Population prevalence and symptom associations of autonomic dysfunction in primary biliary cirrhosis


  • Potential conflict of interest: Nothing to report.


Patients with primary biliary cirrhosis (PBC) frequently experience significant fatigue thought to result from as-yet-unidentified central nervous system (CNS)-mediated processes. Pilot studies have suggested that autonomic dysfunction is a frequent occurrence in PBC and may contribute to the pathogenesis of this fatigue. The degree to which autonomic dysfunction affects the PBC population as a whole, and its interrelationship with other symptoms experienced by PBC patients remains unstudied. In this study, we used a geographically defined, fully representative PBC patient cohort to study the prevalence of symptoms of autonomic dysfunction and its relationship with other symptoms of PBC. Symptoms of cardiovascular autonomic dysfunction (as assessed using the Orthostatic Grading Scale [OGS]) were significantly more frequently reported and significantly more severe in PBC patients than in both matched normal controls (40% versus 6% with moderate or worse orthostasis (P < .0001), mean OGS score 3.2 ± 3.4 versus 1.3 ± 1.9, P < .005) and in patients with primary sclerosing cholangitis and in severity were independently associated with severity of fatigue and cognitive symptoms (both r2 = 0.3, P < .0001). Thirteen of 20 patients with an OGS value > 4 (moderate severity and worse) had significant abnormality in autonomic regulation of blood pressure, which was identified on dynamic testing. Conclusion: Symptoms suggestive of autonomic dysfunction frequently occur in PBC patients and reflect dysregulation of actual blood pressure. Autonomic dysfunction is independently associated with both fatigue and, importantly, symptoms of cognitive dysfunction, suggesting the potential for significant organic sequelae. (HEPATOLOGY 2007;45:1496–1505.)

Primary biliary cirrhosis (PBC) is a chronic cholestatic liver disease affecting up to 1 in 700 women over the age of 40 in western European populations.1 It has become widely accepted in the last few years that this condition is characterized by profound fatigue that can often significantly impair quality of life.2–6 The severity of fatigue experienced by PBC patients appears to be unrelated to the degree of hepatocellular dysfunction.3, 4, 6 Studies using animal models of cholestasis (exemplified by the bile duct ligated rodent) have suggested that fatigue results from a combination of the biochemical and inflammatory consequences of cholestasis development.7–10 Studies using these animal models have further suggested that the processes responsible for fatigue pathogenesis are mediated in the central nervous system. More limited studies performed in human PBC patients have indicated the likelihood that central nervous system–mediated processes are similarly important in fatigue pathogenesis in human cholestatic subjects.11, 12 Despite increasing appreciation of the importance of fatigue in the experience of PBC patients and localization of key pathogenetic processes to the central nervous system, the understanding of the precise mechanisms responsible for fatigue expression remains limited. To date, this has impeded progress in developing therapies able to reduce the impact of fatigue in PBC. Recent studies performed by our group and others in selected PBC patients have begun to identify biological processes that may make a significant contribution to the mechanisms of fatigue expression. One example is the recent demonstration that abnormality in sleep patterns, in particular, excessive daytime somnolence, was strongly associated with fatigue in a subgroup of PBC patients.12

Another, less extensively studied biological process showing a potential association with fatigue in PBC is abnormality in regulation of the autonomic nervous system, particularly manifested as cardiovascular instability. In studies performed to date, abnormalities of autonomic regulation of the cardiovascular system have been found at increased incidence in patients with disease at all stage,13–17 and appear to be associated with fatigue.16, 17 The possibility that autonomic dysfunction may be a centrally mediated process that plays a role in the expression of fatigue in PBC is supported by the observation that a number of fatigue-associated conditions including Parkinson's disease, multiple sclerosis, chronic fatigue syndrome/ME (myalgic encephalomyelitis), and primary autonomic failure are characterized by similar degrees of autonomic dysfunction.18–21 What is not clear from the previous PBC studies, which were based on clinic-recruited patient series, is the true extent to which autonomic dysfunction is experienced by the wider PBC population and its specificity to that population. It is also unclear from these studies how autonomic dysfunction relates to abnormality in the other biological processes already implicated in PBC fatigue pathogenesis such as abnormal daytime sleep regulation.

In this study, we set out to explore the population impact of the symptoms of cardiovascular autonomic dysfunction in PBC and the interrelationship between such symptoms and fatigue and sleep disturbance in a unique, representative cohort of PBC patients defined by area of residence. This geographically defined cohort, which by its nature is representative of the PBC population as a whole, has previously provided important insights into the clinical associations of this disease.5, 22–24 We then went on to study the extent to which the symptoms suggestive of cardiovascular autonomic dysfunction experienced by PBC patients were actually associated with abnormalities of blood pressure regulation as shown by dynamic testing.


