- Email: [email protected]
Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease: A meta-analyses study
Author’s Accepted Manuscript
image
Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A metaanalyses study Sella A. Provan, Daniela Schäfer Olstad, Erik E. Solberg, Geir Smedslund, Hanne Dagfinrud www.elsevier.com/locate/bios
PII: DOI: Reference:
S0049-0172(17)30517-6 https://doi.org/10.1016/j.semarthrit.2017.11.010 YSARH51278
To appear in: Seminars in Arthritis and Rheumatism Cite this article as: Sella A. Provan, Daniela Schäfer Olstad, Erik E. Solberg, Geir Smedslund and Hanne Dagfinrud, Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A meta-analyses study, Seminars in Arthritis and Rheumatism,doi:10.1016/j.semarthrit.2017.11.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A meta-analyses study.
Sella A Provan1, Daniela Schäfer Olstad2, Erik E Solberg3, Geir Smedslund2, Hanne Dagfinrud2 1
Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway, 2 National Resource
Centre for Rehabilitation in Rheumatology, Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway 3 Department of Medicine, Diakonhjemmet Hospital, Oslo, Norway.
Corresponding author: Sella Aarrestad Provan MD. PhD. Department of Rheumatology Diakonhjemmet Hospital PB 23 Vindern 0319 Oslo, Norway [email protected]
This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation Disclosures: The authors SAP, EES, GS, HD declare no conflict of interest. DSO is currently employed by Polar Electro, Finland. Funding statement: This work was sponsored by the Research Fund at Diakonhjemmet Hospital and by Extra-Stiftelsen, Norway. The sponsors have not been involved in the data-analyses, writing or submission of this paper. .
1
Abstract Background: Rheumatoid arthritis (RA) and spondyloarthritis (SpA) are inflammatory joint disorders (IJD) with increased risk of cardiovascular disease (CVD). Autonomic dysfunction (AD) is a risk factor for CVD, and parasympathetic AD is linked to key features of IJD such as inflammation, physical inactivity and pain. Heart rate variability (HRV) is a marker of cardiac AD. The study objective was to compare parasympathetic cardiac AD, measured by HRV, between patients with IJD and healthy controls, using metaanalysis methodology, and to examine the impact of inflammation, physical inactivity and pain on HRV in IJD.
Methods: Medline, Embase and Amed were searched. Inclusion criteria were adult casecontrol studies published in English or a Scandinavian language, presenting HRV data in IJD. Two measures of HRV and 3 from the Ewing protocol were selected: Square root of mean squared difference of successive R-R intervals (RMSSD), high frequency (HF), Ewing protocol; standing (E-S), breathing (E-B) and Valsalva (E-V). Patients with RA, SpA and healthy controls were compared separately using random-effects meta-analyses of standardized mean differences (SMD). Results: 35 papers were eligible for inclusion. For RMSSD the pooled SMD (95% CI) RA vs. controls was -0.90 (-1.35 to -0.44), for SpA vs. controls; -0.34 (-0.73 to 0.06). For HF pooled SMD RA vs. controls was -0.78 (-0.99 to -0.57), for SpA vs. controls; -0.04 (-0.22 to 0.13). All Ewing parameters were significantly lower in cases, except for E-V which was comparable between cases and controls in patients with RA.
2
Conclusion: Patients with IJD have cardiac parasympathetic AD which is related to inflammation. Key words: Autonomic dysfunction, heart rate variability, Inflammatory joint disease, rheumatoid arthritis, spondyloarthritis.
1. Background Cardiac autonomic dysfunction (AD) is linked to arrhythmias, all-cause mortality and death following myocardial infarction [1, 2]. There is a complex interplay between the immune and parasympathetic autonomic systems, with evidence of mutual regulation [3]. Parasympathetic AD is also linked to physical inactivity and is more prevalent in patients who suffer from chronic pain [4]. Rheumatoid arthritis (RA) and spondyloarthritis (SpA) are inflammatory joint disorders (IJD) with an excess risk of cardiovascular disease (CVD), arrhythmias and sudden death [5]. IJDs are characterised by joint inflammation and symptoms include pain, functional decline and restricted movement. A previous systematic review concluded that there is evidence of AD in approximately 60 % of patients with RA, but a meta-analysis of the data was not performed due to the heterogeneity of outcomes included in the review. [6]. The heart rate of the individual will vary on a beat to beat level throughout the day, influenced by the body’s physiological and psychological state and essentially under the control of the autonomic nervous system. The variation in heart rate has been the subject of academic interest for decades, and several techniques for quantifying the variation have been developed. Heart rate variability (HRV) is a measure of heart-rate variation and an estimate of cardiac autonomic dysfunction (AD) [1, 2] The objective of this study was to compare parasympathetic cardiac AD, measured by HRV, between patients with IJD and healthy controls, using meta-analyses methodology, and to examine the impact of inflammation, physical inactivity and pain on HRV in IJD. 3
2. Methods These meta-analyses were performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA statement).
2.1 Search Strategy A research librarian developed and conducted systematic searches in the following databases: Medline (from 1946 to August 2016), Embase (from 1980 to August 2016) and Amed (from 1985 to August 2016). 2.2 Inclusion criteria: Case-control studies on human subjects of any age, published in English, Norwegian, Danish or Swedish which present data on HRV according to an established methodology. Population: Patients who were diagnosed with an inflammatory rheumatic joint disease (RA or SpA). Included under the SpA diagnosis in this study are also patients with psoriatic arthritis (PsA) or ankylosing spondylitis (AS). 2.3 Exclusion criteria: Studies not fulfilling the inclusion criteria, double publications and studies presenting outcomes graphically or as categorical variables only were excluded. 2.4 HRV as a marker of autonomic dysfunction
HRV is the result of complex interactions between sympathetic and parasympathetic stimulation on the sinus node of the heart, coupled with input from pacemaker cells and the geometry of the heart in the individual [1]. HRV can be defined as the beat-to-beat variation in heart-rate (R-R interval (RRI)) quantified in time and frequency domains. Frequently 4
reported variables in the time domain are the square root of the variance of the RRI (SDNN) and Square root of the mean squared difference of successive RRI (RMSSD). In the frequency domain, also named power spectral analysis, the RRI (y-axis) is plotted against time of the corresponding heart-beat (x-axis). The spectral power for a given frequency is quantified by determining the area under the curve within the specific frequency range [7]. The spectral power can be estimated by the non-parametric fast Fourier transformation or by parametric autoregressive modelling. In the frequency domain, low frequency (LF) or high frequency (HF) components are frequently reported variables. A five minute recording is sufficient for analyses in both domains, although additional information can be gathered from 24-hour recordings[1]. The Ewing battery is a collection of five non-invasive tests originally designed to assess autonomic neuropathy in patients with diabetes [8]. Three tests quantify cardiac parasympathetic integrity by determining changes in RRI in response to the following stressors; standing up (E-S), deep breathing (E-B) and Valsalva manoeuvre (E-V), blood pressure response to standing up, and sustained handgrip. A summary of abbreviations of different HRV parameters described in this report is outlined in supplementary table 1. In order to limit the scope of the study we selected central markers of markers of parasympathetic cardiac autonomic function. Two HRV variables; RMSSD from the time domain and HF component from the frequency domain, and three tests from the Ewing battery; E-S, E-B and E-V, were selected
2.4 Screening Two of the authors independently judged each record from the search as either not relevant or possibly relevant. If at least one person scored a record as being possibly relevant, we ordered it in full text. The full texts were read independently by two of the authors (SAP and GS) and 5
each person rated them as either ‘include’ or ‘exclude’. A final decision was made by the third author (HD) if there was a conflict. 2.5 Risk of bias Two of the authors independently assessed risk of bias for each included study. Conflicts were solved through an independent assessment by a third author and through group discussions. For the risk of bias assessments we used a modified version of “The quality assessment for case-control studies checklist” from the National Heart, Lung, and Blood Institute [9] (Supplementary table 2). 2.6 Data extraction The following data were extracted from each included study: Publication year and country in which the study was conducted, number of subjects, matching procedure, diagnosis of patients, percent females included, average age, HRV outcomes, apparatus used, measures of disease activity when included in the analysis, other measures of autonomic function measured. When data were presented only as categorical variables or in figures, the authors were contacted whenever possible and continuous data were requested.
