Serum Catecholamines and Dysautonomia in Diabetic Gastroparesis and Liver Cirrhosis

Abstract

Background

Plasma catecholamine influences autonomic function and control, but there are few reports correlating them. In this study, 47 individuals (mean age, 38 years) were studied: 19 diabetes mellitus (DM) patients with gastroparesis, 16 with liver disease and 12 control subjects.

Methods

Noninvasive autonomic function was assessed for sympathetic adrenergic functions as peripheral vasoconstriction in response to cold stress test and postural adjustment ratio (PAR) and cholinergic function as Valsalva ratio, represented by change in R-R intervals. Measurements were compared by analysis of variance and Spearman’s correlation, and results were reported as mean ± standard error.

Results

Plasma norepinephrine (1902.7 ± 263.3; P = 0.001) and epinephrine (224.5 ± 66.5; P = 0.008) levels, as well as plasma dopamine levels (861.3 ± 381.7), and total plasma catecholamine levels were highest for patients with liver disease, who also had significant negative correlation between norepinephrine level and vasoconstriction (P = 0.01; r = −0.5), PAR1 (P = 0.01; r = −0.5), sympathetic adrenergic functions (P = 0.005; r = −0.6), total autonomic index (P = 0.01–0.5) and total autonomic function (P = 0.01; r = −0.2) and also negative correlation between epinephrine plasma level and total autonomic function (P = 0.04; r = 0.4). DM patients were next highest in norepinephrine level (133.26 ± 7.43), but lowest for plasma catecholamine; a positive correlation between dopamine level and PAR1 (P = 0.008; r = 0.6) was also seen in this group. Plasma dopamine levels and spider score correlated negatively (P = 0.04; r = −0.5) and total plasma catecholamine positively with encephalopathy (P = 0.04; r = 0.5) in patients with liver disease.

Conclusions

Plasma catecholamine levels correlated with adrenergic functions in control subjects and patients with DM and liver disease, with no significant correlation seen for cholinergic function.

In the autonomic nervous system (ANS), the adrenergic and cholinergic systems work in harmony in maintaining health; however, this ANS may become disordered in a variety of disease states. Diabetes mellitus (DM) and liver cirrhosis provide particularly interesting insight into system disharmony because they show inconsistent alterations in serum catecholamine that may trigger autonomic responses.^1–4

The main sources of plasma catecholamine are the adrenal glands, which serve as the major supplier of epinephrine and norepinephrine, and the ANS, the major supplier of dopamine. Abnormalities in autonomic function in patients with DM^4–6 or liver disease⁷ and the abnormal catecholamine levels⁸ seen in these 2 groups, as compared with those in normal healthy individuals, raise several questions about the cause of dysautonomia in these 2 patient groups.

Heart rate (HR) regulation is mainly controlled by the cholinergic system, while the vascular response to stress, postural and temperature changes are mainly controlled by the adrenergic system. The ANS regulation of the HR, blood pressure, vascular response to stress, postural changes and temperature changes is mediated by the direct effect of the postganglionic nerves on the targeted receptors through neurohormones.

In this study, the authors aimed to correlate total plasma catecholamine levels, including epinephrine, norepinephrine and dopamine, with the autonomic functions measured by the noninvasive autonomic function test (AFT) in 3 groups of individuals: normal healthy subjects as a control group and patients with a detailed history of either DM or liver cirrhosis. They also aimed to analyze the correlation of serum catecholamine and autonomic dysfunction with clinical status in patients with DM and liver cirrhosis.

METHODS

Patients

Forty-seven individuals (mean age, 38 years; 18 men and 29 women) were studied. Individuals were stratified into normal healthy control (12 of 47) subjects, patients with DM (19 of 47) with the clinical evidence of autonomic failure often presenting with the symptoms of gastroparesis and patients with liver cirrhosis (16 of 47). The DM and liver cirrhosis groups were accrued from consecutive patients referred to an autonomic clinic in a tertiary/transplant hospital, who agreed to having catecholamine levels drawn in addition to having autonomic testing performed. For patients with DM, many of who were being evaluated for pancreas transplant, the mean HgbA1c was 8.7, although these results were for 10 of 19 patients. The patients with liver cirrhosis, many awaiting possible liver transplant, were subclassified based on Child-Turcotte-Pugh (Child) scores as A (3 of 16) patients, B (10 of 16) patients and C (3 of 16) patients (Table 1). Of the 16 patients with liver cirrhosis, 7 had clinically significant hepatic encephalopathy and 5 had ascites. Of these 16 patients, 13 had alcoholic cirrhosis, 2 had cirrhosis secondary to viral hepatitis and 1 had primary biliary cirrhosis, and all were also assessed for a history of encephalopathy or the presence of ascites or cutaneous spiders on physical examination. Routine screening of cardiac history eliminated those patients with recent acute ischemic events. The control subjects were volunteers responding to a notice placed in the hospital and were all healthy.