ESS, Epworth Sleepiness Scale; FIS, Fatigue Impact Scale HADS, Hospital Anxiety and Depression Scale; OGS, Orthostatic Grading Scale; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis.

Patients and Methods


The potential cohort of the PBC symptom study consisted of all 164 patients identified as having the disease residing in the geographical area defined by postal codes NE1-NE25 (Newcastle-upon-Tyne and surrounding suburbs) on the study date. To be eligible for inclusion in the study cohort, patients had to definitely (all 3 of the antimitochondrial antibodies or PBC-specific antinuclear antibodies at a titer of ≥1:40 as determined by immunofluorescence, cholestatic liver function tests, and compatible liver histology) or probably (2 of these 3 criteria) have the disease as previously defined.3, 5 This cohort is continuously updated, and the rationale for studying this patient population and the methods used to identify its members have previously been described.5, 22–24 Ongoing full clinical follow-up records, including histological and laboratory parameters of disease severity, parameters of portal hypertension, and medication history are maintained on the patients in this cohort. Patients in the potential study population were contacted by mail and asked to complete and return a series of questionnaires in a prepaid envelope. Twenty consecutive PBC patients experiencing symptoms of orthostatic hypotension subsequently went on to have a full cardiovascular assessment. Two control groups were studied. Normal control subjects (n = 50) were recruited from the same geographical area as the PBC patient cohort by notices in the local press and in the hospital asking for healthy volunteers to participate in research projects. Our ethical board conditions prevented us from actively screening these study volunteers for undiagnosed disease. The liver disease control cohort consisted of the complete cohort of patients with primary sclerosing cholangitis (PSC) living in the same geographical area as the PBC patients (NE1-NE25) and identified as using the same approaches as the PBC patients (n = 31). Information was gathered for all participants about medications they were taking that could potentially alter regulation of blood pressure (antihypertensive, antianginal, and diuretic agents and antidepressants). The study was approved by the local research ethics committee, and return of a questionnaire was considered consent for the use of data. All data were coded and returned anonymously.

Symptom Assessment Tools


Health-related quality of life and the symptoms that contribute to its impairment in PBC of the PBC patient cohort were assessed using the PBC-40, a fully validated PBC-specific, multidomain, quality-of-life measure.25, 26 The PBC-40 contains 40 questions in 5 domains (fatigue; itch; cognitive; social and emotional; and other symptoms, the latter made up of several PBC-related symptoms that do not map to the other domains). Previous studies performed during the validation of the PBC-40 demonstrated strong correlation of the fatigue domain with the other symptoms, cognitive, and social and emotional domains (which in turn all correlated with each other), but no correlation with Itch scores.26 To complete the PB-40, participants rate items on a 5-point scale (1 = “never”, 2 = “rarely”, 3 = “sometimes”, 4 = “most of the time”, and 5 = “always”). To develop clinically meaningful bands of symptom severity in this study, symptoms reported by a participant in a domain were assessed as follows: (1) None: All responses “never”. (2) Mild: Average of all responses is“rarely,” with up to half of items rated “sometimes”. (3) Moderate: Average of all responses is “sometimes”, with at least half of items rated “most of the time”. (4) Severe: All scores above the cutoff for moderate.

The specific numerical cutoffs for each of these 4 levels of impact for all domains of the PBC-40 calculated using this approach are given in Table 1.

Table 1. Defined Score Ranges for the PBC-40 Domains
PBC-40 domainNoneMildModerateSevere
Symptoms< 78-1819-25> 26
Itch< 34-89-11> 12
Fatigue< 1112-2829-39> 40
Cognitive< 67-1516-21> 22
Social and emotional< 1314-3435-49> 50

Orthostatic Grading Scale.

Studies suggest a role for autonomic dysfunction and fatigue in both nonhepatic diseases and PBC. Participants therefore completed the Orthostatic Grading Scale (OGS), a fully validated self-report assessment tool for the symptoms of orthostatic intolerance due to orthostatic hypotension (e.g., severity, frequency, and interference with daily activities), It consists of 5 items, each rated on a scale from 0 to 4.27 The total score is the sum of the scores of the individual items. Studies have shown that scores from the OGS correlate with conventional tests of integrity of the autonomic nervous system. A score above 9 is considered consistent with a formal diagnosis of orthostatic hypotension. A score of at least 4 but below 9 is considered indicative of moderate orthostatic hypotension

Epworth Sleepiness Scale.