3. Theory and calculations For the comparison of the disease groups with the healthy controls, we conducted randomeffects meta-analyses using standardized mean differences. The analyses were run separately for patients with RA and SpA. Within each diagnostic category we ran separate meta-analyses for each of the variables RMSSD, HF, E-S, E-B, and E-V. Sensitivity analyses were performed: Forest plots were examined for heterogeneity and when significant heterogeneity was displayed outliers were removed in descending order from the pooled results until the level of heterogeneity became insignificant (p ≥ 0.05) and the meta-analyses repeated.
6
Additionally papers with a high risk of bias were removed and the meta-analyses repeated separately. Funnel plots were constructed to assess the risk of publication bias in the subgroups in which 10 or more studies had been identified. Suspected publication bias would appear as asymmetrical funnel plots due to the non-publication of small studies that found nonsignificant difference in HRV between patient and controls groups.
4. Results The final literature search was performed in August 2016. Supplementary table 3 presents the complete search strategy. Supplementary figure 1 presents a summary of the findings of literature search, screening and selection process. The database search identified 1015 records. In addition 120 records were found through searching grey literature (conference abstracts, internet searches) and bibliographies. After duplicates had been removed, 847 records remained, and these titles and abstracts were reviewed according to our study selection criteria. We considered 112 full texts in more detail. Two papers presented results from the Ewing battery as categorical outcomes and these outcomes were not included in the meta-analyses [10, 11]. One paper did not report standard confidence intervals of data and was excluded [12]. 35 papers were finally selected. Of the 35 papers included, 21 studied RA patients, 11 studied SpA and 3 papers studied both patient groups. Supplementary tables 4 and 5 present the characteristics of the studies concerning patients with RA and SpA respectively. Of the studies including patients with SpA, ten included patients diagnosed with AS, three included PsA patients and one included patients diagnosed with SpA specifically. The studies came from 18 countries of which Turkey and India had the greatest number of publications with nine and seven papers respectively. The average number of cases included was low, with a mean of 38 (range 10 to 7
207). The RA studies included more women than men, reflecting the epidemiology of the disease. Nine RA studies only included female subjects. Gender distribution was more similar in the SpA studies. The average age of cases in the RA studies was 48 years (range 31-63 years), while the SpA cases were younger [average 38 years (range 29-51 years)]. The Ewing battery was the most frequent HRV outcome, measured in 17 studies, short-term HRV was reported in 12, while 24-hour HRV was the outcome in 7 studies. (One study measured both short-term HRV and Ewing Battery). 4.1 Risk of bias Risk of bias decisions are presented in supplementary figure 2. All papers clearly differentiated between patients and controls, and the majority of papers had recruited patients and controls from a similar population and applied consistent criteria for the selection of participants. In three papers there was a high risk of bias regarding item 3, and in these cases there was a discrepancy in the exclusion criteria between the patient and controls populations. Blinding of the assessors was not described in the majority of studies and, thus, only four studies had a low risk of bias regarding item 4. We defined matching or adjustment for confounders as adequate if age and gender were statistically similar between groups, or adjusted for in statistical analyses. Ten studies had a high risk of bias due to insufficient matching or adjustments, the highest number in any item. For the last item we looked at how the examinations were standardised, and we considered whether the examinations were performed at the same time of day and under fasting conditions. Here three papers were judged to have a high risk of bias due to poor standardisation.
8
4.2 Studies comparing the HRV time and frequency domain variables between patients and controls The results are presented in Figure 1 (RA) and Figure 2 (SpA). In patients with RA eight studies reported RMSSD and the pooled results gave a SMD (95% CI) of -0.90 (-1.35- -0.44). For HF there were ten studies, with a pooled SMD of -0.78 (-0.99-0.57). The level of heterogeneity was significant for the RMSSD analyses. In patients with SpA, four studies reported RMSSD and the pooled SMD (95% CI) was -0.34 (-0.73- 0.06). Six studies reported HF with a pooled SMD of -0.04 (-0.22- 0.13). There was a non-significant level of heterogeneity for all three outcomes. 4.3 Studies comparing results of the Ewing battery between patients and controls The results are presented in Figure 3 (RA) and Figure 4 (SpA). In patients with RA nine studies reported E-S. The pooled SMD (95% CI) was -1.01 (-1.61- 0.42). Twelve studies reported E-B in RA, pooled SMD was -1.06 (-1.59- -0.52), seven studies reported E-V with a pooled SMD -0.84 (-1.81- 0-14). There was significant heterogeneity for all three analyses. In patients with SpA six studies reported E-S. The pooled SMD (95% CI) was -0.79 (-1.07- 0.51). Six studies reported E-B, the pooled SMD was -0.99 (-1.28- -0. 69), two studies reported E-V -0.93 (-1.38- -0.48). The heterogeneity was not significant for these studies. 4.4 Correlation to inflammation and disease duration The correlations between markers of HRV, inflammation and disease duration are presented in supplementary table 6. 4.5 Sensitivity analyses RA vs controls RMSSD, E-S, E-B and E-V all displayed significant levels of heterogeneity. Results of the meta-analyses after removal of papers according to level of heterogeneity are presented in supplementary figures 3 and 4. Removing papers with a high risk of bias from 9
the meta-analyses of each HRV variable in the two disease categories did not significantly alter the results (supplementary figures 5-8). 4.6 Publication bias Publication bias was considered for the two subgroups, HF in RA vs controls, and E-B in RA vs controls (supplementary figures 9 and 10). The subgroups presented results of 10 and 12 more papers respectively. By visual examination there was no evidence of publication bias. The funnel plots were not distinctly asymmetric. There were a few outliers, but they were also roughly symmetric.
5. Discussion This study reports significantly lower markers of cardiac parasympathetic modulation, measured as both RMSSD and HF, in patients with RA compared to healthy controls. RMSSD and HF values for SpA patients were comparable between patients and controls. For the Ewing battery the heart rate variation was also significantly lower for patients vs. controls, except for the heart rate variation under the Valsalva test in patients with RA which although numerically lower in RA patients did not reach significance. To the best of our knowledge this is the first meta-analyses performed addressing HRV in patients with IJD. The results substantiate the findings of the systematic review performed by Adlan et al. which examined the evidence for AD, measured by a range of biomarkers, in patients with RA [6]. In that review, 15 studies reporting HRV were identified, 13 of these measured markers of cardiac parasympathetic modulation and 62% of these were in favour of lower markers of cardiac parasympathetic activity in patients with RA. Autonomic function can be altered by several physiological processes, many of which have relevance to patients with IJD. The parasympathetic cholinergic anti-inflammatory pathway has been described by many authors [13]. Studies in experimental models report that 10
acetylcholine, acting through nicotinic receptors, prevent NF-kappaB pathways and attenuate several cytokines including Tumour Necrosis Factor (TNF) [14]. There is evidence that the pathway has a bidirectional nature, whereby inflammation also can impact on AD, although the evidence for this reverse chain of events is more sparse [3, 15]. Of the 35 papers included in this review we found 18 which compared measures of inflammation and/or disease activity to HRV [10, 16-35] of which 13 [16, 18-23, 26-29, 31, 34] described an inverse crosssectional relationship between the parameters of inflammation and/or disease activity and HRV. Most of these studies examined ESR and/or CRP, but Kosek et al. reported an inverse relationship between serum IL-6 and HRV parameters from both time and frequency domains [18]. Three papers examined the relationship between surrogates of cumulative disease activity and HRV; Steinbrocker functional classification in RA [17], articular destruction in RA [30], and AS disease stage on spine and sacroiliac joint radiographs [36], none reporting significant associations. No prospective longitudinal studies examined inflammation as a determinant of AD. Eighteen papers examined the relationship between disease duration and HRV [10, 16, 17, 19, 20, 22, 25, 27-30, 32-35, 37, 38]. Of these only three found an inverse relationship between disease duration and HRV [20, 25, 34], the remainder reporting no significant association except for Anichov et al. who found a significant positive correlation between disease duration and HRV[16]. These results suggest that the inflammation present at the time of HRV measurement plays a greater role in determining HRV than the disease duration or cumulative disease activity. This is further illustrated by the findings of Koopman et al., who examined a population of patients at risk of RA (positive serology for rheumatoid factor and/or anti citrullinated protein antibodies and arthralgia), without increased levels of inflammation, and found that this group had a resting heart rate that was significantly higher than healthy controls, but overall no increased HRV when compared to healthy controls [39].