TABLE 1.

Patient characteristics

	Control	DM	Cirrhosis
Mean age, yr	30 (20–49)	39 (23–66)	43 (21–60)
Mean HgbA1c		8.7 (5.7–12.7)
Mean albumin			3 (1.5–3.7)
Mean total bilirubin			9.3 (0.9–34.8)
Mean prothrombin   time			4 (0.1–12.5)
Mean creatinine			1.7 (0.5–3.7)

Methods

After reviewing the medical history, 47 individuals were enrolled in this study. Each individual was instructed not to take any medication(s) for at least 24 hours before the autonomic test. No alcohol, cigarettes or any caffeine products were allowed before the evaluation. The subject was instructed to wear comfortable clothes and abstain from all forms of excessive activities on the day of evaluation. Each participating subject was measured with a standard blood pressure and HR monitor using a Dinamap (model 8100; Critikon, Inc, Tampa, FL), a thermometer (Mon-a-Therm, Inc, St Louis, MO) secured with a tape on the tip of the middle finger on the right hand and an infrared light–emitting diode probe connected to a MedaSonics photoplethysmograph (model PPG-13; MedaSonics, Mountain View, CA) secured with a nonallergic double stick-tape to maintain good skin contact and prevent any additional pressure on the index finger tip of the left hand. The blood flow probe was captured as a waveform recorded on polygraph (model 79D1E; Grass Medical Instruments, Warwick, RI). The change in amplitude and frequency of the formed waves at 1-minute intervals, in response to cold stress and arm posture, represented the blood vessel capacity. No patients showed clinical symptoms of dehydration when studied. The AFT was conducted in the same manner after the initial skin temperature and baseline measurements showed stability for at least 10 to 20 minutes of relaxation in a quiet room with a warm temperature between 22 and 26°C. With subject in a supine position, he/she was instructed to breathe deeply in and out for approximately 6 cycles per minute. During deep inhalation, maximum HR was recorded, and during deep expiration, the lowest HR was also recorded. Subject then was instructed to blow continuously through a mouthpiece connected to a 1-valve tube and a blood pressure monitor, for 10 to 15 seconds with a pressure of 40 mm Hg. Maximum and minimum HRs were recorded.

The individual then assumed a sitting position with left hand and arm in a 45° angle above the heart level rested on a comfortable chair armrest for approximately 5 minutes. After recording a stable peripheral blood flow, the subject was asked to drop his/her arm down freely next to the side of the chair without touching or holding anything for 1 minute. The change in blood vessel capacity was then calculated. Subject was asked to return the hand and arm back to the original position above the heart level and another measurement for blood flow was calculated. Maximum blood flow in 1 minute was recorded and used as a baseline flow at rest. The individual was then instructed to immerse the right hand in ice-cold water for 1 minute, and again, the blood flow on the opposite hand was calculated as before.

The values which were recorded from this standard noninvasive AFT evaluation were used to calculate cardiovagal arch function and adrenergic arch function. While the cardiovagal arch function was represented by the change in HR to either deep or forced respiration, the adrenergic arch function was assessed through the change in peripheral cutaneous blood flow in response to posture and cold stress testing.

Two measurements of vagal cholinergic function were performed through minute-to-minute blood pressure and HR monitoring for HR: the differences in HR in response to deep respiration and forced respiration (Valsalva maneuver) were tabulated. The sum of the percent of change in HR with deep respiration (lowest HR with expiration subtracted from maximum HR during inspiration divided by the lowest HR) is represented by R-R intervals and the ratio of maximum (phase II) to minimum (phase IV) HR ratio during Valsalva (Valsalva ratio).

