Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004), pp. 1584–1588 (C© 2004)

Patients with Autonomic Neuropathy Are MoreLikely to Develop Hepatic Encephalopathy

ANURAG MAHESHWARI, MD, ANIL THOMAS, MD, and PAUL J. THULUVATH MD, FRCP

Autonomic neuropathy (AN) is common in patients with chronic liver disease. For hitherto unknownreasons, in longitudinal studies, the presence of AN has been found to be an independent predictor ofmortality in patients with cirrhosis. We hypothesized that patients with AN are more likely to develophepatic encephalopathy (HE) due to prologed intestinal transit time. In this study, we examined theincidence of new-onset HE in patients with and without AN. Seventy-two patients (Child A/B/C=35/31/6) without evidence of HE at the time of autonomic function testing (AFT) were followedfor 39.5± 27.3 months. The end point of the study was the development of new onset HE. Patientswere followed until death or liver transplantation. Of the 72 patients, 42 (58%) patients did notdevelop HE (group A) while 30 (42%) developed HE (group B) during the follow-up. Both groupshad similar baseline demographics. AN was more common in group B (27/30; 90%) compared togroup A (28/42; 67%) (P = 0.02). Kaplan–Meier analysis showed a trend toward a higher incidenceof HE in patients with AN. Mortality was higher in group B (12/30; 40%) compared to group A(8/42; 19%) (P = 0.04). Patients with AN were more likely to develop new onset HE. Althoughan independent effect of AN on HE was not established in this study, we speculate that delayedintestinal transit secondary to AN may explain the higher incidence of HE in patients with AN.

KEY WORDS: autonomic neuropathy; cirrhosis; hepatic encephalopathy.

Autonomic neuropathy (AN) is equally common in al-coholic and nonalcoholic liver disease (1). Few studieshave shown an increased prevalence of AN and peripheralneuropathy with advancing liver disease (1–5). However,other investigators have not found a consistent relationshipbetween the severity of liver disease and AN (6, 7). We hadpreviously reported that AN is an independent predictorof survival in patients with early and advanced cirrhosis(2, 3). However, the reasons for the higher mortality inthose with AN remain unknown. Prolonged QTc intervalhas been suggested as a possible explanation, but in a re-cent study, we did not observe any relationship between

Manuscript received January 16, 2004; accepted April 2, 2004.From the Division of Gastroenterology and Hepatology, The Johns

Hopkins University School of Medicine, Baltimore, Maryland 21205,USA.

Address for reprint requests: P. J. Thuluvath, MD, FRCP, Divisionof Gastroenterology and Hepatology, The Johns Hopkins UniversitySchool of Medicine, 1830 East Monument Street, Suite 429, Baltimore,Maryland 21205, USA; pjthuluv@jhmi.edu.

AN and QTc interval (8). Moreover, prolonged QTc in-terval has not been shown to be an independent predictorof survival in patients with cirrhosis (8, 9). We speculatedthat the presence of AN, as in a diabetic population withAN, may result in prolonged gastrointestinal transit time,small bowel bacterial overgrowth (SBBO), endotoxemia,and hepatic encephalopathy (HE) (10–18), and a combi-nation of these factors may result in a higher mortality.

To our knowledge, there have been no previous studiesexamining the role of AN and the presence of HE. Theobjective of this study was to determine whether there isa higher incidence of HE in patients with AN and chronicliver disease compared to those without AN.

PATIENTS AND METHODS

In this study, we conducted a retrospective analysis of 156 pa-tients with established cirrhosis who had undergone autonomicfunction testing (AFT) between 1993 and 2002. We excluded pa-tients (n = 23) who had a history or clinical evidence of HE at

1584 Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004)0163-2116/04/1000-1584/0C© 2004 Springer Science+Business Media, Inc.

PATIENTS WITH AN ARE MORE LIKELY TO DEVELOP HE

the time of AFT. We examined patient records to determine new-onset HE. Hepatic encephalopathy was diagnosed when therewas new onset of somnolence, disorientation, or confusion withor without asterexis. Neurpsychometric testing was not done inthese patients. For patients who were diagnosed with the firstepisode of HE in an outside facility, the time point at whichlactulose or neomycin therapy was first instituted was taken asthe first episode of HE. The investigator who examined patientrecords was not aware of the results of AFT. Forty-eight patientswho were not followed up at our institution and 18 patients whohad less than 6 months from the time of AFT to death or trans-plantation were also excluded due to short follow-up. The rest(n = 72) were followed longitudinally for development of new-onset HE.