11
Pain is a common symptom in IJDs and may also impact on HRV. Fibromyalgia (FM) is a disease of chronic widespread pain with AD, and patients with RA and SpA have a greater prevalence of FM compared to the general population. Indeed studies have identified that between 10 and 20% of patients with IJD have FM, although the frequency varies according to the diagnostic method used [40]. A meta-analyses performed by Tracy in 2015 reported a significantly lower HRV in patients with FM, compared to healthy controls [4]. Possible scenarios are that the presence of AD makes the individual more likely to develop FM, viceversa, or a circular relationship. Nevertheless, the co-existence of FM with IJD is a possible confounding factor that could explain at least part of the unfavourable HRV seen in IJD. Holman et al measured HRV in a cohort of 33 patients with IJD (25 with RA, 8 with PsA) prior to anti-TNF alpha therapy. He found that patients with low markers of cardiac parasympathetic activity had a 40% chance of reaching ACR 20 response on anti-TNF alpha, compared to a 100% chance in the quartile of patients with high markers of cardiac parasympathetic activity [41]. While these findings may be taken as evidence of the interaction between inflammation and ANS, an alternative explanation of the findings of Holman et al. could be that low HRV is a marker of FM in patients with RA. Fibromyalgic RA has previously been shown to have a lower chance of achieving remission, possibly due to the chronicity of pain in these patients [42]. We did not find any study investigating the relationship between symptoms of fibromyalgia and HRV in patients with IJD. Physical inactivity is another possible confounder for the association between HRV and IJD. Resting heart rate is closely and inversely related to cardio-respiratory fitness, but also inversely related to HRV variables [43]. Patients with IJDs are less physically active compared to healthy controls and could therefore be expected to have a higher resting heart rate than control subjects [44 ]. The studies included in this review did not adjust the results for possible difference in resting heart rate between cases and controls. Continuous 12
monitoring of HRV and physical activity reveal a correlation between short term RRI fluctuations and physical activity, even though mean long-term HRV values are in general not significantly related to the mean measures of physical activity on a group level [45]. Borman et al. did not find a significant relationship between Nottingham Health Profile measure of physical activity and RRIV in 20 patients with AS [19]. Janse van Rensburg et al. [46] reported improved short-term HRV parameters following a 12-week exercise intervention in females with RA, in agreement with the conclusions of a review from 2010 showing an improved HRV in patients with CVD who started performing physical exercise [47]. 6.1 Methodological considerations In this meta-analysis, RMSSD was lower in patients with RA compared to controls, but not in patients with SpA. RMSSD is a commonly used time domain variable which is higher during rest at low HR, than during activities that increase HR. It is an established marker for cardiac parasympathetic activity [4]. It is however, interesting to note that in this meta-analysis of RMSSD, short [17, 20, 33, 39, 48, 49] and long-term recordings [11, 16, 18, 36, 50, 51] are grouped together. When studies utilising short-term recordings are excluded both patient groups display significantly lower RMSSD compared to population controls (ad hoc analysis). The HF component of HRV is generally accepted to predominantly reflect cardiac parasympathetic nerve activity which has a rapid dynamic control of the heart through the action of acetylcholine on muscarinic receptors. The fact that HF variability can be blocked by atropine in human volunteers confirm the validity of this parameter as a measure of parasympathetic activity [52]. In our meta-analysis HF was significantly lower in patients with RA compared to controls, but not in patients with SpA. Here the exclusion of short-term recordings [20, 22, 34, 35] did not change the results.
13
We did not find any other explanations for the heterogeneity of the results, but the studies included have used a variety of apparatus which may have introduced variance. The supplementary tables 4 and 5 present the study procedures and although many have referred to the HRV standard of measures [1] there are small variations in the methodology that may impact on the validity of the results. The Ewing battery is also a measure of parasympathetic activity [8] and has the advantage of standardisation of conditions, thus removing some possible sources of bias. The heterogeneity was mainly evident in patients with RA and the reason for this is uncertain. There are a large number of tests of autonomic function, and in this paper we have chosen to focus on measures that chiefly reflect cardiac parasympathetic function. Low frequency power (LF) is often included in studies of frequency domain HRV, but as the sympathetic nervous system is an important regulator of LF the measure was not included in this study. The papers included in the meta-analyses were of variable quality. Very few studies were for example assessor blinded. Sensitivity analyses were performed by excluding studies with a high risk of bias in any category from the analyses confirming the results. A variety of apparatus and procedures were employed but the majority of the studies used a methodology that was in accordance with the guidelines, although the limitations of fiveminute recordings must be borne in mind. Future studies should focus on the longitudinal impact of pain and longitudinal effect of physical activity and IJD treatment on HRV.
Conclusion Patients with IJD have lower cardiac parasympathetic modulation compared to healthy controls. There is a cross-sectional inverse relationship between level of inflammation and parasympathetic AD.
14
References [1] Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Eur Heart J. 1996;17:354-81. [2] Tsuji H, Larson MG, Venditti FJ, Jr., Manders ES, Evans JC, Feldman CL, et al. Impact of reduced heart rate variability on risk for cardiac events. The Framingham Heart Study. Circulation. 1996;94:2850-5. [3] Tracey KJ. The inflammatory reflex. Nature. 2002;420:853-9. [4] Tracy LM, Ioannou L, Baker KS, Gibson SJ, Georgiou-Karistianis N, Giummarra MJ. Metaanalytic evidence for decreased heart rate variability in chronic pain implicating parasympathetic nervous system dysregulation. Pain. 2016;157:7-29. [5] Agca R, Heslinga SC, van Halm VP, Nurmohamed MT. Atherosclerotic cardiovascular disease in patients with chronic inflammatory joint disorders. Heart. 2016;102:790-5. [6] Adlan AM, Lip GY, Paton JF, Kitas GD, Fisher JP. Autonomic function and rheumatoid arthritis: a systematic review. Semin Arthritis Rheum. 2014;44:283-304. [7] Billman GE. Heart rate variability - a historical perspective. Front Physiol. 2011;2:1-9. [8] Ewing DJ, Clarke BF. Diagnosis and management of diabetic autonomic neuropathy. Br Med J (Clin Res Ed). 1982;285:916-8. [9] National Institutes of Health NH, Lung and Blood Institute. Quality Assessment of Case-Control Studies http://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-riskreduction/tools/case-control. 2014. [10] Aydemir M, Yazisiz V, Basarici I, Avci AB, Erbasan F, Belgi A, et al. Cardiac autonomic profile in rheumatoid arthritis and systemic lupus erythematosus. Lupus. 2010;19:255-61. [11] Milovanovic B, Stojanovic L, Milicevik N, Vasic K, Bjelakovic B, Krotin M. Cardiac autonomic dysfunction in patients with systemic lupus, rheumatoid arthritis and sudden death risk. Srp Arh Celok Lek. 2010;138:26-32. [12] Bruchfeld A, Goldstein RS, Chavan S, Patel NB, Rosas-Ballina M, Kohn N, et al. Whole blood cytokine attenuation by cholinergic agonists ex vivo and relationship to vagus nerve activity in rheumatoid arthritis. J Intern Med. 2010;268:94-101. [13] Huston JM, Tracey KJ. The pulse of inflammation: heart rate variability, the cholinergic antiinflammatory pathway and implications for therapy. J Intern Med. 2011;269:45-53. [14] Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405:458-62. [15] Singh P, Hawkley LC, McDade TW, Cacioppo JT, Masi CM. Autonomic tone and C-reactive protein: a prospective population-based study. Clin Auton Res. 2009;19:367-74. [16] Anichkov DA, Shostak NA, Ivanov DS. Heart rate variability is related to disease activity and smoking in rheumatoid arthritis patients. Int J Clin Pract. 2007;61:777-83. [17] Evrengul H, Dursunoglu D, Cobankara V, Polat B, Seleci D, Kabukcu S, et al. Heart rate variability in patients with rheumatoid arthritis. Rheumatol Int. 2004;24:198-202. [18] Kosek E, Altawil R, Kadetoff D, Finn A, Westman M, Le Maitre E, et al. Evidence of different mediators of central inflammation in dysfunctional and inflammatory pain--interleukin-8 in fibromyalgia and interleukin-1 beta in rheumatoid arthritis. J Neuroimmunol. 2015;280:49-55. [19] Borman P, Gokoglu F, Kocaoglu S, Yorgancioglu ZR. The autonomic dysfunction in patients with ankylosing spondylitis: a clinical and electrophysiological study. Clin Rheumatol. 2008;27:126773. [20] Gaydukova I, Rebrov A, Nikitina N, Poddubnyy D. Decreased heart rate variability in patients with psoriatic arthritis. Clin Rheumatol. 2012;31:1377-81. [21] Gunay T, Yilmaz O, Bodur H. Autonomic functions and their relations with disease activity in ankylosing spondylitis. Acta Reumatol Port. 2012;37:234-9. [22] Lazzerini PE, Acampa M, Capecchi PL, Hammoud M, Maffei S, Bisogno S, et al. Association between high sensitivity C-reactive protein, heart rate variability and corrected QT interval in patients with chronic inflammatory arthritis. Eur J Intern Med. 2013;24:368-74.