Two measurements of sympathetic adrenergic function were performed through the changes in peripheral circulation each by using capillary photoplethysmography.¹⁰ Postural adjustment ratio (PAR) was calculated by dividing the capacity of blood vessel with the arm in down position by the value with the hand and arm in a 45° above the heart level and percent vasoconstriction (%VC), in response to 1-minute hand immersion in ice-cold water with a temperature of 14°C, while rate of blood flow was monitored through the opposite hand and calculated as follows: 100 − capacity of blood vessel during cold stress/capacity of blood vessel at rest × 100. %VC is a measure of the change in capillary pulse amplitude caused by reflex vasoconstriction and is expressed as a percentage of change from baseline. Blood vessel capacity during rest was represented by total pulse amplitude “hand” and measured from the amplitude and frequency of wave recorded on the polygraph before the arm movements or the opposite hand’s immersion in ice-cold water.

Total autonomic score was calculated as the sum of % VC, PAR and R-R intervals. Total autonomic index and total autonomic function (TAF), calculated in a similar manner, were also tabulated.

All 3 individual groups performed AFT during the daytime. Patients with liver disease also underwent a clinical evaluation for Child scoring, Maddrey scoring and symptoms/signs scoring, including cutaneous evidence of cirrhosis and plus additional laboratory studies (sodium, potassium, albumin, bilirubin, alkaline phosphatase, prothrombin time, gamma glutamine transpeptidase). Measurements were compared by analysis of variance and Spearman’s correlation to compare catecholamine levels either between or within the 3 groups, and results were reported as mean ± standard error. This study was approved by the Institutional Review Board of the University of Tennessee-Memphis.

RESULTS

Catecholamine Levels by Group

Individuals were stratified as follows: control group (Cont), patients with DM and patients with liver cirrhosis (Liver).

Analysis of variance showed that plasma norepinephrine level was highest in the Liver group followed by DM and Cont groups (1902.7 ± 263.3, 133.26 ± 7.43 and 110.7 ± 23.3, respectively; P = 0.001). Plasma epinephrine levels showed a similar pattern; the Liver group was again the highest followed by DM and Cont groups (224.5 ± 66.5, 37.8 ± 3.2 and 2 ± 0.6, respectively; P = 0.008). Plasma dopamine levels were higher in the Cont group (155.7 ± 37.2) than in the DM group (55.6 ± 4.8) but were still highest in the Liver group (861.3 ± 381.7) (P = 0.026). Total plasma catecholamine levels were highest in the Liver group (2960.7 ± 502.6) than in the control (271.2 ± 56.6) or DM (226.6 ± 10) group (P < 0.005) (Table 2 and Figure 1).

TABLE 2.

Plasma catecholamines

	Group	Mean	Standard error	P
Norepinephrine (supine)   (ANOVA)	Control	110.667	23.31	0.001
	DM	133.26	7.43
	Liver	1902.75	263.26
Epinephrine (supine)   (ANOVA)	Control	2.00	0.57	0.008
	DM	37.79	3.17
	Liver	224.5	66.48
Dopamine (supine)   (ANOVA)	Control	155.67	37.23	0.026
	DM	55.58	4.78
	Liver	861.3	381.73
Total catecholamine	Control	271.2	56.6	<0.005
	DM	226.6	10
	Liver	2960.7	502.6

ANOVA, analysis of variance; DM, diabetes mellitus.

FIGURE 1.

Serum norepinephrine, epinephrine and dopamine level in the 3 groups (controls, diabetes mellitus [DM] and liver disease) in supine position.

Autonomic Function Test Results by Group

AFT measures showed that vagal cholinergic function was below normal in DM group (14.22 ± 1.7) compared with Cont (35.7 ± 6.55), while Liver group showed borderline low-normal (16.7 ± 2.3) (P = 0.001). Total pulse amplitude “hand” in both DM and Liver groups (Figure 3) (4853.07 ± 885.2 and 4066.94 ± 772.18, respectively) was lower than that in the Cont group (7198.3 ± 627.3) (P = 0.03). The same pattern was also noticed with both %VC0020 and PAR (Figure 2). Percent of vasoconstriction “hand” for DM and Liver groups was 53.15 ± 7.6 and 66.7 ± 5.9, respectively, and 92.18 ± 1.2 for Cont group (P = 0.001). It is worth noting that DM group average range was below normal (>65%), while Liver group was borderline low-normal for %VC. PAR was 26.75 ± 6.6 for the DM group and 13.5 ± 2.1 for the Liver group and both were below the Cont group 32.03 ± 4.8 (P = 0.03). The Liver group was below normal range (>16) (Table 3 and Figure 4). Finally, the systolic blood pressure was not clinically different between the 3 groups: 112.67 ± 1.7 for Controls, 135.6 ± 6.6 for DM patients with clinical evidence of autonomic failure and 112.4 ± 2.5 for the liver disease group (Figure 5).