In view of the recently identified association of excessive daytime sleep with fatigue in PBC in several cohorts of clinic patients,12 all participants completed the Epworth Sleepiness Scale (ESS), whose score can range from 0 to 24. This fully validated tool assesses daytime hypersomnolence, with a score of 10 or more indicative of significant daytime hypersomnolence.28 A score of at least 5 but below 10 is considered an indication of moderate daytime hypersomnolence.

Hospital Anxiety and Depression Scale.

The Hospital Anxiety and Depression Scale (HADS) is a 14-item measure of current anxiety (HADS-A) and depression (HADS-D). The HADS was specifically developed for use with individuals who have physical illness because it excludes items related to somatic symptoms. The HADS scores are categorized according to “caseness” (probable presence of the mood disorder), with 0–7 = noncaseness, 8–10 = borderline caseness, and 11–21 = caseness.

Fatigue Impact Scale.

The Fatigue Impact Scale (FIS), a 40-item generic fatigue impact scale, was used to assess the severity of the fatigue of those in the PSC control group because it is only appropriate to use the PBC-40, a disease-specific quality-of-life measure, with patients who have PBC. The FIS has been validated and extensively used in both PBC and PSC populations.3, 5, 29

Cardiovascular Assessment of PBC Patients with Symptoms of Orthostatic Hypotension

To determine the relationship between symptoms suggestive of cardiovascular autonomic dysfunction and actual autonomic regulation of blood pressure, 20 consecutive PBC patients who had described orthostatic dizziness underwent formal autonomic assessment in the cardiovascular laboratory. These patients were assessed for neurally mediated hypotension (defined as either orthostatic hypotension or vasovagal syncope). All subjects refrained from smoking and ingestion of caffeine on the day of investigation and ate only a light breakfast. All investigations were performed at the same time of day and took place in a warm, quiet room. All cardiovascular assessments were carried out with continuous heart rate and beat-to-beat blood pressure measurement (Task Force, CNSystems, Graz, Austria). Orthostatic hypotension and vasovagal syncope were diagnosed using recognized diagnostic criteria.30, 31 The prevalence of neurally mediated hypotension in the PBC study population was compared with that in our whole clinic cohort between January 1997 and August 2005 (3729 consecutive patients).

To detect orthostatic hypotension, heart rate and blood pressure responses to standing more than 2 minutes were measured. In addition, all patients underwent a head-up tilt test to examine the responses of heart rate and blood pressure to prolonged standing and to diagnose vasovagal syncope. In all head-up tilt tests patients were monitored supine for 10 minutes (phase 1) and then brought to standing at 70° by tilt bed. Patients then stood quietly for 20 minutes (phase 2). At the end of this period 400 μg of the vasodilator glyceryl trinitrate was administered sublingually, and the test was continued for another 15 minutes (phase 3). At the end of this time, the bed was returned to the supine position, and hemodynamics were monitored for an additional 10 minutes (phase 4).32 The test was terminated early at the patient's request if syncopal or presyncopal symptoms were reproduced or if systolic blood pressure was below 80 mm Hg for longer than 3 minutes. The integrity of the autonomic nervous system was assessed during the head-up tilt test using baroreflex sensitivity (calculated by the Task Force using the sequence method)33 and heart rate variability (using spectral analysis34) to derive total power (power spectral density), low-frequency heart rate variability (predominantly sympathetic), high-frequency heart rate variability (predominantly parasympathetic), and very-low-frequency heart rate variability. The low-frequency/high-frequency ratio was considered an indicator of balance between the sympathetic and parasympathetic nervous systems. In this study, head-up tilt test data were included in an analysis of baroreflex sensitivity and heart rate variability when the patient was able to complete all 4 phases of the head-up tilt test.

Data Analysis

All data were normally distributed and are presented as means and standard deviations. Data were analyzed using Graphpad software (Prism, San Diego, CA). Groups were compared with the Student t test, and proportions of groups were compared with the chi-square test. Correlations between variables were determined using the Spearman rank correlation test. Regression analysis was performed using SPSS (SPSS Corp.)


Study Cohorts.