15
[23] Leden I, Eriksson A, Lilja B, Sturfelt G, Sundkvist G. Autonomic nerve function in rheumatoid arthritis of varying severity. Scand J Rheumatol. 1983;12:166-70. [24] Louthrenoo W, Ruttanaumpawan P, Aramrattana A, Sukitawut W. Cardiovascular autonomic nervous system dysfunction in patients with rheumatoid arthritis and systemic lupus erythematosus. QJM. 1999;92:97-102. [25] Sandhu V, Allen SC. The effects of age, seropositivity and disease duration on autonomic cardiovascular reflexes in patients with rheumatoid arthritis. Int J Clin Pract. 2004;58:740-5. [26] Syngle V, Syngle A, Garg N, Krishan P, Verma I. Predictors of autonomic neuropathy in rheumatoid arthritis. Auton Neurosci. 2016. [27] Syngle A, Verma I, Krishan P, Syngle V. Disease-modifying antirheumatic drugs improve cardiovascular autonomic neuropathy in psoriatic arthritis. Ther Adv Musculoskelet Dis. 2016;8:42-8. [28] Syngle A, Verma I, Krishan P, Garg N, Syngle V. Disease-modifying anti-rheumatic drugs improve autonomic neuropathy in arthritis: DIANA study. Clin Rheumatol. 2015;34:1233-41. [29] Syngle A, Verma I, Garg N, Krishan P. Autonomic dysfunction in psoriatic arthritis. Clin Rheumatol. 2013;32:1059-64. [30] Toussirot E, Serratrice G, Valentin P. Autonomic nervous system involvement in rheumatoid arthritis. 50 cases. J Rheumatol. 1993;20:1508-14. [31] Toussirot E, Bahjaoui-Bouhaddi M, Poncet JC, Cappelle S, Henriet MT, Wendling D, et al. Abnormal autonomic cardiovascular control in ankylosing spondylitis. Ann Rheum Dis. 1999;58:4817. [32] Verma IS, A.; Garg, N.; Krishan, P. Association between cardiovascular autonomic neuropathy and peripheral sympathetic neuropathy in rheumatoid arthritis and ankylosing spondylitis. The Pharma Research. 2014;11:1-7. [33] Yadav RK, Gupta R, Deepak KK. A pilot study on short term heart rate variability & its correlation with disease activity in Indian patients with rheumatoid arthritis. Indian J Med Res. 2012;136:593-8. [34] Wei CY, Kung WM, Chou YS, Wang YC, Tai HC, Wei JC. Cardiac Autonomic Function in Patients With Ankylosing Spondylitis: A Case-Control Study. Medicine (Baltimore). 2016;95:e3749. [35] Yildirir A, Aksoyek S, Calguneri M, Aytemir K, Apras S, Kiraz S, et al. No evidence of cardiac autonomic involvement in ankylosing spondylitis, as assessed by heart rate variability. Clin Rheumatol. 2001;20:185-8. [36] Kazmierczak J, Peregud-Pogorzelska M, Biernawska J, Przepiera-Bedzak H, Goracy J, Brzosko I, et al. Cardiac arrhythmias and conduction disturbances in patients with ankylosing spondylitis. Angiology. 2007;58:751-6. [37] Maule S, Quadri R, Mirante D, Pellerito RA, Marucco E, Marinone C, et al. Autonomic nervous dysfunction in systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA): possible pathogenic role of autoantibodies to autonomic nervous structures. Clin Exp Immunol. 1997;110:4237. [38] Saraswathi PV, Neelambikai N, Mahesh A, Govindarajan K. Cardiovascular parasympathetic nervous system dysfunction in female rheumatoid arthritis patients. Indian J Physiol Pharmacol. 2013;57:23-30. [39] Koopman FA, Tang MW, Vermeij J, de Hair MJ, Choi IY, Vervoordeldonk MJ, et al. Autonomic Dysfunction Precedes Development of Rheumatoid Arthritis: A Prospective Cohort Study. EBioMedicine. 2016;6:231-7. [40] Bello N, Etcheto A, Beal C, Dougados M, Molto A. Evaluation of the impact of fibromyalgia in disease activity and treatment effect in spondyloarthritis. Arthritis Res Ther. 2016;18:42. [41] Holman AJ, Ng E. Heart rate variability predicts anti-tumor necrosis factor therapy response for inflammatory arthritis. Auton Neurosci. 2008;143:58-67. [42] Duran J, Combe B, Niu J, Rincheval N, Gaujoux-Viala C, Felson DT. The effect on treatment response of fibromyalgic symptoms in early rheumatoid arthritis patients: results from the ESPOIR cohort. Rheumatology (Oxford). 2015;54:2166-70. [43] Monfredi O, Lyashkov AE, Johnsen AB, Inada S, Schneider H, Wang R, et al. Biophysical characterization of the underappreciated and important relationship between heart rate variability and heart rate. Hypertension. 2014;64:1334-43.
16
[44] Halvorsen S, Vollestad NK, Fongen C, Provan SA, Semb AG, Hagen KB, et al. Physical fitness in patients with ankylosing spondylitis: comparison with population controls. Phys Ther. 2012;92:298309. [45] Hautala AJ, Karjalainen J, Kiviniemi AM, Kinnunen H, Makikallio TH, Huikuri HV, et al. Physical activity and heart rate variability measured simultaneously during waking hours. Am J Physiol Heart Circ Physiol. 2010;298:H874-80. [46] Janse van Rensburg DC, Ker JA, Grant CC, Fletcher L. Effect of exercise on cardiac autonomic function in females with rheumatoid arthritis. Clin Rheumatol. 2012;31:1155-62. [47] Routledge FS, Campbell TS, McFetridge-Durdle JA, Bacon SL. Improvements in heart rate variability with exercise therapy. Can J Cardiol. 2010;26:303-12. [48] Jahan KB, N; Ferdousi, S. Heart rate variability in patients with rheumatoid arthritis. Journal of Bangladesh Society of Physiology. 2012. [49] Janse van Rensburg DC, Ker JA, Grant CC, Fletcher L. Autonomic impairment in rheumatoid arthritis. Int J Rheum Dis. 2012;15:419-26. [50] Gunes Y, Tuncer M, Guntekin U, Sahin M, Yazmalar L. Effects of ankylosing spondylitis on the heart. Acta Cardiol. 2009;64:385-92. [51] Kaya EB, Okutucu S, Aksoy H, Karakulak UN, Tulumen E, Ozdemir O, et al. Evaluation of cardiac autonomic functions in patients with ankylosing spondylitis via heart rate recovery and heart rate variability. Clin Res Cardiol. 2010;99:803-8. [52] Pomeranz B, Macaulay RJ, Caudill MA, Kutz I, Adam D, Gordon D, et al. Assessment of autonomic function in humans by heart rate spectral analysis. Am J Physiol. 1985;248:H151-3.
