of intermediate-term contextual and spatial memories because the hippocampus receives dense input from adrenergic terminals (Ross et al., 2015). We cannot exclude that our data may be influenced by the use of β-blockers. However, our results were independent of the use of β-blockers, although used by patients in therapy for a long time, in turn producing pharmacological control of the disease itself.

Although females and males did not differ in numbers, our results show that males have better cognitive performance than females. Furthermore, according to gender, significant differences in the frequency-domain indices and in the analysis of the time-domain parameters were found. Recent studies showed gender influences on HRV indices (Voss et al., 2015). The mechanisms underlying the gender differences in autonomic function are not clear. The difference could be due to a decline in autonomic balance related to the expected decrement in estrogen (Voss et al., 2015), suggesting that, at least partly, hormonal factors play a role. This result is consistent with our data, which found higher SDNN and LF in males than in females.

Obesity is another aspect that needs to be emphasized. Indeed, it is known that obesity, cardiovascular disease, and diabetes have a detrimental effect on cognitive functioning (Yaffe et al., 2004; Whitmer et al., 2005). Moreover, the association between obesity and HRV is not completely clear. Most studies in the literature have shown that obesity is associated with a reduced HRV, indicative of a poorer ANS function. However, many other studies have shown that fat with central distribution, assessed with the measurement of waist circumference (Licht et al., 2010), is strongly associated with HRV compared to the general adiposity detected with BMI measurement. Our results demonstrated that all patients participating in the study were slightly overweight, while no patient was obese. Also, a normal BMI or being slightly overweight was associated with better cognitive performance. These data are in agreement with literature data showing that a low BMI reflecting poor health conditions and a malnutrition status are associated with a lower MMSE score or a decline in cognitive functioning (Sanders et al., 2016). Moreover, adipose tissue secretes a number of hormones (leptin, adiponectin, resistin, and visfatin) and inflammatory markers (TNF-α and IL-6) having the control function on the sympathetic nervous system and immune system, directly stimulating the sympathetic nervous system in the hypothalamus and/or inducing a state of low-grade inflammation that, in turn, can also stimulate the sympathetic nervous system (Ouchi et al., 2011).

Although the specific role of hypertension on cognitive function is known, our data do not provide any evidence for a negative impact of BP on cognition. Observational studies suggest that hypertension causes a vascular injury effect and cerebrovascular damage (Iadecola et al., 2016). In addition, chronically elevated BP exacerbates AD, contributing to dementia. However, judicious treatment of hypertension seems justified to protect vascular health and, consequently, brain health (Iadecola et al., 2016). These findings are important because they increase the possibility that treatment of hypertension may also contribute to reducing the development or progression of AD (Gorelick et al., 2011). Indeed, our results could be justified by compensated BP levels, values that were not far from the targets suggested by the American College of Cardiology guidelines (Whelton et al., 2018), thus exerting a neutral effect on cognition.

Our study is strengthened by the rigorous assessment of cognitive function that included the use of two validated cognitive screening instruments. The MMSE has been the most widely used cognitive screening instrument, while the MoCA is a relatively new cognitive screening instrument (Nasreddine et al., 2005; Mitchell, 2009; Santangelo et al., 2015; Trzepacz et al., 2015). Conversely to MMSE, the MoCA may be more sensitive to the assessment of complexity in executive function (Saczynski et al., 2015). Our data show potential advantages of the MoCA compared to the MMSE both in the assessment of cognitive performance and in affirming a stronger correlation between better cognitive performance and activation of the sympathetic system.

To the best of our knowledge, this is the first study investigating the association between long-term HRV analysis, as a measure of autonomic functioning, focusing on time-domain indices and global cognitive performance in a representative real-life community elderly population. In the current study, cognitive performance was significantly correlated with all the parameters representative of the sympathetic nervous system (SDNN, SDANN, and LF/HF ratio), but not with the parameters representative of the parasympathetic nervous system (HF and RMSSD). This finding suggests that long-term changes in heart rate, mediated by ANS, affect cognitive functioning mainly through the sympathetic system branch than the parasympathetic system branch. Thus, the evaluation of how sympathetic autonomic regulation was associated with the overall cognitive function in elderly patients would provide an opportunity to explain how the cognitive neuronal framework functions.

Finally, the present study has several limitations that must be considered. First, this is a cross-sectional study that does not allow for causal interpretation, and, second, this study includes a small sample size of participants. However, the findings need to be confirmed in larger and longitudinal studies.

The datasets for this manuscript are not publicly available due to agreements involved within informed consent from participants. Requests to access the datasets should be directed to MR, mariarosaria.rizzo@unicampania.it.

The studies involving human participants were reviewed and approved by Ethics Committee of University of Campania Luigi Vanvitelli—Napoli. The patients/participants provided their written informed consent to participate in this study.

All authors’ specific areas of contributions are listed below. MB and MR: study concept, design, analysis and interpretation of data. AD, RP, RF, and MR: acquisition of data. AD and MR: drafting of the manuscript. MB, AG, and MR: critical revision of the manuscript for important intellectual content.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.