One hundred and sixty-four people residing in the geographical area NE1-NE25 were identified as having definite or probable PBC using the comprehensive case-finding protocol and were invited to participate in the study. One hundred and thirty-one patients who had not received transplants returned fully completed assessment tool packs and were included in the study (80%). Twelve patients (9.2%) in the participating cohort were men. The median age of those in the study cohort was 67 years (range 31–92 years). Of the 31 patients in the study area identified as having PSC, 22 returned completed assessment tool packs (71%). Clinical characteristics of the PBC and PSC study participants are outlined in Table 2. Median FIS score in the PSC subject group was 7.5 (range 0–46). As previously reported, fatigue severity in the PSC cohort was significantly correlated with HAD score (P = 0.02).29 The normal control group consisted of 50 healthy individuals living in the NE1-NE25 area, of whom 45 (90%) were women. The median age of this group was 67 years (31–85 years). In the 3 study groups, 52 patients in the PBC group (40%), 6 in the normal control group (12%), and 7 in the liver disease control (PSC) group (29%) were taking medications that had the potential to alter blood pressure regulation.

Table 2. Clinical Characteristics of the PBC and PSC Patients in the NE1-NE25 Geographically Defined Cohorts Participating in This Study
Geographical cohort size16431
Number participating in study (% of entirecohort)131 (80%)22 (71%)
Age (years), median (range)67 (31-92)60 (35-82)
Male sex (%)12 (9%)18 (82%)
Diagnosis confirmed by biopsy (%)117 (89%)21 (94%)
Cirrhosis confirmed by biopsy (% of patients biopsied)20 (17%)3 (14%)
Bilirubin (μmol/l), mean13 ± 421 ± 25
Albumin (g/l), mean39 ± 442 ± 3
Alkaline phosphatase (IU/l), mean196 ± 154263 ± 261
Previous diagnosis of hypertension30 (23%)6 (18%)
Previous diagnosis of diabetes30 (23%)0 (0%)
Clinical diagnosis of portal hypertension23 (17%)2 (9%)

Symptomatic Characterization of the PBC Study Cohort.

In the PBC patient group, scores for all domains of the PBC-40 were similar to those recorded in previous studies (symptoms domain, 17 ± 5 [possible range 7-35]; itch domain, 4 ± 3 [possible range 3–15]; fatigue domain, 29 ±12 [possible range 11-55]; cognitive domain, 14 ± 6 [possible range 6–30]; social and emotional domain, 31 ± 12 [possible range 13–65]).25, 26 There was no significant difference in PBC-40 fatigue domain scores between those PBC patients taking potentially blood-pressure-altering medication and those who were not (30 ± 11 versus 29 ± 12, P = ns), with no correlation between the number of such medications taken by a patient and fatigue severity (r2 = 0.00001, P = ns). The distributions of scores of each of the PBC-40 domains in the study population are shown in Fig. 1. The most frequently seen status was none for “itch”; mild for “other symptoms”, “cognitive”, and “social and emotional”; and moderate for “fatigue”, confirming previous observations that fatigue is the symptom with the greatest apparent impact on the quality of life of PBC patients. Of the 131 participants in the study, 78 (60%) described moderate or severe fatigue, defined using the new criteria. In this non–clinic based cohort 32 of the PBC patients (24%) had ESS scores diagnostic of pathological daytime hypersomnolence (ESS ≥ 10), confirming that daytime somnolence is a significant problem in the PBC population as a whole. Anxiety and depression as determined by HADS were significant problems, with 46% of the cohort having scores on the HADS-A that would be borderline or diagnostic of an anxiety disorder (borderline or case) and 33% having scores on the HADS-D borderline or diagnostic for depression (borderline or case). Univariate analysis showed that severity of symptoms of daytime somnolence assessed using the ESS was significantly correlated with scores for all domains of the PBC-40 except the itch domain and with HAD-D (Table 3). When considered in multivariate analysis, the only association with ESS that remained significant was with the fatigue domain of the PBC-40 (β = 0.226, P < .001).

Figure 1.

Proportions of the geographically defined PBC patient population with scores within the newly defined symptom severity bands for the 5 domains of the PBC-40.

Table 3. Correlations Between ESS and Other Assessment Tool Scores for the Community PBC Cohort (All Values Are Normally Distributed)
Assessment toolCorrelation with ESS (r2)
  • *

    Significant at P < 0.0001. Significant values after correction for multiple testing (P value for significance P < 0.003) are denoted in bold.

PBC-40 symptom domain0.2*
PBC-40 itch domain0.04
PBC-40 fatigue domain0.3*
PBC-40 cognitive domain0.2*
PBC-40 social and emotional domain0.1*

Symptoms of Orthostatic Hypotension.

