Clinical Science

Research article

Low-grade inflammation and arterial wave reflection in patients with chronic fatigue syndrome

Vance A. Spence, Gwen Kennedy, Jill J. F. Belch, Alexander Hill, Faisel Khan


Some of the symptoms reported by people with CFS (chronic fatigue syndrome) are associated with various cardiovascular phenomena. Markers of cardiovascular risk, including inflammation and oxidative stress, have been demonstrated in some patients with CFS, but little is known about the relationship between these and prognostic indicators of cardiovascular risk in this patient group. In the present study, we investigated the relationship between inflammation and oxidative stress and augmentation index, a measure of arterial stiffness, in 41 well-characterized patients with CFS and in 30 healthy subjects. AIx@75 (augmentation index normalized for a heart rate of 75 beats/min) was significantly greater in patients with CFS than in control subjects (22.5±1.7 compared with 13.3±2.3% respectively; P=0.002). Patients with CFS also had significantly increased levels of CRP (C-reactive protein) (2.58±2.91 compared with 1.07±2.16 μg/ml respectively; P<0.01) and 8-iso-prostaglandin F isoprostanes (470.7±250.9 compared with 331.1±97.6 pg/ml respectively; P<0.005). In patients with CFS, AIx@75 correlated significantly with logCRP (r=0.507, P=0.001), isoprostanes (r=0.366, P=0.026), oxidized LDL (low-density lipoprotein) (r=0.333, P=0.039) and systolic blood pressure (r=0.371, P=0.017). In a stepwise multiple regression model, including systolic and diastolic blood pressure, body mass index, CRP, tumour necrosis factor-α, interleukin-1, oxidized LDL, high-density lipoprotein-cholesterol levels, isoprostanes, age and gender, AIx@75 was independently associated with logCRP (β=0.385, P=0.006), age (β=0.363, P=0.022) and female gender (β=0.302, P=0.03) in patients with CFS. The combination of increased arterial wave reflection, inflammation and oxidative stress may result in an increased risk of future cardiovascular events. Assessment of arterial wave reflection might be useful for determining cardiovascular risk in this patient group.

  • arterial stiffness
  • augmentation index
  • cardiovascular risk
  • chronic fatigue syndrome
  • inflammation
  • pulse wave analysis


CFS (chronic fatigue syndrome) is a heterogeneous disorder of unknown aetiology affecting neurological, immunological and cardiovascular systems. The syndrome, which is poorly understood, is characterized by disabling exercise intolerance with significantly reduced activity levels that have been linked to intracellular immune deregulation [1]. There is also accumulating evidence that the cardiovascular system is compromised in many patients with CFS, with reports of autonomic dysfunction [2], attenuated HR (heart rate) and BP (blood pressure) regulation [3] with increased vasomotor tone and loss of beat-to-beat HR control [4]; such abnormalities contribute to destabilization of BP and orthostatic intolerance [5] and, in the more severe cases of CFS, a reduced cardiac output [6].

Oxidative processes are increasingly being linked to symptoms in CFS [79], and we have demonstrated previously [10] that patients with CFS have increased in vivo lipid peroxidation with elevated levels of F2-isoprostanes which correlate with post-exertional myalgia, the dominating symptom that characterizes most patients. We have also shown [11] that our patients have increased concentrations of active TGFβ1 (transforming growth factor β1), and we have postulated that CFS is pro-inflammatory, with many patients in a pro-oxidant state consistent with significant cardiovascular risk [10].

CRP (C-reactive protein) is an acute-phase protein and a sensitive, but non-specific, biochemical marker of chronic inflammation. Highly sensitive assays have consistently shown CRP to be a robust and independent predictor of cardiovascular risk [12], including in healthy individuals without any evidence of cardiovascular disease [13]. Markers of inflammation, including increased CRP levels, have already been demonstrated in some patients with CFS [14], but little is known about the relationship between chronic inflammation and prognostic indicators of cardiovascular risk in this patient group. In other populations, elevated CRP has been consistently linked to increased pulse wave velocity [15,16] and in some, but not all, studies with AIx (augmentation index) [1518], which is a combined measure of elastic and muscular arterial stiffness and wave reflection.

Van de Putte et al. [19] reported increased arterial stiffness and hypotension in a cohort of 32 adolescent patients with CFS, age range 12–18 years, which could not be explained by changes in arterial wall characteristics or lifestyle changes. Vascular stiffness, however, has an impact upon resting and exercise-induced haemodynamics and may contribute to orthostatic hypotension by blunting cardiovagal baroreflex sensitivity in response to changes of arterial pressure in barosensitive regions, such as the aorta and carotid arteries [19], and may also worsen the cardiac performance under conditions of augmented preload in patients with CFS [6].

