2.5. Evaluation of autonomic nervous system function

2.5.1. General aspects

In a patient with a suspected autonomic disorder the major aims of the investigation are to determine the normality of autonomic function, to assess the degree of dysfunction with an emphasis on the site of the lesion and on the functional deficit, and also to ascertain whether the abnormality is of the primary variety or secondary to recognized disorders, as the prognosis and management may depend on the diagnostic category. Investigatory methods are available for recording cardiovascular, sudomotor, gastrointestinal, genitourinary, respiratory and pupillary autonomic functions. They are in routine clinical as well as in research use to study both central and peripheral ANS functions.

It has been observed that the heart is one of the most delicate organs reflecting cardiovascular autonomic regulatory function. It is well known that the HR is not uniform but the interval between the two R spikes in the ECG fluctuates. Respiration, neurohumoral factors and cardiovascular regulation are the main effectors on the RR interval. HRV reflects the competence of this regulatory system, diminished HRV being a sign of dysfunction of the ANS. During the past decades, noninvasive bedside techniques have been developed to study the regulation of HR and BP. (Benarroch 1997c)

2.5.2. Cardiovascular autonomic reflexes

Assessment of cardiovascular autonomic reflexes is a critical element in evaluation of autonomic function in humans. Cardiovascular autonomic reflexes in humans are essential for the maintenance of arterial BP during the orthostatic stress adopting a standing posture, and for preventing wide fluctuations of arterial BP in response to stress, exercise, and other adaptive responses. (Benarroch 1997d). To examine cardiovascular reflexes, a standardized laboratory test pattern, in which HR and BP responses at rest and to certain stimuli are measured, is often used. The method is easy to perform, standardized measures make interindividual comparisons possible and it has indeed an established role as the most used method assessing autonomic functions. The test pattern includes measurement of HR and BP during normal and deep breathing, the Valsalva maneuver, the tilt test and the isometric work test.

The arterial baroreflexes exert buffering influence on the magnitude of centrally induced variations of arterial pressure during day and night. They act by reducing arterial BP oscillations, and their activity is manifested by beat-to-beat variations of the HR opposite in direction to the changes in arterial BP (Shepherd & Mancia 1986). If the arterial baroreflex is acting normally, there are small arterial BP oscillations and large HR oscillations, whereas the opposite occurs in patients with baroreflex failure.

Assumption of an upright posture produces blood to shift downwards, creating a decrease in stroke volume. The circulatory adjustment to orthostatic stress is rapid in healthy subjects. Several mechanisms including the arterial baroreflexes, cardiopulmonary reflexes, venoarteriolar reflexes, and the vasopressin (AVP) and renin-angiotensin systems contribute to the maintenance of postural normotension (Wieling & Van Lieshout 1993). Mechanoreceptors in the atria and ventricles innervated by vagal afferents exert tonic inhibitory influence on sympathetic outflow and AVP release (Shepherd & Shepherd 1992).

The respiratory sinus arrythmia is evaluated while at rest and during deep breathing at a rate of 6 breaths per minute that produces the maximum sinus arrythmia (Angelone & Goulter 1964, Borgdorff 1975, Hirsch & Bishop 1981, Piha et al. 1988). The registration of HR fluctuation during normal and deep breathing is a sensitive detector of autonomic dysfunction (Mackay et al. 1980).

In the Valsalva maneuver the respiratory strain increases intra-thoracic and intra-abdominal pressure altering hemodynamic and cardiac functions (Nishimura & Tajik 1986, Benarroch 1991). The Valsalva ratio is the most commonly used test parameter that calculates the ratio of the longest RR interval after the blowing and to the shortest RR interval during the blowing (Levin 1966). Both SNS and PNS control autonomic responses during the Valsalva maneuver (Sandroni et al. 1991) and continuous BP monitoring increases the sensitivity of the test (Ravits 1997).