Symptoms consistent with orthostatic intolerance were experienced frequently by PBC patients. Of those in the PBC cohort, 90 patients (69%) had at least 1 symptom of orthostatic hypotension identifiable by OGS, 46 patients (40%) had moderate or greater symptoms of orthostasis (OGS ≥ 4), and 10 patients (8%) had severe orthostasis (OGS ≥ 9, compatible with a formal diagnosis of orthostatic hypotension). There was no significant difference in OGS scores between those PBC patients taking potentially blood-pressure-altering medication and those not (3 ± 4 versus 3 ± 3, P = ns), with no correlation between an increasing number of culprit medications and OGS score (r2 = 0.06, P = ns). In contrast, only 18 of 50 (36%) normal controls (P < 0.0001 versus PBC patients) and 4 of 22 (18%) PSC patients (P < 0.0001 versus PBC patients) showed any features suggestive of orthostatic hypotension, and only 3 of 50 (6%) normal controls (P < 0.0001 versus PBC patients) and 1 of 22 (5%) PSC patients (P < 0.005 versus PBC patients) had moderate or greater symptoms of orthostasis. Symptoms suggestive of severe orthostasis were not seen in any of the normal or PSC controls. Overall, the mean OGS score was significantly higher in the PBC patient group than in either the normal or the PSC controls (Fig. 2). Univariate analysis showed that in the PBC patient group OGS score was significantly correlated with all PBC-40 domain scores and with both HADS-A and HADS-D scores (Table 4). However, no association was seen with daytime somnolence as assessed using the ESS. When all the variables (individual domains of the PBC-40, age, sex, HADS-A, HADS-D, and ESS) were considered in multivariate analysis, the only associations with OGS that remained significant were those with the cognitive and fatigue domains of the PBC-40 (β = 0.161, P = 0.005, and β = 0.096, P = 0.003, respectively). When the association between OGS score and the independently associated cognitive and fatigue domain scores was considered in terms of the newly defined symptom bands, a clear stepwise progression across the symptom band range was seen (Fig. 3). In the PBC group, 58% of the patients with moderate or severe orthostatic symptoms (OGS ≥ 4) had moderate or severe fatigue as assessed using the PBC-40 fatigue domain compared with only 20% of patients with mild or no orthostatic symptoms (OGS < 4, P < .0005; odds ratio [OR] 4.9, 95% CI 2.1–11.9; Fig. 4). In contrast, in the PSC control group the single patient with moderate or severe orthostatic symptoms had only mild fatigue (FIS < 60).

Figure 2.

Comparison of symptoms of orthostasis between the geographically defined comprehensive cohort of PBC patients, normal controls, and the geographically defined PSC patient cohort, as assessed by OGS.

Table 4. Correlations Between OGS and Other Assessment Tool Scores for the Community PBC Cohort (All Values Are Normally Distributed)
Assessment toolCorrelation with OGS (r2)
  • *

    Significant at P < 0.001;

  • Significant at P < 0.0001. Significant values after correction for multiple testing (P value for significance P < 0.003) are denoted in bold.

PBC-40 Symptom Domain0.1*
PBC-40 itch domain0.1*
PBC-40 fatigue domain0.3
PBC-40 cognitive domain0.3
PBC-40 social and emotional domain0.2
Figure 3.

Orthostatic grading scale (OGS) scores for PBC patients in the geographical population with scores in the 4 severity bands of the fatigue and cognitive domains of the PBC-40. The fatigue and cognitive domain scores were the only parameters independently associated with OGS scores on multivariate analysis (*P < .05, **P < .01, ***P < .001).

Figure 4.

Frequency of moderate or severe fatigue in PBC patients with a moderate or high level (OGS ≥ 4) or a low-level (OGS < 4) of symptoms of orthostatic hypotension.

When ESS score data were included in the analysis, it was found that the vast majority of patients in the PBC cohort with moderate or significant fatigue described moderate or worse symptoms of orthostasis (OGS ≥ 4), moderate or worse symptoms of daytime somnolence (ESS ≥ 5), or both (OR versus mild or no fatigue 9.2, 94% CI 3.3–25.3; Fig. 5). Although there was some overlap between orthostatic hypotension and daytime somnolence, with a subgroup of fatigued patients experiencing both symptoms of orthostatic hypotension and daytime somnolence, there was a distinct group of patients with moderate or severe fatigue who experienced moderate or severe orthostatic symptoms in the absence of any daytime somnolence (6 of the 78 patients [8%] describing moderate or severe fatigue). Fifteen of 78 patients (19%) with moderate or severe fatigue appeared to have neither excessive daytime somnolence nor symptoms of orthostatic hypotension, suggesting there is another, as-yet-unidentified mechanism of fatigue pathogenesis in PBC. There were no patients in the PSC group who described moderate or severe fatigue (5 described no fatigue, and 17 described mild fatigue).