All patients had documented cirrhosis by histology or imagingcriteria, and data at baseline and follow-up were collected onthe following parameters: presence of ascites, etiology of liverdisease, age, sex, diabetic status, and biochemical parametersincluding albumin, bilirubin, INR, creatinine, and CPT (ChildPugh Turcotte) score. Patients were followed until death or livertransplantation. For those who had liver transplantation, datawere censored at the time of liver transplantation.

All patients underwent AFT at baseline in our neurology lab-oratory using a battery of four tests (1–5). The standardized AFT,first described by Ewinget al.(19), included the following: heartrate variability with deep breathing at a rate of 6/min (the dif-ference between the maximum and the minimum heart rate in1 minute), R–R ratio during and after the Valsalva maneuver (theratio of the longest R–R interval after the Valsalva maneuver tothe shortest R–R ratio during the maneuvre), blood pressure re-sponse to sustained handgrip (change in diastolic BP before andafter the maneuver), and the 30:15 ratio (ratio of the R–R intervalat the 30th beat to the R–R interval at the 15th beat after stand-ing from a supine position). These tests have been validated inour previous studies and were categorized as abnormal if one offour tests was abnormal. Patients were instructed to hold theirdiuretic and antihypertensive medications (such asβ-blockers)for 24 hr prior to the test. However, patients with type II diabetesmellitus (DM) were not excluded from the study.

The end point of the study was the development of new-onsetHE. Patient records were examined at each visit or hospitaliza-tion for the presence of HE. Patients without HE at the timeof AFT were grouped by the absence (group A) or presence(group B) of new-onset HE during follow-up.

Statistical Analysis. Comparison of groups was made bychi-square test for categorical variables and pairedt-test for con-tinuous variables. The incidence of new-onset HE in patientswith and without AN was tested by Kaplan–Meier analysis. Allanalysis were done using SPSS software version 11.0 (SPSS Inc.,Chicago, IL).

RESULTS

Seventy-two patients (Child A, 35; B, 31; C, 6) did nothave evidence of HE at the time of AFT, and of these 55(55/72) had AN present. Of these 72 patients, 42 patientsdid not develop HE during follow-up (group A), while30 patients developed new-onset HE (group B). The meanduration of follow-up was 39.5± 27.3 months (group A,43± 29 months; group B, 34.5± 24 months;P = 0.2).

TABLE 1. DEMOGRAPHICS OFPATIENTS AT BASELINE

Groups A+ B Group A Group B(n= 72) (n= 42) (n= 30) P value*

Age (years) 51.1± 9.2 52.1± 10.4 49.7± 7.6 0.3Male 44 (61%) 24 (57%) 20 (66%) 0.3Diabetes 18 (25%) 9 (21%) 9 (30%) 0.2Ascites 29 (40%) 16 (38%) 13 (43%) 0.4

Severity of liver disease

Albumin 3.5± 0.6 3.6± 0.7 3.4± 0.6 0.2Bilirubin 2.4± 2.0 2.3± 2.1 2.6± 2.0 0.5INR 1.29± 0.33 1.26± 0.32 1.33± 0.35 0.3

Etiology of liver disease

Alcoholic 17 (24%) 7 (17%) 10 (33%)HCV 22 (30%) 14 (33%) 8 (24%)HBV 2 (3%) 1 (2%) 1 (3%)PBC 4 (6%) 3 (7%) 1 (3%)PSC 7 (10%) 7 (17%) 0AIH 5 (7%) 4 (9%) 1 (3%)Others 15 (21%) 6 (14%) 9 (30%)Autonomic 55 (76%) 28 (67%) 27 (90%) 0.02

neuropathy

Note.Group A—patients who did not develop encephalopathy on follow-up; group B—patients who developed encephalopathy on follow-up;groups A+ B—patients without encephalopathy at initial presentation.HCV, hepatitis C; HBV, hepatitis B; PBC, primary biliary cirrhosis; PSC,primary sclerosing cholangitis; AIH, autoimmune hepatitis; INR, Inter-national Normalized Ratio; MELD, model for end-stage liver disease.* P value when groups A and B were compared.