Normally, orthostatic stress produces an increase in HR, increase in diastolic and decrease in systolic BP accompanied by increase in plasma norepinephrine and muscle sympathetic activity. Thereafter a relative bradycardia follows due to vagal reflexes (Ewing et al. 1978, Borst et al. 1982). The HR changes to standing are expressed as the 30:15 ratio (Ewing et al. 1980). The BP is monitored during the test continuously or serially and the largest drop (or lowest increase) is quantified. Orthostatic hypotension means a reduction of systolic BP of at least 20 mmHg or diastolic BP of at least 10 mmHg within 3 minutes (Consensus Statement 1996). Even though the physiological responses to passive tilting are not identical to standing up, the pulse response is seen if the tilting is quick and extended up to 90° (Sundquist et al. 1980, Myllylä et al. 2000).

During the isometric work test the BP reaction to sustained handgrip is measured. The mechanism involves the exercise reflex that withdraws parasympathetic activity and increases sympathetic tone. Normally the diastolic BP raise is more than 15 mmHg. The participant´s age does not affect the BP responses to isometric work (Goldstraw & Warren 1985), but the responses are greater in male than female (Piha 1993, Khurana & Setty 1996).

The cardiovascular reflex parameters based on HR fluctuation are age and HR dependent and must therefore be adjusted for age and baseline RR interval. The results of two subtests outside the 95% confidential limits of the control subjects has been considered as a clinically abnormal finding, but also a marked decrease of BP with fainting after standing or tilting, as a sole finding, is enough for the diagnosis.

Abnormalities in the cardiovascular reflexes have been detected in a majority of patients with multisystem atrophy and primary autonomic failure (Cohen et al. 1987, Ravits et al. 1995). Cardiovascular responses have also been used to evaluate autonomic dysfunction in diabetic, uremic and hereditary neuropathies (Bennet et al. 1977, Ewing et al. 1978, Ewing et al. 1981, Low et al. 1986, Shahani et al. 1990, Wang et al. 1994), pulmonary diseases (Pagani et al. 1996), amyotrophic lateral sclerosis (Pisano et al. 1995), cerebellar and extrapyramidal disorders (Turkka et al. 1987, Sandroni et al. 1991, Haapaniemi et al. 2000), stroke (Korpelainen et al. 1994), migraine (Havanka-Kanniainen et al. 1988), multiple sclerosis (Senaratne et al. 1984, Vita et al. 1993) and chronic alcoholism (Yokohama et al. 1991, Monforte et al. 1995).

Power spectral analysis of HR and arterial BP variations, including analysis of signal coherence both at rest and in response to tilt, and other maneuvers, provides additional information on the sympathetic and cardiovagal components of cardiovascular reflex responses.

2.5.3. Ambulatory ECG and analysis of heart rate variability

Although respiratory sinus arrythmia has been noticed already in 1733 by Hales (Singer & Underwood 1962), only the development of high resolution ECG and digital computers has made it possible to measure more subtle fluctuations of HR and BP. The HRV analysis from ambulatory ECG recording is now an important tool in evaluating cardiovascular autonomic regulation. Information about tonic autonomic effects on the heart can be obtained by the traditional time and frequency domain measures based on linear fluctuations of HR (Huikuri et al. 1995). However, as the behaviour of the heart is not only linear but also chaotic, new methods based on non-linear dynamics and fractal analysis have been developed (Goldberger & West 1987, Denton et al 1990, Pincus & Goldberger 1994).

Diminished HRV and loss of its circadian oscillation are particularly associated with cardiac arrythmogenic death in patients with prior heart problems, but the predictive value is not very good, with the sensitivity being 50%-80% and the specificity 60-90% in an individual patient (Huikuri et al. 1995c). Age, gender, physical activity and certain drugs affect HRV (Huikuri et al 1995c, Task Force 1996).

The time domain analysis of HR fluctuation is conventionally based on indices on statistical operations on RR intervals. The most widely used index is the standard deviation of all normal to normal RR intervals (SDNN) over a 24-hour period, reflecting primarily the very low frequency (VLF) fluctuation in HR behaviour (Rosenbaum & Race 1968). It has been used as a predictor of mortality in post myocardial infarction patients (Kleiger et el. 1987).

Spectral analysis of HRV inspects the frequency-specific oscillations of HR fluctuation and decomposes a series of sequential RR intervals into a sum of sinusoidal functions at different amplitudes and frequencies (Akselrod et al. 1981). The amplitude of the HR fluctuations at different oscillation frequencies is presented as power spectrum.