Figure 5.

Frequency of moderate or severe fatigue in PBC patients with a moderate or high level of symptoms of orthostatic hypotension (OGS ≥ 4), a moderate or high level of daytime somnolence (ESS ≥ 5), defined as “symptoms,” or both compared with patients with a low level of symptoms of both (OGS < 4 and ESS < 5), defined as “no symptoms.”

Relationship of Symptoms of Orthostatic Hypotension with Dynamic Blood Pressure Regulation.

To examine the nature of the relationship between symptoms suggestive of orthostatic hypotension and objective features of blood pressure control, 20 consecutive PBC patients who described symptoms of orthostatic hypotension via the OGS (all OGS ≥ 4) underwent full cardiovascular assessment including dynamic investigation of blood pressure regulation through the head-up tilt test. All patients in this phase of the study were women. Their mean subject age was 63 ± 11 years. The median PBC-40 fatigue domain score of these 20 patients was 40 (16–55). None of the 20 patients had significant abnormalitiesin their 12-lead electrocardiograph. Fourteen participants (70%) described both postural dizziness and blackouts, 5 (25%) described postural dizziness alone, and 1 (5%) described falls. Six patients (30%) were taking potentially blood-pressure-altering medications at the time of their cardiovascular assessment: 2 were taking diuretics 3 were taking calcium antagonists, and 1 was taking a β-blocker. These medications were stopped after clinical review. Of the 20 patients assessed, significant neurally mediated hypotension was found in 13 (65%). The frequency of significant neurally mediated hypotension in this PBC group was significantly higher than that in equivalent non-PBC populations (of the 3,729 patients who presented to the Falls and Syncope clinic between 1997 and 2005 with symptoms of autonomic dysfunction of the type identified using the OGS, 1,263 [34%] were found to have neurally mediated hypotension [P < 0.005 versus the PBC group]). Of the 13 patients found to have neurally mediated hypotension, 9 had vasovagal syncope, 2 had orthostatic hypotension, and 2 had both orthostatic hypotension and vasovagal syncope. Significant abnormality in autonomic regulation as assessed using head-up tilt test was observed in all patients, including those falling short of a diagnosis of vasovagal syncope (Table 5). Total power and low-frequency and high-frequency heart rate variability were substantially below the lower limit of normal through all phases of the head-up tilt test, confirming the presence of significantly reduced heart rate variability. Baroreflex sensitivity was below the lower threshold after tilting (phase 2) and in response to vasodilation with glyceryl trinitrate (phase 3), confirming that baroreceptor response to provocation by orthostasis and vasodilation was impaired. When PBC-40 fatigue domain score was correlated with hemodynamics of each phase of the head-up tilt test, heart rate was significantly related to fatigue score for all phases of the head-up tilt test, suggesting reflex compensation (Table 6). In addition, there was a significant relationship between increasing fatigue and impaired heart rate variability (very low frequency, low-frequency, and high-frequency heart rate variability in response to glyceryl trinitrate was significantly related to increasing fatigue in phases 3 and 1 (resting), and low-frequency and high-frequency heart rate variability was significantly related to increasing fatigue in phase 4 (recovery), confirming the association of reduced heart rate variability with fatigue. There was no relationship between low-frequency/ high-frequency ratio, suggesting no shift in sympathovagal balance with increasing fatigue. Furthermore, there was a significant relationship of impaired baroreflex sensitivity (total mean slope) with fatigue on passive tilting (phase 2) and recovery. These studies suggest that cardiovascular assessment would show that a significant proportion of PBC patients reporting symptoms suggestive of orthostatic hypotension would have an actual abnormality diagnostic of neurally mediated hypotension (orthostatic hypotension and/or vasovagal syncope) and that with a head-up tilt test, even those patients falling short of a formal diagnosis of neurally mediated hypotension would have significant abnormality diagnostic of autonomic dysfunction.

Table 5. Parameters of Autonomic Cardiovascular Regulation in PBC Patients Who Described Symptoms of Orthostatic Hypotension Assessed During Head-Up Tilt Test Compared with International Normative Values
 Phase 1 (resting)Phase 2 (passive HUT)Phase 3 (post-GTN)Phase 4 (recovery)Normative values
  1. NOTE. Bold denotes values lying outside the normal range.