The relatively shorter duration duration of follow-up Bwas due to censoring of data when patients developedHE. Baseline demographic data of the entire cohort andgroups A and B are listed in Table 1. The groups weresimilar at baseline with regard to age, sex, presence of DMand ascites, etiology and severity of liver disease (assessedby CPT score), albumin, INR (International StandardizedRatio), and bilirubin. Eighteen patients had type II DMand they were equally distributed in both groups.

Table 2 shows changes from baseline characteristics forthe entire cohort, group A, and group B. These data werecollected at the last follow-up visit in group A and aroundthe time of development of new-onset HE in group B.Eighteen patients (group A, 5 patients; group B, 13 pa-tients) underwent liver transplantation, and in these pa-tients, follow-up data were censored at transplantation.Bilirubin and INR did not change significantly from base-line in both groups, but serum albumin decreased signif-icantly in group B (Table 2). Prevalence of ascites washigher in both groups at follow-up compared to baseline;9 patients in group A and 11 in group B developed new-onset ascites during the follow-up. Mortality was higherin patients who developed HE; 20 patients died duringfollow-up, 8 from group A and 12 from B (P = 0.04).

Fifty-five patients had evidence of AN and 17 pa-tients had no AN. The characteristics of patients with andwithout AN were similar as shown in Table 3. Among

Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004) 1585

MAHESHWARI ET AL.

TABLE 2. PATIENT CHARACTERISTICS ATFOLLOW-UP COMPARED TOBASELINE

Groups A+ B (n= 72) Group A (n= 42) Group B (n= 30)

Baseline Follow-up P value Baseline Follow-up P value Baseline Follow-up P value

Ascites 29 (40%) 49 (68%) <0.001 16 (38%) 25 (60%) 0.004 13 (43%) 24 (80%)<0.001Albumin 3.5± 0.6 3.2± 0.5 0.02 3.6± 0.7 3.5± 0.5 0.3 3.5± 0.4 2.9± 0.3 <0.001Bilirubin 2.4± 2.0 3.7± 7.2 0.1 2.3± 2.1 3.4± 7.9 0.1 2.6± 2.0 4.2± 5.9 0.4INR* 1.29± 0.3 1.46± 0.6 0.02 1.26± 0.3 1.35± 0.5 0.2 1.33± 0.3 1.62± 0.8 0.06

*International Standardized Ratio.

the 30 patients who developed new-onset HE, 27 had(90%) had AN (P = 0.02) (Table 1); HE developed in49% of patients with AN (versus 18% without AN). Onlythree (3/17) patients without AN developed new-onsetHE. Kaplan–Meier analysis of patients with new-onsetHE grouped by the presence or absence of AN is shown inFigure 1. There was a clear trend toward a higher incidenceof HE in those with autonomic neuropathy, however, thisdid not reach statistical significance (log rank test). Whendata were analyzed after excluding patients who had livertransplantation, all patients (100%) with HE had evidenceof AN.

DISCUSSION

In this longitudinal study, we have shown that, de-spite the similar severity of liver disease at baseline, pa-tients with AN were more likely to develop HE during amean follow-up of 3–4 years. However, serum albumindecreased significantly in those who had HE, and there-

TABLE 3. BASELINE DEMOGRAPHICS OFPATIENTS BASED ON THE

PRESENCE OFAN IN PATIENTS WITHOUT HE (n = 72) AT THE

TIME OF AFT

Patients with Patients withoutAN (n= 55) AN (n= 17) P value

Age (months) 51.6± 9.4 49.6± 8.7 0.4Male 33 (60%) 11 (65%) 0.8Diabetes 14 (25%) 4 (23%) 0.5Ascites 23 (42%) 6 (35%) 0.7HE at follow-up 27 (49%) 3 (18%) 0.02

Severity of liver disease

Bilirubin 2.4± 1.9 2.6± 2.6 0.6Albumin 3.5± 0.7 3.7± 0.5 0.2INR 1.3± 0.3 1.2± 0.4 0.4CPT score 7± 1.9 6.7± 1.5 0.5

Etiology of liver disease 0.5

Alcoholic 13 (24%) 4 (23%)HCV 16 (29%) 6 (35%)HBV 2 (4%) 0PBC 4 (7%) 0PSC 4 (7%) 3 (18%)AIH 3 (5%) 2 (12%)Others 13 (24%) 2 (12%)

fore, we could not prove an independent effect of AN onnew-onset HE. Despite this limitation, we feel our dataare interesting and merit further attention and, moreover,believe that our findings may partly explain the highermortality observed in patients with AN. Causal effect re-lationship could not be established in our study, but if ourfindings are corroborated by other investigators, they maypartly explain why some patients are more prone to de-velop severe HE while others with a similar severity ofliver disease do not develop HE.

