Introduction
Primary Sjögren's syndrome (pSS) is a systemic autoimmune disease that is characterized by exocrine failure of salivary and lacrimal glands, in addition to a wide range of extraglandular features. Many clinical features of pSS are also features of autonomic neuropathy, which has been documented in pSS [
1]. Functional autoantibodies that target muscarinic acetylcholine receptors have been identified in the sera of patients with pSS, and these may represent an important mechanism in the production of sicca symptoms, bladder irritability and gastrointestinal symptoms [
2,
3]. Cardiovascular autonomic responses are a validated measure of autonomic nervous system function, and analysis of heart rate variability (HRV) provides additional information about parasympathetic and sympathetic activity. Studies in this area have yielded variable results in pSS patients, depending on the population studied and methodology applied. Some reported no autonomic dysfunction [
4,
5] whereas others found disturbance of the parasympathetic system [
6,
7] or both parasympathetic and sympathetic nervous systems [
8‐
10]. The aim of this study was to compare objectively autonomic nervous system function in pSS patients with that in control individuals, and to correlate the findings with objective measures of secretomotor function and self-reported symptoms.
Materials and methods
Study participants
Female pSS patients were recruited consecutively from the Rheumatology Clinic at The Queen Elizabeth Hospital. All patients met the revised 2002 American European Consensus criteria [
11] for pSS. Age-matched, population-based female control individuals were recruited from the local community. Exclusion criteria for the study included diabetes, ischaemic heart disease, current anticholinergic medication, or a serious medical illness. Seven participants from each group were taking antihypertensive medications, which were withheld for 24 hours before testing. Classes of medications used by pSS patients and control individuals (respectively) were as follows: angiotensin-converting enzyme inhibitors (one and two participants), β-blockers (one and two participants), angiotensin receptor blockers (three and four participants), diuretics (two and three participants), calcium channel blockers (three and no participants) and hydrallazine (one and no participants). Five pSS patients were using pilocarpine, which was withheld for 24 hours before testing. Two control individuals in whom cardiac arrhythmias were detected during cardiovascular reflex testing were excluded from the analysis, and 27 pSS patients and 25 control individuals were included in the final study.
Eighteen (67%) of the pSS patient sera were positive for Ro/La autoantibodies, and of those tested nine out of nine (100%) were negative for cryoglobulins and two out of 17 (12%) had low C3 or C4 levels. Eleven patients (41%) had Raynaud's phenomenon, and of those tested 13 out of 14 (93%) had a positive labial salivary gland biopsy. The average age of onset of disease was 48 years (range 29 to 73 years) and the average disease duration was 13 years (range 2 to 29 years).
All participants gave informed, written consent for the study, and the study was approved by the North Western Adelaide Health Service Ethics of Human Research Committee.
Study protocol
All testing was conducted in the morning and in a standardized manner. Hypertensive medications were withheld 24 hours before testing, participants abstained from caffeine and cigarettes from the previous evening, and artificial tears from waking that morning. Patients were instructed to have an early light breakfast, and testing did not commence until the patients had been fasting for more than 1 hour.
Participants initially completed the FACIT-F (Functional Assessment of Chronic Illness Therapy-Fatigue), a 13-item assessment of fatigue [
12], and the COMPASS (Composite Autonomic Symptom Scale) questionnaire [
13]. A 15-minute unstimulated whole salivary flow and Schirmer's-I test were performed as objective measures of dryness. Sicca was defined as an unstimulated salivary flow test of under 1.5 ml in 15 minutes and/or a Schirmer's test with under 5 mm wetting in both eyes over 5 minutes.
Physical activity levels were measured using the short telephone form of the International Physical Activity Questionnaire [
14]. Participants were classified as HEPA (health enhancing physical activity) active if they achieved either of the following: activity of vigorous intensity on at least 3 days, achieving a minimum of at least 1,500 multiples of resting metabolic state (MET)-minutes/week; or 7 or more days of any combination of walking, or activity of moderate intensity or vigorous intensity achieving a minimum of at least 3,000 MET-minutes/week. All study participants were examined by a neurologist. Abnormalities were observed in five pSS patients: bilateral carpal tunnel syndrome (in one patient), an old minor cerebrovascular accident (in one), unilateral benign essential tremor (in one), peripheral neuropathy (in two) and facial numbness (in one).
Cardiovascular reflex testing
Noninvasive, beat-to-beat measurements of systolic blood pressure (SBP), diastolic blood pressure, mean blood pressure (MBP), heart rate and heart period (RR interval) were recorded during all manoeuvres using the Finapres™ (Ohmeda. Louisville, Colorado, USA) finger arterial pressure monitoring system [
15].
Manoeuvres (described below) were performed in the following order: supine rest, postural change, Valsalva manoeuvre (seated), isometric grip (seated) and controlled breathing (seated). Participants rested for several minutes between successive manoeuvres and between replicates of manoeuvres.