17
Figure legends. Figure 1 Time and frequency domain HRV for patients with RA vs. controls Figure 2 Time and frequency domain HRV for patients with SpA vs. controls Figure 3 Ewing protocol for patients with RA vs. controls Figure 4 Ewing protocol for patients with SpA vs. controls
18
19
20
image
Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A metaanalyses study Sella A. Provan, Daniela Schäfer Olstad, Erik E. Solberg, Geir Smedslund, Hanne Dagfinrud www.elsevier.com/locate/bios
PII: DOI: Reference:
S0049-0172(17)30517-6 https://doi.org/10.1016/j.semarthrit.2017.11.010 YSARH51278
To appear in: Seminars in Arthritis and Rheumatism Cite this article as: Sella A. Provan, Daniela Schäfer Olstad, Erik E. Solberg, Geir Smedslund and Hanne Dagfinrud, Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A meta-analyses study, Seminars in Arthritis and Rheumatism,doi:10.1016/j.semarthrit.2017.11.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Evidence of reduced parasympathetic autonomic regulation in inflammatory joint disease. A meta-analyses study.
Sella A Provan1, Daniela Schäfer Olstad2, Erik E Solberg3, Geir Smedslund2, Hanne Dagfinrud2 1
Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway, 2 National Resource
Centre for Rehabilitation in Rheumatology, Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway 3 Department of Medicine, Diakonhjemmet Hospital, Oslo, Norway.
Corresponding author: Sella Aarrestad Provan MD. PhD. Department of Rheumatology Diakonhjemmet Hospital PB 23 Vindern 0319 Oslo, Norway [email protected]
This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation Disclosures: The authors SAP, EES, GS, HD declare no conflict of interest. DSO is currently employed by Polar Electro, Finland. Funding statement: This work was sponsored by the Research Fund at Diakonhjemmet Hospital and by Extra-Stiftelsen, Norway. The sponsors have not been involved in the data-analyses, writing or submission of this paper. .
1
Abstract Background: Rheumatoid arthritis (RA) and spondyloarthritis (SpA) are inflammatory joint disorders (IJD) with increased risk of cardiovascular disease (CVD). Autonomic dysfunction (AD) is a risk factor for CVD, and parasympathetic AD is linked to key features of IJD such as inflammation, physical inactivity and pain. Heart rate variability (HRV) is a marker of cardiac AD. The study objective was to compare parasympathetic cardiac AD, measured by HRV, between patients with IJD and healthy controls, using metaanalysis methodology, and to examine the impact of inflammation, physical inactivity and pain on HRV in IJD.
Methods: Medline, Embase and Amed were searched. Inclusion criteria were adult casecontrol studies published in English or a Scandinavian language, presenting HRV data in IJD. Two measures of HRV and 3 from the Ewing protocol were selected: Square root of mean squared difference of successive R-R intervals (RMSSD), high frequency (HF), Ewing protocol; standing (E-S), breathing (E-B) and Valsalva (E-V). Patients with RA, SpA and healthy controls were compared separately using random-effects meta-analyses of standardized mean differences (SMD). Results: 35 papers were eligible for inclusion. For RMSSD the pooled SMD (95% CI) RA vs. controls was -0.90 (-1.35 to -0.44), for SpA vs. controls; -0.34 (-0.73 to 0.06). For HF pooled SMD RA vs. controls was -0.78 (-0.99 to -0.57), for SpA vs. controls; -0.04 (-0.22 to 0.13). All Ewing parameters were significantly lower in cases, except for E-V which was comparable between cases and controls in patients with RA.
2
Conclusion: Patients with IJD have cardiac parasympathetic AD which is related to inflammation. Key words: Autonomic dysfunction, heart rate variability, Inflammatory joint disease, rheumatoid arthritis, spondyloarthritis.
1. Background Cardiac autonomic dysfunction (AD) is linked to arrhythmias, all-cause mortality and death following myocardial infarction [1, 2]. There is a complex interplay between the immune and parasympathetic autonomic systems, with evidence of mutual regulation [3]. Parasympathetic AD is also linked to physical inactivity and is more prevalent in patients who suffer from chronic pain [4]. Rheumatoid arthritis (RA) and spondyloarthritis (SpA) are inflammatory joint disorders (IJD) with an excess risk of cardiovascular disease (CVD), arrhythmias and sudden death [5]. IJDs are characterised by joint inflammation and symptoms include pain, functional decline and restricted movement. A previous systematic review concluded that there is evidence of AD in approximately 60 % of patients with RA, but a meta-analysis of the data was not performed due to the heterogeneity of outcomes included in the review. [6]. The heart rate of the individual will vary on a beat to beat level throughout the day, influenced by the body’s physiological and psychological state and essentially under the control of the autonomic nervous system. The variation in heart rate has been the subject of academic interest for decades, and several techniques for quantifying the variation have been developed. Heart rate variability (HRV) is a measure of heart-rate variation and an estimate of cardiac autonomic dysfunction (AD) [1, 2] The objective of this study was to compare parasympathetic cardiac AD, measured by HRV, between patients with IJD and healthy controls, using meta-analyses methodology, and to examine the impact of inflammation, physical inactivity and pain on HRV in IJD. 3
2. Methods These meta-analyses were performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA statement).
2.1 Search Strategy A research librarian developed and conducted systematic searches in the following databases: Medline (from 1946 to August 2016), Embase (from 1980 to August 2016) and Amed (from 1985 to August 2016). 2.2 Inclusion criteria: Case-control studies on human subjects of any age, published in English, Norwegian, Danish or Swedish which present data on HRV according to an established methodology. Population: Patients who were diagnosed with an inflammatory rheumatic joint disease (RA or SpA). Included under the SpA diagnosis in this study are also patients with psoriatic arthritis (PsA) or ankylosing spondylitis (AS). 2.3 Exclusion criteria: Studies not fulfilling the inclusion criteria, double publications and studies presenting outcomes graphically or as categorical variables only were excluded. 2.4 HRV as a marker of autonomic dysfunction
HRV is the result of complex interactions between sympathetic and parasympathetic stimulation on the sinus node of the heart, coupled with input from pacemaker cells and the geometry of the heart in the individual [1]. HRV can be defined as the beat-to-beat variation in heart-rate (R-R interval (RRI)) quantified in time and frequency domains. Frequently 4
reported variables in the time domain are the square root of the variance of the RRI (SDNN) and Square root of the mean squared difference of successive RRI (RMSSD). In the frequency domain, also named power spectral analysis, the RRI (y-axis) is plotted against time of the corresponding heart-beat (x-axis). The spectral power for a given frequency is quantified by determining the area under the curve within the specific frequency range [7]. The spectral power can be estimated by the non-parametric fast Fourier transformation or by parametric autoregressive modelling. In the frequency domain, low frequency (LF) or high frequency (HF) components are frequently reported variables. A five minute recording is sufficient for analyses in both domains, although additional information can be gathered from 24-hour recordings[1]. The Ewing battery is a collection of five non-invasive tests originally designed to assess autonomic neuropathy in patients with diabetes [8]. Three tests quantify cardiac parasympathetic integrity by determining changes in RRI in response to the following stressors; standing up (E-S), deep breathing (E-B) and Valsalva manoeuvre (E-V), blood pressure response to standing up, and sustained handgrip. A summary of abbreviations of different HRV parameters described in this report is outlined in supplementary table 1. In order to limit the scope of the study we selected central markers of markers of parasympathetic cardiac autonomic function. Two HRV variables; RMSSD from the time domain and HF component from the frequency domain, and three tests from the Ewing battery; E-S, E-B and E-V, were selected
2.4 Screening Two of the authors independently judged each record from the search as either not relevant or possibly relevant. If at least one person scored a record as being possibly relevant, we ordered it in full text. The full texts were read independently by two of the authors (SAP and GS) and 5
each person rated them as either ‘include’ or ‘exclude’. A final decision was made by the third author (HD) if there was a conflict. 2.5 Risk of bias Two of the authors independently assessed risk of bias for each included study. Conflicts were solved through an independent assessment by a third author and through group discussions. For the risk of bias assessments we used a modified version of “The quality assessment for case-control studies checklist” from the National Heart, Lung, and Blood Institute [9] (Supplementary table 2). 2.6 Data extraction The following data were extracted from each included study: Publication year and country in which the study was conducted, number of subjects, matching procedure, diagnosis of patients, percent females included, average age, HRV outcomes, apparatus used, measures of disease activity when included in the analysis, other measures of autonomic function measured. When data were presented only as categorical variables or in figures, the authors were contacted whenever possible and continuous data were requested.