  2. Abbreviations: BRS, baroreflex sensitivity; LF, low-frequency heart rate variability; HF, high-frequency heart rate variability; TP, total power.

LF (ms2)493 ± 527231 ± 271286 ± 318616 ± 561750-1590
HF (ms2)251 ± 238207 ± 15399 ± 95249 ± 283775-1175
TP (ms2)912 ± 996548 ± 579410 ± 1350144 ± 17942400-4600
LF/HF ratio1.3 ± 1.01.8 ± 1.62.8 ± 2.43.7 ± 2.4< 2
BRS (ms/mm Hg)12 ± 79 ± 56 ± 210 ± 6> 9.3
Table 6. Correlations Between PBC-40 Fatigue Domain Score and Hemodynamic Parameters During the 4 Phases of the Head-Up Tilt Test
  Phase 1 (resting)Phase 2 (passive HUT)Phase 3 (post-GTN)Phase 4 (recovery)
  • NOTE. Data are denoted as r values, with significant values indicated in bold;

  • *

    P < 0.05,

  • P < 0.005.

  • Abbreviations: BRS, baroreflex sensitivity; HRV, heart rate variability; HF, high-frequency heart rate variability; LF, low-frequency heart rate variability; VLF, very-low-frequency heart rate variability; HR, heart rate.

 Mean HR0.60.80.6*0.8


In this study, we demonstrated that in a PBC population defined by area of residence (and therefore likely to be representative of the distribution of PBC patients in UK populations), symptoms associated with cardiovascular autonomic dysfunction were present in almost 70% of patients. Symptoms suggestive of severe orthostatic hypotension, sufficient to warrant referral to a specialist syncope clinic, were present in almost 10% of patients. The incidence and severity of these symptoms was significantly greater in the study PBC population than in age- and sex-matched normal controls from the same study area and in PSC patients defined using the same geographical criteria as the PBC study population (and who, interestingly, were characterized by very low fatigue, consistent with other published observations).29 The clinical assessment tool used in this study, the Orthostatic Grading Scale (OGS), has been validated against objective assessment of cardiovascular status of patients with other diseases and shown to be predictive of true abnormalities in blood pressure regulation.27 Through the use of the head-up tilt test for the first time in PBC patients, in this study we also confirmed32 that this association between symptoms suggestive of blood pressure lability, as detected by the OGS, and true abnormality in regulation of blood pressure remains true for PBC patients. The extent of the symptoms of orthostatic hypotension described by PBC patients showed a significant correlation with severity of both the fatigue experienced by PBC patients and symptoms suggestive of cognitive function, as assessed by the cognitive domain of the PBC-40. Moreover, severity of fatigue also showed a significant correlation with objective parameters of degree of blood pressure dysregulation, assessed using the head-up tilt test. In contrast, symptoms of orthostatic hypotension were apparently independent of degree of daytime somnolence experienced by PBC patients, a symptom that itself has recently been linked with fatigue in PBC.12 Our findings fully support those of previous studies identifying the objective features of autonomic dysfunction in PBC13–17 and extend our understanding of the impact of this effect by suggesting that cardiovascular autonomic dysfunction is both common in PBC patients and significantly associated with several of the key symptoms of the disease. The absence of an association with sleep abnormality would suggest that any role autonomic dysfunction plays in the expression of fatigue in PBC is largely independent of the processes associated with daytime somnolence.

The aim of this study was to explore the extent to which the PBC patient population as a whole has symptoms of cardiovascular autonomic dysfunction, the association of these symptoms with other key disease symptom parameters, and the extent to which such symptoms are indicative of objective cardiovascular instability, rather than to explore the mechanisms actually responsible for such instability. Thus, it is not clear what the responsible mechanisms are. However, several potential mechanisms are known. Autonomic dysfunction, including cardiovascular instability, has been demonstrated to be associated with advanced liver disease in a number of settings, including in patients with PBC, and this undoubtedly will be found to contribute to the symptoms described in at least some of the patients.13–15, 35–39 It is most unlikely, however, that such features of advanced disease are responsible for a symptom complex present in almost three quarters of patients given the predominance of early-stage disease and the very low prevalence of advanced liver disease and/or portal hypertension in the study population. Another potential explanation for a fatigued population having such symptoms and for the apparent correlation of the severity of these symptoms with the severity of fatigue would be that autonomic dysfunction of the type identified here is a consequence of the physical deconditioning that results from the impaired function caused by fatigue, rather than a cause of the fatigue. This mechanism has been postulated to contribute to the association between fatigue and autonomic dysfunction observed in chronic fatigue syndrome.40 We previously demonstrated that the perception of fatigue severity in PBC patients is significantly correlated with actual degree of physical activity undertaken.26 A third and intriguing possibility is that the pathways contributing to autonomic dysfunction in PBC are of central origin. We previously demonstrated deposition of manganese in the basal ganglia of PBC patients, the degree of which correlated significantly with the degree of fatigue experienced.11 The anatomical areas showing this associated effect are adjacent to those responsible for control of the autonomic nervous system.41 It is possible, therefore, that one of the postulated central effects of cholestasis, that of deposition of factors normally cleared in the bile and retained in cholestasis, is directly responsible for autonomic dysfunction.