Given the association between inflammation and increased arterial stiffness in other populations and the recent emerging evidence that increased arterial stiffness is an independent predictor of adverse cardiovascular outcome [20], the aim of the present study was to investigate the relationship between CRP levels and AIx in well-characterized patients with CFS. AIx has been shown to be a useful marker of cardiovascular risk both in healthy subjects and in patients with cardiovascular disease [21,22], and is independently associated with adverse cardiovascular events [2327].



A total of 41 subjects (19–63 years old) were recruited from a local register of patients with CFS/ME (myalgic encepthalopathy) and underwent a medical examination by the same physician; all were found to fulfil the CDC (Centres for Disease Control) classification for CFS [28]. Any patient with either an inflammatory or systemic disease history was excluded. The mean length of illness was 9.2 years (S.D., 5.7 years). A total of 17 patients were on more than one medication; 10 patients were on low-dose amitriptyline taken at night for sleeping problems, eight were on thyroxine, 13 patients were on paracetamol-based drugs, three were on β-blockers and five were on benzodiazepine derivatives. A total of 30 healthy volunteers (19–63 years old) were also recruited from the local community and served as the control group. Details of the patients with CFS and control subjects are given in Table 1.

View this table:
Table 1 Physical and clinical characteristics of patients with CFS and control subjects

Values are means±S.D. *P<0.05, **P<0.01 (using non-parametric tests) and †P<0.005 compared with patients with CFS.

The local Medical Ethics Committee approved the study, and all volunteers gave written informed consent.

Sample collection and analysis

Blood samples were obtained from the antecubital fossa and collected into tubes containing EDTA for plasma or clotting beads for serum preparation. All blood samples were taken at the same time of day. The blood was centrifuged immediately for EDTA plasma preparation for 15 min at 2000 g at 4 °C. The serum tube was placed in a water bath at 37 °C for 1 h before it was centrifuged for 15 min at 2000 g at 4 °C. The plasma or serum was removed, aliquoted and stored at −70 °C until assayed for 8-iso-prostaglandin F isoprostanes, CRP, TNF-α (tumour necrosis factor-α), IL-1 (interleukin-1), total cholesterol, HDL (high-density lipoprotein)-cholesterol and oxLDL [oxidized LDL (low-density lipoprotein)]. Plasma isoprostanes were measured by GC/MS, following the method described by Roberts and Morrow [29]. Total cholesterol and HDL-cholesterol levels were measured on a Cobas Bio centrifugal analyser using products from Roche. Plasma oxLDL levels were measured by ELISA (Mercodia), serum levels of CRP was measured using a high-sensitivity ELISA (Kalon Biological), and serum levels of TNF-α and IL-1 were measured by high-sensitivity ELISAs (R&D Systems).

Assessment of AIx: pulse waveform analysis

Measurements were conducted in a temperature-controlled room (23±1 °C). Subjects were rested in a supine position for at least 10 min after which BP was measured in triplicate using an automated BP monitor (Omron 705 CPII). An index of arterial stiffness was assessed non-invasively by measuring the AIx, a composite measure of arterial stiffness and wave reflections, using the validated SphygmoCor pulse waveform analysis system (Scanmed Medical Instruments) [30,31]. Peripheral pressure waveforms were recorded at the radial artery by applanantion tonometry using a high-fidelity micromanometer (Miller Instruments). At least 15 high-quality pressure waveform recordings were obtained from which the central aortic pressure waveform was calculated using a validated generalized transfer function. From the averaged aortic pulse wave, the following parameters were calculated: (i) AIx, defined as the augmented pressure divided by the pulse pressure and expressed as a percentage; (ii) AIx@75 (AIx normalized for an HR of 75 beats/min), which was calculated to take into account the effect of HR on AIx [31,32]; and (iii) Tr (time to return of the reflected wave), which was used as a marker of pulse wave velocity [31].

Statistical analysis

Values are expressed as means±S.D., unless stated otherwise. Differences in values between patients with CFS and control subjects were tested using unpaired Student's t tests or non-parametric tests if the data were not normally distributed. The independent effect of potential determinants [SBP (systolic BP), DBP (diastolic BP), BMI (body mass index), CRP, TNF-α, IL-1, oxLDL, HDL-cholesterol levels, isoprostanes, age and gender) on AIx were assessed using stepwise multiple regression models. For non-normally distributed data, values were log-transformed for regression analysis. A P value of <0.05 was considered significant. All analyses were performed using SPSS statistical package (version 13).


Patients with CFS and control subjects were matched for age, gender distribution, smoking status, height and weight (Table 1). SBP, DBP and HR were significantly greater in patients with CFS compared with control subjects (P<0.005, P<0.005 and P<0.05 respectively).