3. Theory and calculations For the comparison of the disease groups with the healthy controls, we conducted randomeffects meta-analyses using standardized mean differences. The analyses were run separately for patients with RA and SpA. Within each diagnostic category we ran separate meta-analyses for each of the variables RMSSD, HF, E-S, E-B, and E-V. Sensitivity analyses were performed: Forest plots were examined for heterogeneity and when significant heterogeneity was displayed outliers were removed in descending order from the pooled results until the level of heterogeneity became insignificant (p ≥ 0.05) and the meta-analyses repeated.
6
Additionally papers with a high risk of bias were removed and the meta-analyses repeated separately. Funnel plots were constructed to assess the risk of publication bias in the subgroups in which 10 or more studies had been identified. Suspected publication bias would appear as asymmetrical funnel plots due to the non-publication of small studies that found nonsignificant difference in HRV between patient and controls groups.
4. Results The final literature search was performed in August 2016. Supplementary table 3 presents the complete search strategy. Supplementary figure 1 presents a summary of the findings of literature search, screening and selection process. The database search identified 1015 records. In addition 120 records were found through searching grey literature (conference abstracts, internet searches) and bibliographies. After duplicates had been removed, 847 records remained, and these titles and abstracts were reviewed according to our study selection criteria. We considered 112 full texts in more detail. Two papers presented results from the Ewing battery as categorical outcomes and these outcomes were not included in the meta-analyses [10, 11]. One paper did not report standard confidence intervals of data and was excluded [12]. 35 papers were finally selected. Of the 35 papers included, 21 studied RA patients, 11 studied SpA and 3 papers studied both patient groups. Supplementary tables 4 and 5 present the characteristics of the studies concerning patients with RA and SpA respectively. Of the studies including patients with SpA, ten included patients diagnosed with AS, three included PsA patients and one included patients diagnosed with SpA specifically. The studies came from 18 countries of which Turkey and India had the greatest number of publications with nine and seven papers respectively. The average number of cases included was low, with a mean of 38 (range 10 to 7
207). The RA studies included more women than men, reflecting the epidemiology of the disease. Nine RA studies only included female subjects. Gender distribution was more similar in the SpA studies. The average age of cases in the RA studies was 48 years (range 31-63 years), while the SpA cases were younger [average 38 years (range 29-51 years)]. The Ewing battery was the most frequent HRV outcome, measured in 17 studies, short-term HRV was reported in 12, while 24-hour HRV was the outcome in 7 studies. (One study measured both short-term HRV and Ewing Battery). 4.1 Risk of bias Risk of bias decisions are presented in supplementary figure 2. All papers clearly differentiated between patients and controls, and the majority of papers had recruited patients and controls from a similar population and applied consistent criteria for the selection of participants. In three papers there was a high risk of bias regarding item 3, and in these cases there was a discrepancy in the exclusion criteria between the patient and controls populations. Blinding of the assessors was not described in the majority of studies and, thus, only four studies had a low risk of bias regarding item 4. We defined matching or adjustment for confounders as adequate if age and gender were statistically similar between groups, or adjusted for in statistical analyses. Ten studies had a high risk of bias due to insufficient matching or adjustments, the highest number in any item. For the last item we looked at how the examinations were standardised, and we considered whether the examinations were performed at the same time of day and under fasting conditions. Here three papers were judged to have a high risk of bias due to poor standardisation.
8
4.2 Studies comparing the HRV time and frequency domain variables between patients and controls The results are presented in Figure 1 (RA) and Figure 2 (SpA). In patients with RA eight studies reported RMSSD and the pooled results gave a SMD (95% CI) of -0.90 (-1.35- -0.44). For HF there were ten studies, with a pooled SMD of -0.78 (-0.99-0.57). The level of heterogeneity was significant for the RMSSD analyses. In patients with SpA, four studies reported RMSSD and the pooled SMD (95% CI) was -0.34 (-0.73- 0.06). Six studies reported HF with a pooled SMD of -0.04 (-0.22- 0.13). There was a non-significant level of heterogeneity for all three outcomes. 4.3 Studies comparing results of the Ewing battery between patients and controls The results are presented in Figure 3 (RA) and Figure 4 (SpA). In patients with RA nine studies reported E-S. The pooled SMD (95% CI) was -1.01 (-1.61- 0.42). Twelve studies reported E-B in RA, pooled SMD was -1.06 (-1.59- -0.52), seven studies reported E-V with a pooled SMD -0.84 (-1.81- 0-14). There was significant heterogeneity for all three analyses. In patients with SpA six studies reported E-S. The pooled SMD (95% CI) was -0.79 (-1.07- 0.51). Six studies reported E-B, the pooled SMD was -0.99 (-1.28- -0. 69), two studies reported E-V -0.93 (-1.38- -0.48). The heterogeneity was not significant for these studies. 4.4 Correlation to inflammation and disease duration The correlations between markers of HRV, inflammation and disease duration are presented in supplementary table 6. 4.5 Sensitivity analyses RA vs controls RMSSD, E-S, E-B and E-V all displayed significant levels of heterogeneity. Results of the meta-analyses after removal of papers according to level of heterogeneity are presented in supplementary figures 3 and 4. Removing papers with a high risk of bias from 9
the meta-analyses of each HRV variable in the two disease categories did not significantly alter the results (supplementary figures 5-8). 4.6 Publication bias Publication bias was considered for the two subgroups, HF in RA vs controls, and E-B in RA vs controls (supplementary figures 9 and 10). The subgroups presented results of 10 and 12 more papers respectively. By visual examination there was no evidence of publication bias. The funnel plots were not distinctly asymmetric. There were a few outliers, but they were also roughly symmetric.
5. Discussion This study reports significantly lower markers of cardiac parasympathetic modulation, measured as both RMSSD and HF, in patients with RA compared to healthy controls. RMSSD and HF values for SpA patients were comparable between patients and controls. For the Ewing battery the heart rate variation was also significantly lower for patients vs. controls, except for the heart rate variation under the Valsalva test in patients with RA which although numerically lower in RA patients did not reach significance. To the best of our knowledge this is the first meta-analyses performed addressing HRV in patients with IJD. The results substantiate the findings of the systematic review performed by Adlan et al. which examined the evidence for AD, measured by a range of biomarkers, in patients with RA [6]. In that review, 15 studies reporting HRV were identified, 13 of these measured markers of cardiac parasympathetic modulation and 62% of these were in favour of lower markers of cardiac parasympathetic activity in patients with RA. Autonomic function can be altered by several physiological processes, many of which have relevance to patients with IJD. The parasympathetic cholinergic anti-inflammatory pathway has been described by many authors [13]. Studies in experimental models report that 10
acetylcholine, acting through nicotinic receptors, prevent NF-kappaB pathways and attenuate several cytokines including Tumour Necrosis Factor (TNF) [14]. There is evidence that the pathway has a bidirectional nature, whereby inflammation also can impact on AD, although the evidence for this reverse chain of events is more sparse [3, 15]. Of the 35 papers included in this review we found 18 which compared measures of inflammation and/or disease activity to HRV [10, 16-35] of which 13 [16, 18-23, 26-29, 31, 34] described an inverse crosssectional relationship between the parameters of inflammation and/or disease activity and HRV. Most of these studies examined ESR and/or CRP, but Kosek et al. reported an inverse relationship between serum IL-6 and HRV parameters from both time and frequency domains [18]. Three papers examined the relationship between surrogates of cumulative disease activity and HRV; Steinbrocker functional classification in RA [17], articular destruction in RA [30], and AS disease stage on spine and sacroiliac joint radiographs [36], none reporting significant associations. No prospective longitudinal studies examined inflammation as a determinant of AD. Eighteen papers examined the relationship between disease duration and HRV [10, 16, 17, 19, 20, 22, 25, 27-30, 32-35, 37, 38]. Of these only three found an inverse relationship between disease duration and HRV [20, 25, 34], the remainder reporting no significant association except for Anichov et al. who found a significant positive correlation between disease duration and HRV[16]. These results suggest that the inflammation present at the time of HRV measurement plays a greater role in determining HRV than the disease duration or cumulative disease activity. This is further illustrated by the findings of Koopman et al., who examined a population of patients at risk of RA (positive serology for rheumatoid factor and/or anti citrullinated protein antibodies and arthralgia), without increased levels of inflammation, and found that this group had a resting heart rate that was significantly higher than healthy controls, but overall no increased HRV when compared to healthy controls [39].