Supine rest
Participants lay quietly on a bed, and once settled recording commenced for a period of 5 minutes. Brachial blood pressure was measured using a digital blood pressure monitor approximately 1 minute before completion of this period.
Postural change
Participants were asked to stand quickly and remain standing quietly for a period of 6 minutes. Brachial blood pressure was measured with a digital blood pressure monitor at 2 and 5 minutes after standing.
Valsalva manoeuvre (seated)
Participants blew into a closed tube with a small leak, maintaining an expiratory pressure of 40 mmHg for 10 seconds. Beat-to-beat measurements were monitored for a period of 1 minute after release of expiratory pressure. This manoeuvre was performed three times.
Isometric grip (seated)
Patients gripped a dynamometer for 3 minutes, maintaining a contraction pressure of one-third of their maximum voluntary contraction pressure.
Controlled breathing (seated)
Participants maintained a controlled, even breathing rate of six breaths/minute over a period of 1 minute. This manoeuvre was performed three times.
Cardiovascular reflex test analysis
Five standard parameters of the cardiovascular autonomic test were estimated [
16]: supine to standing ΔSBP, supine to standing 30/15 ratio, isometric grip ΔMBP, Valsalva ratio, and breathing E/I ratio.
Supine to standing ΔSBP
The supine to standing ΔSBP was calculated as the brachial SBP 5 minutes after standing minus the supine brachial SBP.
Supine to standing 30/15 ratio
The supine to standing 30/15 ratio is the ratio of the longest RR interval near to the 30th beat after standing to the shortest RR interval near to the 15th beat after standing.
Isometric grip ΔMBP
The isometric grip ΔMBP is the MBP (Finapres™; Ohmeda) at the end of the 3-minute grip period minus the MBP just before commencing grip.
Valsalva ratio
The Valsalva ratio is the ratio of the longest RR interval immediately after strain to the shortest RR interval during strain. The geometric mean was estimated from three replicates.
Breathing E/I ratio
The breathing E/I ratio is the mean of the longest RR intervals during each expiration divided by the mean of the shortest RR intervals during each inspiration [
17]. The geometric mean was estimated from three replicates.
Heart rate variability analysis
Time domain analysis
The mean and time domain HRV parameters of the RR interval [
18], including the standard deviation (SDNN), the proportion of successive intervals differing by more than 50 ms (pNN50) and the standard deviation of the differenced RR series (RMSSD), were calculated for both supine and standing (beginning 1 minute after standing) positions over a 5-minute recording interval.
Frequency domain (spectral) analysis
Cross-spectral analysis of the beat-to-beat RR and SBP data was performed on both supine and standing 5-minute recording intervals. The data were interpolated at a frequency of 2 Hz using cubic spline interpolation. Exact length cross-spectral Fourier analysis was performed using the Time Series module of Statistica (v6.1; Statsoft Inc., Tulsa, Oklahoma, USA), with a taper of 15% and a Hamming window of width five to estimate the spectral densities. Power was calculated by integration of the spectral densities over the frequency ranges of 0.04 to 0.15 Hz (low frequency [LF]) and 0.15 to 0.4 Hz (high frequency). The gain, essentially a regression coefficient for the SBP variability as a predictor of RR variability, was used as a measure of baroreflex function [
19]. This was estimated as the total cross-amplitude power divided by the total SBP power over the relevant frequency range.
Statistical analysis
The cardiovascular reflex test scores were analyzed as continuous variables rather than classified as normal, borderline and abnormal, as initially described [
16]. This is because there was a substantial age dependence in these scores, also recognized in other studies [
20], that is not incorporated into the classification criteria. All analyses were performed by analysis of variance (ANOVA), multivariate ANOVA and repeated measures ANOVA, as indicated. With the exception of blood pressure, measurements of autonomic function and HRV, which were either ratios or rate measurements, were log-transformed before analysis to normalize their distribution. Reported results are for an age-unadjusted analysis, but they did not differ from results for an age-adjusted analysis. Many of the cardiovascular reflex test and COMPASS domain scores were highly correlated with each other. Therefore principal component factor analysis was employed to detect structure in the relationships between parameters. All factors with a minimum Eigan value of 1 were extracted. All analyses were performed using Statistica (v6.1; Statsoft Inc.).
Discussion
In this study we demonstrated evidence, obtained from both self-reported symptoms and objective cardiovascular reflex testing, of mild autonomic dysfunction in pSS. From the cardiovascular reflex testing, there was evidence of multiple autonomic disturbances in pSS relating to decreased HRV, decreased blood pressure variability and an increased heart rate (tachycardia), which were most evident in response to postural change. There was a strong trend toward an association between decreased HRV and increased severity of the secretomotor, orthostatic, bladder, gastroparesis and constipation self-reported symptom cluster in pSS patients. This symptom cluster was also associated with fatigue and reduced unstimulated salivary flow, and therefore may be an important component of the clinical spectrum of this disease. Of note, we [
21] and others [
23] previously reported increased bladder symptoms in pSS patients; furthermore, a high frequency of both delayed gastric emptying and decreased bladder detrusor muscle tone has also been observed in pSS patients [
24].