Total cholesterol was similar in the two groups, but HDL-cholesterol was significantly lower in patients with CFS compared with control subjects (P<0.005). Patients with CFS had significantly increased levels of CRP (P<0.01, as determined using a Mann–Whitney U test) and 8-iso-prostaglandin F isoprostanes (P<0.005). There were no significant differences in IL-1 and TNF-α between the two groups.

AIx was significantly greater in patients with CFS than in control subjects (24.4±1.8 compared with 16.0±2.8 respectively; P=0.017). AIx was 52% higher in patients with CFS than in control subjects; however, as AIx is influenced by HR, which was significantly higher in patients with CFS, an index normalized for 75 beats/min was used to provide a better comparison of arterial wave reflection between patients with CFS and control subjects. This index, AIx@75, was also significantly greater in patients with CFS than in control subjects (22.5±1.7 compared with 13.3±2.3 respectively; P=0.002), but the difference (69%) was greater than for the unadjusted AIx (52%). Tr was significantly shorter in patients with CFS compared with control subjects (137.6±21.4 compared with 146.5±12.3 ms respectively; P=0.039), indicating a faster pulse wave velocity.

Table 2 shows the significant univariate associations between markers of arterial stiffness and physical and biochemical variables in patients with CFS. AIx@75 had the strongest correlations with SBP, logCRP, isoprostanes and oxLDL. Tr was significantly correlated with logCRP and isoprostanes. There were borderline associations between AIx@75 and age (r=0.265, P=0.099) and gender (r=0.297, P=0.087). In a stepwise regression model for the patients with CFS, the independent associations with AIx@75 were logCRP (β=0.385, P=0.006), age (β=0.363, P=0.022) and female gender (β=0.302, P=0.03). For Tr, the only independently associated parameter was logCRP (β=−0.420, P=0.012).

View this table:
Table 2 Univariate associations between markers of arterial stiffness and physical and biochemical parameters in patients with CFS

In the control group, independent associations with AIx@75 were female gender (β=0.516, P<0.001) and brachial DBP (β=0.301, P=0.026). There were no independent associations with Tr.


The results of the present study have shown that patients with CFS have higher serum CRP levels, elevated levels of isoprostanes and oxLDL, and significantly increased AIx@75, a composite measure of arterial stiffness and wave reflection. AIx@75 significantly correlated with CRP and isoprostanes. Patients with CFS also had higher BP and lower HDL-cholesterol levels than control subjects, which could potentially affect AIx@75. However, the independent determinants of AIx@75 in patients with CFS were CRP, age and female gender. The combination of increased arterial wave reflection, inflammation and oxidative stress may result in unfavourable haemodynamics and an increased risk of a future cardiovascular event in these patients.

Very few long-term follow-up studies have been carried out in patients with CFS and none on the occurrence of other health conditions, so it is not possible to estimate cardiovascular risk in this patient group. However, Lerner et al. [33] documented repetitively abnormal T-wave oscillations, indicative of cardiomyopathy, in many of their patients with CFS and, using radionuclide ventriculography, post-exertional abnormal cardiac wall motion was demonstrated in approx. one-quarter of patients with CFS, indicative of cardiac dysfunction [34]. Furthermore, in a recent report [35], heart failure was seen as a major cause of death in a sample of CFS mortalities and to occur at a considerably younger age than would have been expected from the general population.

Although increased levels of CRP have limited diagnostic value in CFS [14], CRP levels are indicative of chronic, low-grade, subclinical inflammation acting as a potential adjunct in the global prediction of cardiovascular risk [16,18]. Prospective epidemiological studies covering 30 years found that a single CRP measurement strongly predicted a cardiovascular event in a dose–response manner, i.e. continuous across the full range of CRP levels, and that the prediction was independent of traditional risk factors, such age, smoking, hypertension, dyslipidaemia and diabetes [36]. Others have recently reported [37] that markers of inflammation in post-infective fatigue syndromes are not sustained into the chronic phase of illness and that they play no role in the development of persistent symptoms in an Australian population; however, CFS is heterogeneous and there is considerable diversity within and between patient cohorts used in research [38].