11
Pain is a common symptom in IJDs and may also impact on HRV. Fibromyalgia (FM) is a disease of chronic widespread pain with AD, and patients with RA and SpA have a greater prevalence of FM compared to the general population. Indeed studies have identified that between 10 and 20% of patients with IJD have FM, although the frequency varies according to the diagnostic method used [40]. A meta-analyses performed by Tracy in 2015 reported a significantly lower HRV in patients with FM, compared to healthy controls [4]. Possible scenarios are that the presence of AD makes the individual more likely to develop FM, viceversa, or a circular relationship. Nevertheless, the co-existence of FM with IJD is a possible confounding factor that could explain at least part of the unfavourable HRV seen in IJD. Holman et al measured HRV in a cohort of 33 patients with IJD (25 with RA, 8 with PsA) prior to anti-TNF alpha therapy. He found that patients with low markers of cardiac parasympathetic activity had a 40% chance of reaching ACR 20 response on anti-TNF alpha, compared to a 100% chance in the quartile of patients with high markers of cardiac parasympathetic activity [41]. While these findings may be taken as evidence of the interaction between inflammation and ANS, an alternative explanation of the findings of Holman et al. could be that low HRV is a marker of FM in patients with RA. Fibromyalgic RA has previously been shown to have a lower chance of achieving remission, possibly due to the chronicity of pain in these patients [42]. We did not find any study investigating the relationship between symptoms of fibromyalgia and HRV in patients with IJD. Physical inactivity is another possible confounder for the association between HRV and IJD. Resting heart rate is closely and inversely related to cardio-respiratory fitness, but also inversely related to HRV variables [43]. Patients with IJDs are less physically active compared to healthy controls and could therefore be expected to have a higher resting heart rate than control subjects [44 ]. The studies included in this review did not adjust the results for possible difference in resting heart rate between cases and controls. Continuous 12
monitoring of HRV and physical activity reveal a correlation between short term RRI fluctuations and physical activity, even though mean long-term HRV values are in general not significantly related to the mean measures of physical activity on a group level [45]. Borman et al. did not find a significant relationship between Nottingham Health Profile measure of physical activity and RRIV in 20 patients with AS [19]. Janse van Rensburg et al. [46] reported improved short-term HRV parameters following a 12-week exercise intervention in females with RA, in agreement with the conclusions of a review from 2010 showing an improved HRV in patients with CVD who started performing physical exercise [47]. 6.1 Methodological considerations In this meta-analysis, RMSSD was lower in patients with RA compared to controls, but not in patients with SpA. RMSSD is a commonly used time domain variable which is higher during rest at low HR, than during activities that increase HR. It is an established marker for cardiac parasympathetic activity [4]. It is however, interesting to note that in this meta-analysis of RMSSD, short [17, 20, 33, 39, 48, 49] and long-term recordings [11, 16, 18, 36, 50, 51] are grouped together. When studies utilising short-term recordings are excluded both patient groups display significantly lower RMSSD compared to population controls (ad hoc analysis). The HF component of HRV is generally accepted to predominantly reflect cardiac parasympathetic nerve activity which has a rapid dynamic control of the heart through the action of acetylcholine on muscarinic receptors. The fact that HF variability can be blocked by atropine in human volunteers confirm the validity of this parameter as a measure of parasympathetic activity [52]. In our meta-analysis HF was significantly lower in patients with RA compared to controls, but not in patients with SpA. Here the exclusion of short-term recordings [20, 22, 34, 35] did not change the results.
13
We did not find any other explanations for the heterogeneity of the results, but the studies included have used a variety of apparatus which may have introduced variance. The supplementary tables 4 and 5 present the study procedures and although many have referred to the HRV standard of measures [1] there are small variations in the methodology that may impact on the validity of the results. The Ewing battery is also a measure of parasympathetic activity [8] and has the advantage of standardisation of conditions, thus removing some possible sources of bias. The heterogeneity was mainly evident in patients with RA and the reason for this is uncertain. There are a large number of tests of autonomic function, and in this paper we have chosen to focus on measures that chiefly reflect cardiac parasympathetic function. Low frequency power (LF) is often included in studies of frequency domain HRV, but as the sympathetic nervous system is an important regulator of LF the measure was not included in this study. The papers included in the meta-analyses were of variable quality. Very few studies were for example assessor blinded. Sensitivity analyses were performed by excluding studies with a high risk of bias in any category from the analyses confirming the results. A variety of apparatus and procedures were employed but the majority of the studies used a methodology that was in accordance with the guidelines, although the limitations of fiveminute recordings must be borne in mind. Future studies should focus on the longitudinal impact of pain and longitudinal effect of physical activity and IJD treatment on HRV.
Conclusion Patients with IJD have lower cardiac parasympathetic modulation compared to healthy controls. There is a cross-sectional inverse relationship between level of inflammation and parasympathetic AD.
14
References [1] Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Eur Heart J. 1996;17:354-81. [2] Tsuji H, Larson MG, Venditti FJ, Jr., Manders ES, Evans JC, Feldman CL, et al. Impact of reduced heart rate variability on risk for cardiac events. The Framingham Heart Study. Circulation. 1996;94:2850-5. [3] Tracey KJ. The inflammatory reflex. Nature. 2002;420:853-9. [4] Tracy LM, Ioannou L, Baker KS, Gibson SJ, Georgiou-Karistianis N, Giummarra MJ. Metaanalytic evidence for decreased heart rate variability in chronic pain implicating parasympathetic nervous system dysregulation. Pain. 2016;157:7-29. [5] Agca R, Heslinga SC, van Halm VP, Nurmohamed MT. Atherosclerotic cardiovascular disease in patients with chronic inflammatory joint disorders. Heart. 2016;102:790-5. [6] Adlan AM, Lip GY, Paton JF, Kitas GD, Fisher JP. Autonomic function and rheumatoid arthritis: a systematic review. Semin Arthritis Rheum. 2014;44:283-304. [7] Billman GE. Heart rate variability - a historical perspective. Front Physiol. 2011;2:1-9. [8] Ewing DJ, Clarke BF. Diagnosis and management of diabetic autonomic neuropathy. Br Med J (Clin Res Ed). 1982;285:916-8. [9] National Institutes of Health NH, Lung and Blood Institute. Quality Assessment of Case-Control Studies http://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-riskreduction/tools/case-control. 2014. [10] Aydemir M, Yazisiz V, Basarici I, Avci AB, Erbasan F, Belgi A, et al. Cardiac autonomic profile in rheumatoid arthritis and systemic lupus erythematosus. Lupus. 2010;19:255-61. [11] Milovanovic B, Stojanovic L, Milicevik N, Vasic K, Bjelakovic B, Krotin M. Cardiac autonomic dysfunction in patients with systemic lupus, rheumatoid arthritis and sudden death risk. Srp Arh Celok Lek. 2010;138:26-32. [12] Bruchfeld A, Goldstein RS, Chavan S, Patel NB, Rosas-Ballina M, Kohn N, et al. Whole blood cytokine attenuation by cholinergic agonists ex vivo and relationship to vagus nerve activity in rheumatoid arthritis. J Intern Med. 2010;268:94-101. [13] Huston JM, Tracey KJ. The pulse of inflammation: heart rate variability, the cholinergic antiinflammatory pathway and implications for therapy. J Intern Med. 2011;269:45-53. [14] Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405:458-62. [15] Singh P, Hawkley LC, McDade TW, Cacioppo JT, Masi CM. Autonomic tone and C-reactive protein: a prospective population-based study. Clin Auton Res. 2009;19:367-74. [16] Anichkov DA, Shostak NA, Ivanov DS. Heart rate variability is related to disease activity and smoking in rheumatoid arthritis patients. Int J Clin Pract. 2007;61:777-83. [17] Evrengul H, Dursunoglu D, Cobankara V, Polat B, Seleci D, Kabukcu S, et al. Heart rate variability in patients with rheumatoid arthritis. Rheumatol Int. 2004;24:198-202. [18] Kosek E, Altawil R, Kadetoff D, Finn A, Westman M, Le Maitre E, et al. Evidence of different mediators of central inflammation in dysfunctional and inflammatory pain--interleukin-8 in fibromyalgia and interleukin-1 beta in rheumatoid arthritis. J Neuroimmunol. 2015;280:49-55. [19] Borman P, Gokoglu F, Kocaoglu S, Yorgancioglu ZR. The autonomic dysfunction in patients with ankylosing spondylitis: a clinical and electrophysiological study. Clin Rheumatol. 2008;27:126773. [20] Gaydukova I, Rebrov A, Nikitina N, Poddubnyy D. Decreased heart rate variability in patients with psoriatic arthritis. Clin Rheumatol. 2012;31:1377-81. [21] Gunay T, Yilmaz O, Bodur H. Autonomic functions and their relations with disease activity in ankylosing spondylitis. Acta Reumatol Port. 2012;37:234-9. [22] Lazzerini PE, Acampa M, Capecchi PL, Hammoud M, Maffei S, Bisogno S, et al. Association between high sensitivity C-reactive protein, heart rate variability and corrected QT interval in patients with chronic inflammatory arthritis. Eur J Intern Med. 2013;24:368-74.