Previous studies addressing cardiovascular autonomic function in pSS have yielded conflicting results, although a pattern is emerging. Two studies using 24-hour Holter monitoring [
4,
5], which reflects tonic balance, both reported negative results. In contrast, a number of studies of provoked cardiovascular responses or short-term HRV [
6‐
10,
25‐
27] identified abnormalities in pSS, although not all of these were controlled studies or used appropriate age-adjusted criteria for interpretation of abnormal test results. Four controlled studies [
8‐
10,
27] found abnormalities in either the 30/15 ratio or blood pressure response to postural challenge, as we also observed, and two controlled studies [
6,
26] identified reduced HRV/blood pressure variability in pSS patients by using spectral analysis. Our observation of a relative tachycardia in pSS patients has not previously been reported. Four studies [
8‐
10,
27] also reported a decreased breathing E/I ratio in pSS patients, which we did not observe. However, this test is also influenced by breathing tidal volume, which may have differed between study participants and potentially confounded the results. Similar to other studies [
10], we did not observe an association between cardiovascular reflex test scores and objective measures of sicca in pSS patients, but this is the first study to both examine and report an association between objective measures of sicca and self-reported autonomic symptoms in pSS patients.
Potential mechanisms of autonomic dysfunction in SS include T-cell infiltration and destruction of ganglions and nerves [
28], cytokine-induced inhibition of neuropeptide secretion from nerve endings [
29], immune complex-mediated inflammation (although few pSS patients in this study exhibited cryoglobulins and/or low C3 or C4, which might indicate immune complex deposition), and pathogenic autoantibodies targeting receptors relevant for autonomic functioning [
30]. IgG autoantibodies, which inhibit the function of type 3 muscarinic receptors (M3Rs), have been described in pSS patients [
3,
31]. Importantly, these autoantibodies inhibit salivary secretion [
32], bladder detrusor muscle contraction [
3], and colon contractions [
33]
in vitro. Evidence that lower urinary tract symptoms in pSS are autoantibody mediated comes from passive transfer of SS immunoglobulin or rabbit anti-M3R to mice, which produces the phenotype of overactive bladder [
2]. Furthermore, neutralization of anti-M3R autoantibodies by intravenous immunoglobulin led to improvement in bladder and bowel autonomic symptoms in patients with autoimmune diseases [
34]. Therefore, pathogenic M3R autoantibodies are strongly implicated in the pathophysiology of the cluster of secretomotor, bladder, gastroparesis, constipation and orthostatic autonomic symptoms in pSS patients observed in the present study.
Pathogenic M3R autoantibodies may also potentially influence cardiovascular autonomic responses. Although the M2R subtype is the numerically and functionally predominant muscarinic receptor in the heart, recent studies have provided compelling and solid evidence in support of the important roles of M3R in regulating and maintaining cardiac function and heart disease [
35]. Furthermore, given the close structural similarity between the M2R and M3R, it is likely that the autoantibodies may be cross-reactive.
Muscarinic receptor-mediated cardiac parasympathetic activity is essential for regulating heart rate [
35] and HRV [
36]. Furthermore, vasodilatory responses to cholinergic stimuli are diminished in M3R knockout mice [
37] and in pSS patients [
38], which may – at least in part – underpin the reduced blood pressure variability observed in the present study. Cardiovascular reflex tests are traditionally interpreted as an indication of parasympathetic or sympathetic function, but our results are better interpreted as multiple autonomic disturbances in pSS relating to decreased HRV, decreased blood pressure variability and increased heart rate, which are likely to reflect a disturbance of parasympathetic/sympathetic balance.
Conclusion
We have confirmed the presence of mild autonomic dysfunction in pSS patients, as measured by both self-reported symptoms and objective assessment. We have identified an important cluster of self-reported secretomotor, orthostatic, bladder, gastroparesis and constipation symptoms in pSS, which correlate with increased fatigue and reduced serum salivary flow. Cardiovascular reflex testing reveals multiple abnormalities that reflect probable disturbance of parasympathetic/sympathetic balance. Although pathogenic M3R autoantibodies remain a strong candidate for the underlying pathophysiology in pSS, it is not yet possible to test this hypothesis, because practical assays for anti-M3R autoantibody detection remain elusive.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
FZC recruited patients, was responsible for data management, and carried out autonomic testing and drafted the manuscript. SL assisted with autonomic testing, performed statistical analysis and assisted with manuscript preparation. TL performed additional patient recruitment and autonomic testing. HK and AT were responsible for test selection and training in autonomic testing. KB assisted with study design and performed neurological examinations. SP assisted with the study design and patient ascertainment. MR conceived of the study, participated in its design and coordination and drafting of the manuscript. All authors read and approved the final manuscript.