Our present study is the first to report CRP levels in patients with CFS, but others have reported increased arterial wave reflection in CFS, albeit in adolescents with the illness [19]. Whether chronic subclinical inflammation is the cause of increased arterial wave reflection in patients with CFS remains to be determined, although it has been positively correlated with indices of arterial stiffness in apparently healthy subjects in the general population [16], patients with systemic vasculitis [15] and children with Kawasaki disease [39], but not in women with active rheumatoid arthritis [40]. Particular lifestyle characteristics, such as smoking, obesity and physical fitness, also play a role in the development of arterial stiffness [41], but only a small percentage of our patients were smokers, and body weight and BMI were similar in patients with CFS and control subjects. Lower levels of physical conditioning are associated with increased arterial stiffness [42] and, by definition, most patients with CFS perform very little physical activity, even when compared with sedentary healthy individuals. Experimentally and epidemiologically, it is evident that long-term physical activity has an anti-inflammatory effect [43]. In a recent study [44], markers of inflammation were shown to be inversely related to physical activity and fitness in sedentary men with well-managed hypertension, and although prescriptive walking regimens improve fitness and reduce cardiovascular risk in sedentary civil servants, they failed to induce any significant reduction in CRP [45]. Furthermore, although physical activity might be effective in modulating pro-inflammatory activity in atherosclerotic disorders [43], many patients with CFS are exercise-intolerant, possibly because of oxidative damage within exercising muscle [46,47] or due to increased plasma levels of oxidation, such as isoprostanes [10], that are associated with post-exertional myalgia in other disorders [48].

Oxidative stress may also be pro-inflammatory and, in adults free of clinical cardiovascular disease, markers of oxidation correlate significantly with CRP levels, independently of other confounders, such as BMI [49]. Although there are significant univariate correlations between measures of oxidation, such as increased levels of isoprostanes and oxLDL, and decreased HDL-cholesterol and AIx in our present study, none of these proved to be independent determinants of arterial wave reflection within the multiple regression model. Patients in our present study group have significantly increased neutrophil apoptosis [11], and there is recent evidence that CRP-degradation products, generated by neutrophil elastase, serve to promote neutrophil apoptosis [50]. In circumstances where CRP is degraded by neutrophil elastase, CRP acquires an anti-inflammatory capacity that functions to promote neutrophil clearance [50] and may offer protection against inflammatory injury.

Development of arterial stiffness results from a complex interaction between structural elements of the vessel wall involving scaffolding proteins, such as collagen and elastin [51], the latter being influenced by serum elastase activity. A significant positive relationship exists between the level of disability in CFS and elastase activity [52], and it is also a key determinant of submaximal and peak exercise intensity in these patients [1]. Pulse wave velocity, but not AIx, correlates with serum elastase activity in apparently healthy individuals and in those with systolic hypertension, giving rise to the possibility that inflammation and increased elastase activity contribute to arterial stiffness [53]. However, not all studies show that elastases are positively correlated with arterial stiffness [54]. Although we did not measure arterial stiffness directly, we cannot exclude the possibility that our measurement of AIx, which is a composite measure of arterial stiffness and wave reflections, might be associated with elastase activity. We did, however, measure the Tr, which is an indirect measure of pulse wave velocity, and found that this was significantly shorter in patients with CFS, suggestive of an increase in pulse wave transmission, and that CRP was independently associated with Tr.

We appreciate that the gold standard measure of arterial stiffness is pulse wave velocity. AIx is mainly determined by the speed of the wave, which is, in turn, determined by arterial stiffness, the amplitude of reflected waves, and the duration and pattern of ventricular ejection. Nevertheless, AIx has been also been shown to be an important determinant of cardiovascular risk and outcome [2127].

A limitation of the present study is that we cannot be certain of the extent to which arterial wave reflections were affected in patients with CFS by drug therapy, such as thyroxine and β-blockers. We did, however, correct for any possible effect on HR by using the AIx@75.

In conclusion, in the present study, we have demonstrated an independent association between increased arterial wave reflection and low-grade inflammation in patients with CFS. The results of the present study are hypothesis-generating and the clinical impact of this association cannot be determined at this point. Nevertheless, since measures of arterial stiffness are predictive of cardiovascular outcome they might be useful for assessing cardiovascular risk in this patient group. There is, however, uncertainty surrounding both the diagnosis and prognosis of CFS, so further studies will be required to address the relationship between inflammation and vascular stiffness in a wider population of patients with CFS. It also remains to be determined whether suppression of the chronic inflammatory process in carefully selected patients with CFS leads to an improvement in arterial stiffness and cardiovascular outcome in the longer term.


We thank Dr Christine Underwood for carrying out the medical examinations. This work was supported by a project grant from ME Research UK.

Abbreviations: AIx, augmentation index; BMI, body mass index; BP, blood pressure; CFS, chronic fatigue syndrome; CRP, C-reactive protein; DBP, diastolic BP; HDL, high-density lipoprotein; HR, heart rate; AIx@75, AIx normalized for a HR of 75 beats/min; IL-1, interleukin-1; LDL, low-density lipoprotein; oxLDL, oxidized LDL; SBP, systolic BP; TNF-α, tumour necrosis factor-α; Tr, time to return of the reflected wave


View Abstract