The reasons for the higher prevalence of HE in patientswith AN remains unproven, but we believe, based on cir-cumstantial evidence, that this may be due to delayed gas-trointestinal transit time, which may result in small bowelbacterial overgrowth (SBBO), hyperammonemia, and en-dotoxemia. Previous studies have shown delayed gastroin-testinal transit time in patients with liver disease (10–13).A study involving 30 patients with cirrhosis and varyingdegrees of encephalopathy found delayed orocecal tran-sit time in patients with HE using the lactulose hydro-gen breath test (11). The authors of this study speculatedthat the intestinal production of neurotransmitter-like sub-stances may be responsible for the delayed intestinal tran-sit time as well as HE. However, other studies have showndelayed intestinal transit time in patients with chronic liverdisease in the absence of HE (10, 12, 13). There has beenno satisfactory explanation for the increase in intestinaltransit time in patients with cirrhosis; it has been specu-lated that the delay in intestinal transit could possibly bedue to deranged sex hormone metabolism.

SBBO is often seen in chronic liver disease (14–16). Ithas been suggested that this condition in patients with cir-rhosis may contribute to endotoxemia (15), spontaneousbacterial peritonitis (SBP) (16), and HE. In a prospec-tive study, Morencoset al. (14) showed that SBBO wasmore common in patients with ascites or advanced liverdisease (Child C cirrhosis), and this was associated witha higher incidence of SBP. In diabetics without liver dis-ease, a close relationship between AN and SBBO has beenshown (17, 18). In our study, we did not exclude patientswith type II DM. We had nine patients in both group A andgroup B with similar prevalence of HE and AN, and our

1586 Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004)

PATIENTS WITH AN ARE MORE LIKELY TO DEVELOP HE

Fig 1. Kaplan–Meier analysis of cumulative incidence of new-onset hepatic encephalopathy in patients with (solid line)and without (. . . .) autonomic neuropathy (AN).

results were similar even when we excluded those withdiabetes. To our knowledge, there have been no studiesthat have examined the relationship among AN, gastroin-testinal transit time, and bacterial overgrowth in chronicliver disease. A prospective study to examine the relation-ship between AN and orocecal transit time is in progressat our institution. Until a close relationship among AN,gastrointestinal transit time, and HE is established, wecan only speculate on the role of AN in the pathogenesisof HE.

In our study, despite similar baseline serum albumin,bilirubin, INR, and ascites, patients who developed HEhad worsening serum albumin at follow-up compared tobaseline. Lower albumin was probably related to wors-ening liver disease, but it is also possible that it was dueto protein restriction in those who developed HE. In ourprevious studies, we showed a higher prevalence of ANin more advanced liver disease (1–5). However, it is im-portant to note that some investigators have not seen anyrelationship between the severity of liver disease and theprevalence of AN (8, 9). In this study, we excluded patientswith HE at the onset of our study, and in those who hadno HE, the severity of liver disease was similar in patientswith and without AN. We therefore could not establish anindependent relationship between AN and HE. We alsorelied on clinical records, in a retrospective manner, to de-

termine the presence of HE and may have missed manysubjects with subclinical encephalopathy. However, samecriteria were used for all patients without the knowledgeof AFT test results.

Despite some of the limitations, our study brings for-ward a novel concept linking the presence of AN to thedevelopment of HE. Our observations may also explainthe inferior survival rates noted in cirrhotics with AN. Ifdelayed small bowel transit time is a major contributoryfactor in the pathogenesis of HE, prokinetic agents may bebeneficial in the management of HE, especially in thosewho are refractory or noncompliant to conventional med-ical treatment. Cisapride has been shown to be effectivein reducing intestinal bacterial overgrowth and bacterialtranslocation in rats with both cirrhosis (20) and acute liverfailure (21). In summary, in this retrospective longitudi-nal study, we have shown an association between AN andthe development of encephalopathy among patients withchronic liver disease.

REFERENCES

1. Thuluvath PJ, Triger DR: Autonomic neuropathy and chronic liverdisease. Q J Med 268:737–747, 1989

2. Hendrickse MT, Thuluvath PJ, Triger DR: Natural history of auto-nomic neuropathy in chronic liver disease. Lancet 339:1462–1464,1992

Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004) 1587

MAHESHWARI ET AL.