15
[23] Leden I, Eriksson A, Lilja B, Sturfelt G, Sundkvist G. Autonomic nerve function in rheumatoid arthritis of varying severity. Scand J Rheumatol. 1983;12:166-70. [24] Louthrenoo W, Ruttanaumpawan P, Aramrattana A, Sukitawut W. Cardiovascular autonomic nervous system dysfunction in patients with rheumatoid arthritis and systemic lupus erythematosus. QJM. 1999;92:97-102. [25] Sandhu V, Allen SC. The effects of age, seropositivity and disease duration on autonomic cardiovascular reflexes in patients with rheumatoid arthritis. Int J Clin Pract. 2004;58:740-5. [26] Syngle V, Syngle A, Garg N, Krishan P, Verma I. Predictors of autonomic neuropathy in rheumatoid arthritis. Auton Neurosci. 2016. [27] Syngle A, Verma I, Krishan P, Syngle V. Disease-modifying antirheumatic drugs improve cardiovascular autonomic neuropathy in psoriatic arthritis. Ther Adv Musculoskelet Dis. 2016;8:42-8. [28] Syngle A, Verma I, Krishan P, Garg N, Syngle V. Disease-modifying anti-rheumatic drugs improve autonomic neuropathy in arthritis: DIANA study. Clin Rheumatol. 2015;34:1233-41. [29] Syngle A, Verma I, Garg N, Krishan P. Autonomic dysfunction in psoriatic arthritis. Clin Rheumatol. 2013;32:1059-64. [30] Toussirot E, Serratrice G, Valentin P. Autonomic nervous system involvement in rheumatoid arthritis. 50 cases. J Rheumatol. 1993;20:1508-14. [31] Toussirot E, Bahjaoui-Bouhaddi M, Poncet JC, Cappelle S, Henriet MT, Wendling D, et al. Abnormal autonomic cardiovascular control in ankylosing spondylitis. Ann Rheum Dis. 1999;58:4817. [32] Verma IS, A.; Garg, N.; Krishan, P. Association between cardiovascular autonomic neuropathy and peripheral sympathetic neuropathy in rheumatoid arthritis and ankylosing spondylitis. The Pharma Research. 2014;11:1-7. [33] Yadav RK, Gupta R, Deepak KK. A pilot study on short term heart rate variability & its correlation with disease activity in Indian patients with rheumatoid arthritis. Indian J Med Res. 2012;136:593-8. [34] Wei CY, Kung WM, Chou YS, Wang YC, Tai HC, Wei JC. Cardiac Autonomic Function in Patients With Ankylosing Spondylitis: A Case-Control Study. Medicine (Baltimore). 2016;95:e3749. [35] Yildirir A, Aksoyek S, Calguneri M, Aytemir K, Apras S, Kiraz S, et al. No evidence of cardiac autonomic involvement in ankylosing spondylitis, as assessed by heart rate variability. Clin Rheumatol. 2001;20:185-8. [36] Kazmierczak J, Peregud-Pogorzelska M, Biernawska J, Przepiera-Bedzak H, Goracy J, Brzosko I, et al. Cardiac arrhythmias and conduction disturbances in patients with ankylosing spondylitis. Angiology. 2007;58:751-6. [37] Maule S, Quadri R, Mirante D, Pellerito RA, Marucco E, Marinone C, et al. Autonomic nervous dysfunction in systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA): possible pathogenic role of autoantibodies to autonomic nervous structures. Clin Exp Immunol. 1997;110:4237. [38] Saraswathi PV, Neelambikai N, Mahesh A, Govindarajan K. Cardiovascular parasympathetic nervous system dysfunction in female rheumatoid arthritis patients. Indian J Physiol Pharmacol. 2013;57:23-30. [39] Koopman FA, Tang MW, Vermeij J, de Hair MJ, Choi IY, Vervoordeldonk MJ, et al. Autonomic Dysfunction Precedes Development of Rheumatoid Arthritis: A Prospective Cohort Study. EBioMedicine. 2016;6:231-7. [40] Bello N, Etcheto A, Beal C, Dougados M, Molto A. Evaluation of the impact of fibromyalgia in disease activity and treatment effect in spondyloarthritis. Arthritis Res Ther. 2016;18:42. [41] Holman AJ, Ng E. Heart rate variability predicts anti-tumor necrosis factor therapy response for inflammatory arthritis. Auton Neurosci. 2008;143:58-67. [42] Duran J, Combe B, Niu J, Rincheval N, Gaujoux-Viala C, Felson DT. The effect on treatment response of fibromyalgic symptoms in early rheumatoid arthritis patients: results from the ESPOIR cohort. Rheumatology (Oxford). 2015;54:2166-70. [43] Monfredi O, Lyashkov AE, Johnsen AB, Inada S, Schneider H, Wang R, et al. Biophysical characterization of the underappreciated and important relationship between heart rate variability and heart rate. Hypertension. 2014;64:1334-43.
16
[44] Halvorsen S, Vollestad NK, Fongen C, Provan SA, Semb AG, Hagen KB, et al. Physical fitness in patients with ankylosing spondylitis: comparison with population controls. Phys Ther. 2012;92:298309. [45] Hautala AJ, Karjalainen J, Kiviniemi AM, Kinnunen H, Makikallio TH, Huikuri HV, et al. Physical activity and heart rate variability measured simultaneously during waking hours. Am J Physiol Heart Circ Physiol. 2010;298:H874-80. [46] Janse van Rensburg DC, Ker JA, Grant CC, Fletcher L. Effect of exercise on cardiac autonomic function in females with rheumatoid arthritis. Clin Rheumatol. 2012;31:1155-62. [47] Routledge FS, Campbell TS, McFetridge-Durdle JA, Bacon SL. Improvements in heart rate variability with exercise therapy. Can J Cardiol. 2010;26:303-12. [48] Jahan KB, N; Ferdousi, S. Heart rate variability in patients with rheumatoid arthritis. Journal of Bangladesh Society of Physiology. 2012. [49] Janse van Rensburg DC, Ker JA, Grant CC, Fletcher L. Autonomic impairment in rheumatoid arthritis. Int J Rheum Dis. 2012;15:419-26. [50] Gunes Y, Tuncer M, Guntekin U, Sahin M, Yazmalar L. Effects of ankylosing spondylitis on the heart. Acta Cardiol. 2009;64:385-92. [51] Kaya EB, Okutucu S, Aksoy H, Karakulak UN, Tulumen E, Ozdemir O, et al. Evaluation of cardiac autonomic functions in patients with ankylosing spondylitis via heart rate recovery and heart rate variability. Clin Res Cardiol. 2010;99:803-8. [52] Pomeranz B, Macaulay RJ, Caudill MA, Kutz I, Adam D, Gordon D, et al. Assessment of autonomic function in humans by heart rate spectral analysis. Am J Physiol. 1985;248:H151-3.
17
Figure legends. Figure 1 Time and frequency domain HRV for patients with RA vs. controls Figure 2 Time and frequency domain HRV for patients with SpA vs. controls Figure 3 Ewing protocol for patients with RA vs. controls Figure 4 Ewing protocol for patients with SpA vs. controls
18
19
20