FIGURE 3.

Vagal cholinergic function (VCF) (R-R intervals and Valsalva ratio) in controls, diabetes mellitus (DM) and liver disease.

FIGURE 2.

Total pulse amplitude (TPA) in controls, diabetes mellitus and liver disease.

TABLE 3.

Autonomic function tests in 3 groups

	Group	Mean	Standard error	P
Vagal cholinergic   function (ANOVA)	Control	35.65	6.55	0.001
	DM	14.22	1.7
	Liver	16.65	2.33
TPA hand (ANOVA)	Control	7198	627.28	0.026
	DM	4853.07	885.2
	Liver	4066.94	772.18
%VC hand	Control	92.18	1.18	0.001
	DM	53.15	7.63
	Liver	66.67	5.9
PAR (ANOVA)	Control	32.03	4.83	0.03
	DM	26.75	6.61
	Liver	13.53	2.1

ANOVA, analysis of variance; DM, diabetes mellitus; TPA, total pulse amplitude; %VC, percent vasoconstriction; PAR, postural adjustment ratio.

FIGURE 4.

Sympathetic adrenergic function (%VC and PAR) in controls, diabetes mellitus (DM) and liver disease.

FIGURE 5.

Systolic blood pressure in controls, diabetes mellitus (DM) and liver disease.

Correlation of Catecholamine Levels With Autonomic Function Measures by Groups

By Spearman’s correlation analysis, in the control group, plasma catecholamine levels showed a negative correlation between dopamine level and VC (P = 0.037; r = −0.53), and total catecholamine level correlates with PAR (P = 0.06; r = 0.5).

In DM patients, there was a positive correlation between dopamine level and PAR (P = 0.008; r = 0.6). Total plasma catecholamine levels also correlated with PAR1 (P = 0.01; r = 0.6).

The liver patients showed a significantly negative correlation between norepinephrine level and VC (P = 0.01; r = −0.5), PAR1 (P = 0.01; r = −0.5), sympathetic adrenergic function (P = 0.005; r = −0.6), total autonomic index (P = 0.01–0.5), TAF (P = 0.01; r = −0.2), Na⁺ (P = 009; r = −0.6) and albumin (P = 0.01; r= −0.6) and positive correlation with ascites score (P = 0.01; r = 0.5), encephalopathy score (P = 0.03; r = 0.5), Maddery score (P = 0.01; r = 0.5) and Child score (P = 0.002; r = 0.7). Also, in liver patients, there was a negative correlation between epinephrine plasma level and TAF (P = 0.04; r = 0.4) and a negative correlation with plasma alkaline phosphatase (P = 0.008; r = − 0.6). Plasma dopamine level correlates negatively with spider score (P = 0.04; r = −0.5), while total plasma catecholamine correlates positively with encephalopathy score (P = 0.04; r = 0.5).

DISCUSSION

Autonomic dysfunction is seen in many medical illnesses like DM, cerebrovascular disease, cirrhosis, spinal cord disease and end-stage renal disease on hemodialysis.¹¹ In this study, patients with DM and cirrhosis compared their autonomic dysfunction and serum catecholamines with normal individuals.

Several studies have shown evidence of autonomic dysfunction in cirrhotic patients.¹² Patients with cirrhosis demonstrate reduced blood pressure despite increased HR and increased cardiac output, which point to autonomic dysfunction.^13,14 Vagal neuropathy is seen in patients with alcohol-induced cirrhosis.¹⁵ Previous work by other investigators has shown increased plasma catecholamines in patients with alcohol-induced liver cirrhosis.¹⁶ The possible causes of this increase in concentration of plasma catecholamine manifest as an increase in synthesis, a reduction in reuptake and impaired metabolism or reduced clearance. Hendrickse et al¹⁷ reported vagal neuropathy in 45% of the 60 patients with chronic liver disease. Gentile et al¹⁸ reported autonomic neuropathy in 60% (71% in the alcoholic group and 57% in the nonalcoholic group) of the 113 cirrhosis studied. Some studies considered the possible role of the significant metabolic disorder of epinephrine and norepinephrine in the pathogenesis of portal hypertension and in the development of hepatic encephalopathy.^8,19 Carey et al found that in patients with end-stage liver disease, more than 86% of study participants had abnormal autonomic testing using a 10-point composite autonomic score. Of note, 63% of patients with cirrhosis and with autonomic dysfunction showed improvement after liver transplantation.²⁰ The authors found that the patients with liver disease showed a significantly negative correlation between plasma norepinephrine levels and albumin and positive correlation with ascites score, encephalopathy score, Maddery score and Child score. In patients with liver disease, there was also a negative correlation between plasma epinephrine level and TAF and with plasma alkaline phosphatase. Plasma dopamine level correlates negatively with spider score, while total plasma catecholamine correlates positively with encephalopathy score.