3. Fleckenstein JF, Frank SM, Thuluvath PJ: Presence of autonomicneuropathy is a poor prognostic indicator in patients with advancedliver disease. Hepatology 23:471–475, 1996

4. Hendrickse MT, Triger DR: Peripheral and cardiovascular auto-nomic impairment in chronic liver disease: Prevalence and relationto hepatic function. J Hepatol 16:177–183, 1992

5. Chaudhry V, Corse AM, O’Brien R, Cornblath DR, Klein AS,Thuluvath PJ: Autonomic and peripheral (sensorimotor) neuropa-thy in chronic liver disease: A electrophysiologic study. Hepatology29:1698–1793, 1999

6. Oliver MI, Miralles R, Rubies-Prat J, Navarro X, Espadaler JM, SolaR, Andreu M: Autonomic dysfunction in patients with non-alcoholicchronic liver disease. J Hepatol 26:1242–1248, 1997

7. Gonzalez-Reimers E, Alonso-Socas M, Santolaria-Fernandez F,Hernandez-Pena J, Conde-Martel A, Rodriguez-Moreno F: Auto-nomic and peripheral neuropathy in chronic alcoholic liver disease.Drug Alcohol Depend 27:219–222, 1991

8. Puthumana L, Chaudhary V, Thuluvath PJ: Prolonged QTc intervaland its relationship to autonomic cardiovascular reflexes in patientswith cirrhosis. J Hepatol 35:733–738, 2001

9. Bal JS, Assouad F, Thuluvath PJ: Prolongation of QTc interval:Relationship with aetiology and severity of liver disease, mor-tality and liver transplantation. Liver International 23:243–248,2003

10. Galati JS, Holdeman KP, Bottjen PL, Quigley EM: Gastric emptyingand oro-cecal transit in portal hypertension and end-stage chronicliver disease. Liver Transpl Surg 3:34–38, 1997

11. Van Thiel DH, Fagiuoli S, Wright HI, Chien MC, Gavaler JS:Gastrointestinal transit in cirrhotic patients: Effect of hepatic en-cephalopathy and its treatment. Hepatology 19:67–71, 1994

12. Wegener M, Scheaffstein J, Dilger U, Coenen C, Wedmann B,Schmidt G: Gastrointestinal transit of solid-liquid meal in chronicalcoholics. Dig Dis Sci 36:917–923, 1991

13. Madrid AM, Cumsille F, Defilippi C: Altered small bowel motilityin patients with liver cirrhosis depends on severity of liver disease.Dig Dis Sci 42:738–742, 1997

14. Morencos FC, Castano DLH, Ramos LM, Arias MJL, LedesmaF, Romero FP: Small bowel bacterial overgrowth in patients withalcoholic cirrhosis. Dig Dis Sci 40:1252–1256, 1995

15. Bauer TM, Schwacha H, Steinbruckner B,et al.: Small intestinalbacterial overgrowth in human cirrhosis is associated with systemicendotoxemia. Am J Gastro 97:2364–2370, 2002

16. Chang CS, Chen GH, Lein HC, Yeh HZ: Small intestine dysmotil-ity and bacterial overgrowth in cirrhotic patients with spontaneousbacterial peritonitis. Hepatology 28:1187–1190, 1998

17. Vinik AL, Maser RE, Mitchell BD, Freeman R: Diabetic autonomicneuropathy. Diabetes Care 26:1553–1579, 2003

18. Zietz B, Lock G, Straub RH, Braun B, Scholmerich J, Palitzsch KD:Small bowel bacterial overgrowth in diabetic subjects is associatedwith cardiovascular autonomic neuropathy. Diabetes Care 23:1200–1201, 2000

19. Ewing DJ, Clarke BF: Diagnosis and management of diabetic auto-nomic neuropathy. Br Med J 285:916–918, 1982

20. Pardo A, Bartoli R, Lorenzo-Zuniga V,et al.: Effect of cisapride onintestinal bacterial overgrowth and bacterial translocation in cirrho-sis. Hepatology 31:858–863, 2000

21. Wang XD, Soltesz V, Andersson R: Cisapride prevents enteric bac-terial overgrowth and translocation by improvement of intestinalmotility in rats with acute liver failure. Eur Surg Res 28(6):402–412, 1996

1588 Digestive Diseases and Sciences, Vol. 49, No. 10 (October 2004)