The Fast Fourier transformation and autoregressive analysis are the most commonly used methods for transforming signals to the frequency domain. The power spectrum is usually divided into three or four frequency bands as follows: ultra low frequency (ULF) <0.0033 Hz, VLF from 0.0033–0.04 Hz, low frequency (LF) from 0.04-0.15 Hz and high frequency (HF) from 0.15-0.4 Hz (Task Force 1996). The HF fluctuation of RR intervals mainly reflects the cardiovagal modulation and the inspiratory inhibition of vagal tone and the LF and VLF bands reflect sympathetic excitation, sympathovagal balance, and arterial BP oscillations and thermoregulation (Rosenbaum & Race 1968, Dwain & Eckberg 1997, Pagani et al. 1997).

The Poincaré plot is a geometrical method of HRV analysis. It is a diagram (scattergram) that plots each RR interval as a function of the previous RR interval. These plots can be interpreted visually and quantitatively, where the instantaneuous beat to beat RR interval variability (SD1) and the SD of continuous long-term RR interval variability (SD2) are analysed (Huikuri et al. 1996, Tulppo et al. 1996). SD1 describes the magnitude of beat to beat RR interval variability reflecting vagal modulation of the HR. SD1 correlates with the HF spectral component relatively strongly. On the other hand, SD2 has correlations with the magnitude of the LF and VLF spectral components and it describes the long-term RR interval fluctuations. Unlike the spectral analysis, the Poincaré plot method is not interferred by stationary irregularities and trends in the RR intervals, and it may therefore be more suitable for HRV analysis from uncontrolled ambulatory ECG recordings (Tulppo et al. 1996).

Analysis of non-linear dynamics based on chaos theory and fractal mathematics have opened new approaches for studying and also understanding the HR behaviour (Goldberger & West 1987, Goldberger 1996). With these methods, estimation of the correlation properties and complexity of HRV can be performed. Analysis of fractal-like properties have been used to detect abnormalities of HR dynamics in various cardiovascular disorders (Bigger et al 1996, Huikuri et al 1998, Mäkikallio et al 1998, Mäkikallio et al 1999a, Mäkikallio et al 1999b). Analysis of 1/f characteristics i.e. the inverse power-law slope has been shown to be an independent predictor of survival in the elderly as well as in patients with impaired left ventricular function (Brouwer et al. 1996, Ho et al. 1997, Huikuri et al. 1998). The physiological backgroud for this method is not completely undestood, but it is influenced by the autonomic input to the heart, as the slope of the power law relation is especially deep in denervated, transplanted hearts (Bigger et al. 1996).

The parameter approximate entropy (ApEn) quantifies the regularity or predictability of time series data. This reduced complexity of HR dynamics has been found in sickness of the neonates and in patients with postoperative complications after cardiac surgery, as well as in patients with chronic liver disease (Pincus & Viscarello 1992, Fleisher et al. 1993, Fleisher et al. 2000). The decrease of these dynamic parameters, especially the value of ApEn, are also suggested to be associated with spontaneous onset of paroxysmal atrial fibrillation (Vikman et al. 1999).

Altered HRV may also be associated with certain neurological diseases. Acute cerebrovascular diseases and brain injuries frequently cause cardiovascular complications and also decrease HRV. It has been shown that all the spectral components of HRV are suppressed after hemispheral and brainstem cerebral infarctions, and this may be long lasting (Korpelainen et al. 1996a, Korpelainen et al. 1996b). A similar pattern of HRV is also seen in brain-dead patients (Kita et al. 1993, Freitas et al. 1996) and in patients with severe brainstem injury (Novak et al. 1995). Moreover, it has been shown that circadian fluctuation of HR variability is reversibly abolished in the acute phase of ischaemic stroke. This reversible abolition and the loss of the relative nocturnal dominance of the HR variability may contribute to the high incidence of cardiac arrythmias and other cardiovascular complications after acute stroke (Korpelainen et al. 1999). Altered HRV has also been observed in Parkinson´s disease (Haapaniemi et al. 2001), and multiple sclerosis (Frontoni et al. 1996). In addition, there is also evidence that HRV is altered in neonates who later experience sudden infant death (Rosenstock et al. 1999).