A key observation in this study was the extent to which the symptoms of orthostatic hypotension, as detected by the OGS, are predictive of true cardiovascular autonomic dysfunction, as detected using the head-up tilt test. It was found that of the 20 serial (and otherwise unselected) patients with symptoms of orthostatic hypotension who underwent the head-up tilt test, almost two thirds exhibited sufficient abnormality in blood pressure regulation to warrant a formal diagnosis of neurally mediated hypotension. Significant abnormality in the parameters of hemodynamic stability, exemplified by heart rate variability, was seen in all patients, including those falling short of a formal diagnosis of neurally mediated hypotension. Strikingly, the severity of these abnormalities, in particular, those relating to heart rate variability, was significantly correlated with fatigue severity. One interesting observation was some patients whose symptoms of orthostatic hypotension were not particularly severe had a very significant objectively determined abnormality in blood pressure regulation, suggesting that even patients with relatively limited symptoms of orthostasis could be susceptible to a very marked abnormality in actual blood pressure regulation.

Our finding that 40% of our geographically defined PBC cohort was taking potentially blood-pressure-altering medications raises an interesting question about the degree to which medication contributes to orthostasis-associated fatigue in the wider PBC population. The similarity in the OGS and PBC-40 fatigue domain scores seen between patients taking such medication and patients not taking such medication, coupled with the number of potentially culprit medication not being correlated with either the OGS or fatigue severity, would argue against this as an overarching mechanism of fatigue in PBC. However, as the head-up tilt test data suggest (6 of the consecutive 20 patients investigated for symptoms of orthostatic hypotension were receiving potentially blood-pressure-altering medications, 5 of whom had neurally mediated hypotension), over- or inappropriate medication may contribute to orthostatic symptoms and potentially be associated with fatigue in individual patients. If, as the findings of this study suggest, PBC is associated with a tendency toward blood pressure dysregulation and hypotension, the ongoing need of all PBC patients to take antihypertensive medication should be critically reviewed, particularly because such therapy has often been prescribed for prolonged periods following a diagnosis of hypertension, which may have been done before the hypotensive effects of PBC were known. In our clinical practice, both the head-up tilt test and 24-hour blood pressure measurement have proved to be invaluable tools for identifying PBC patients receiving, typically for historical reasons, excessive and inappropriate antihypertensive treatment that has contributed significantly to their fatigue.

At present it is unclear how, if at all, blood pressure dysregulation actually contributes to the expression of fatigue. Ongoing studies by our group are addressing the role played by impaired or abnormal muscle perfusion secondary to altered hemodynamics in the expression of fatigue in PBC. However, our findings that the symptoms of orthostatic hypotension correlate significantly with symptoms associated with cognitive dysfunction, as assessed by the cognitive domain of the PBC 40, suggest there may be an additional, potentially important organic consequence of blood pressure lability in PBC. Previous studies of non-PBC populations identified a link between blood pressure lability (in particular, hypotension) and organic cognitive dysfunction, up to and including dementia.42 The possibility that there may be an organic consequence over and above fatigue associated with the autonomic dysfunction in PBC makes understanding the mechanisms underpinning this blood pressure lability particularly important. Once again studies are ongoing in this area.

The final aim of this study was to address the extent to which symptoms of orthostatic hypotension and autonomic dysfunction in PBC were related to those of the recently described fatigue-associated abnormality in sleep regulation. Strikingly, no significant independent association was seen between degree of daytime somnolence and degree of orthostatic hypotension in the PBC population, although each was independently associated with fatigue severity. This observation raises the possibility that the pathways responsible for daytime somnolence and autonomic dysfunction are independent mechanisms underpinning the expression of fatigue, with both potentially originating from the CNS effects of cholestasis and inflammation. The identification of a second potentially independent pathway for the expression of fatigue in PBC gives us further optimism that the development of therapeutic approaches able to improve this important and debilitating symptom will occur in the near future.