A variety of alterations have been described in the nervous system of diabetic animals,²¹ although different studies have shown variable results. One study showed decreased activity of the nigrostriatal dopaminergic system. Changes found at the sympathoadrenal level could be explained by the reduced norepinephrine and epinephrine synthesis and/or increased storage as a result of a release from synaptic vesicle.²² Considerable debate exists regarding interactions of congestive heart failure, DM and the ANS. Diabetes can cause myocyte hypertrophy, interstitial fibrosis, impaired myocardial blood flow and increased turnover of free fatty acids, leading to the development of cardiomyopathy. However, heart failure increases catecholamine plasma levels, and it may cause insulin resistance, leading to the development of diabetes.²³

Another study showed a comparatively higher level of circulating norepinephrine in painful than in painless diabetic neuropathy.²⁴ Painful neuropathy is associated with a relatively higher number of functioning sympathetic fibers that may contribute to pain.²⁴ Norepinephrine basal levels were found to be lower in patients with severe neuropathy. Norepinephrine basal level and cold responses are diminished in patients with definite and severe autonomic neuropathy.²⁴ These data provide further evidence of the patient groups’ impaired response to stress.

However, the comparable epinephrine levels in patients with or without autonomic neuropathy indicate that adrenal medullary function may not be significantly altered.

Plasma noradrenaline and adrenaline levels reflect sympathetic nervous activity and are based on the following suppositions. Norepinephrine is a main neurotransmitter in the sympathetic nervous system and is released from the sympathetic nerve terminals as a consequence of the activation of the sympathetic neuron.

Epinephrine is released from the sympathetic medulla upon activation of the gland (the adrenal medulla is the equivalent of the postganglionic sympathetic neuron). The application of plasma norepinephrine and epinephrine measurements as a marker of sympathetic activity is based on the relationships between spillover rates from sympathetic nerve terminals into the circulation. It is assumed that a constant fraction escapes the reuptake mechanism in the terminal parts of the nerve, which means that an increased activity in the nerve will release a proportional amount of norepinephrine and that a proportional amount will escape to the circulation.²⁵

In DM patients, there was a positive correlation between dopamine level and PAR, and total plasma catecholamine levels also correlated with PAR. In this study, plasma norepinephrine levels in the DM group were higher than those in the control group and were highest in liver patients. The plasma epinephrine levels showed similar results. Conversely, plasma dopamine levels showed slightly different results with the DM group being lower than the control group and the liver patients’ levels being the highest. The total plasma catecholamine levels were lower in the DM group compared with the control group, and liver patients’ levels were the highest. As summarized in Figure 1, plasma catecholamine levels were lowest and within the normal range for the control group and higher (abnormal range) in both the DM and liver groups (Figures 2–5).

The present data suggest that catecholamine accumulation, which is later accompanied by an impairment in catecholamine secretion, as seen in diabetic rats, gives a basis for an inference that similar changes may play a role in the pathogenesis of diabetic autonomic neuropathy in man.²⁶ There is evidence to suggest that there are some disturbances in catecholamine secretion in diabetic patients’ hearts before more typical microangiopathic changes occur²⁷; however, it remains to be seen if these changes are secondary to the changes in plasma catecholamine levels in the serum leading to secondary changes in the local hormonal milieu of the cardiac myocytes. Abnormalities noted in this study are similar to those found in congestive heart failure. In addition, diabetic patients presenting with hyperadrenergic orthostatic hypotension have an initial stage of autonomic neuropathy, with overtly abnormal vagal function and early signs of sympathetic impairment.²⁸

CONCLUSIONS

Plasma catecholamine levels correlate with physiologic adrenergic functions in control subjects and in patients with DM and liver disease, while no significant correlation was found with cholinergic functions. In addition, plasma catecholamines correlate with clinical scoring in patients with liver disease. These findings may explain abnormalities in the ANS seen in patients with diabetes and liver disease with signs of organ failure via abnormal catecholamine levels secondary to their primary disease. The noninvasive methods demonstrated in this study might have applicability as surrogate measures of catecholamines and thus may be of use in clinical care.