Methods

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Patients

We evaluated 14 persons, including 7 patients (3 men, 4 women) 43 to 63 years of age (mean age, 53.4 years) with biopsy-proven cirrhosis of the liver and evidence of portal hypertension (1 had hepatitis B, 3 had hepatitis C, 1 had primary biliary cirrhosis, and 2 had autoimmune disease). Patients were classified as having Child A (n = 2), B (n = 4), or C (n = 1) liver disease. Exclusion criteria were defined on the basis of factors that could affect melatonin levels and included therapy with ß-blockers or corticosteroids, current therapy for hepatic encephalopathy, diabetes mellitus, alcoholic liver disease, nightshift work, or recent intercontinental travel. Controls (n = 7) were healthy volunteers who were not receiving any medication at the time of the study and who were matched to the study patients by age, sex, and educational level. Each control was studied within 1 month of his or her matched patient to exclude seasonal variability. The protocol was approved by the Institutional Review Board of Northwestern University.

Overall Protocol Design

One week before admission (study day 1), neuropsychological testing was done, medical history was taken, and blood was drawn for baseline laboratory values. At this time, patients were instructed to keep a sleep diary for 1 week, recording the times they went to bed, the times they woke up, activities, naps, meals, and awakenings during the night. During the week, patients continued their normal daily activities. Dietary recommendations formulated by a dietitian included the ingestion of 1 g of protein/kg body weight per day for 1 week before neuropsychological testing and during the entire evaluation period.

Patients were admitted to the Clinical Research Center at Northwestern Memorial Hospital at 9:00 a.m. on study day 8 so that they could become acclimated to the new environment. They received three meals and a bedtime snack daily; only noncaffeinated beverages were allowed. On study day 9, patients were asked to remain recumbent throughout the day, and naps were prevented. Beginning at 8:00 a.m. and for 24 hours thereafter, 3 mL of blood were drawn from each patient every 30 minutes through an indwelling catheter. During the night, a plastic tube was extended outside the sleeping compartment. Plasma was stored at –70°C until further analysis. The room light was kept at less than 300 lux and was turned off at 11:00 p.m., at which time the patients were allowed to sleep.

Neuropsychological Testing

Visual reaction time was measured as the time to response, in milliseconds, to a visual stimulus appearing on a computer screen. In Trail Making Test A and B, the time needed to connect numbered or lettered circles on a sheet of paper was recorded [2]. The Digit symbol test is a subtest of the Wechsler Adult Intelligence Scale-Revised [3].

Hormonal Assay

In each sample, plasma melatonin levels were measured in duplicate with a commercially available radioimmunoassay [4] that uses (Iodine-125)melatonin (Elias USA, Inc., Osceola, Wisconsin) with a detection limit of 1.5 pg/mL. The intra-assay coefficient of variation was 1.8%. The time of onset of an increase in melatonin level was defined as the time point in the evening at which levels had increased at least threefold from the previous sample. The time of peak melatonin levels was defined as the time after which values consistently decreased.

Statistical Analysis

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Patients and controls were compared using the Student paired t-test with a level of significance set at P < 0.05. When appropriate, the Wilcoxon signed-rank test was used. Values are expressed as mean ±SE.

Results

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Neuropsychological Testing

Our patients met criteria that may indicate the presence of subclinical hepatic encephalopathy [5]. Prolonged visual reaction time was seen in patients with cirrhosis during the third set of observations (391 ±23 ms for patients and 332 ±19 ms for controls), and these patients showed a significant prolongation of Trail B (93 ±14 s for patients and 54 ±8 s for controls).

Melatonin Levels

Three variables of the melatonin profile were analyzed [Figures 1 and 2]. Time of onset of melatonin secretion Figure 1, top) was significantly displaced from 7:50 p.m. ±26 min to 9:30 p.m. ±13 min (P = 0.013). The melatonin profile of one patient with cirrhosis did not show a diurnal variation, but absolute values were markedly increased; plateau levels were three times higher than the SE of the group mean time. This patient with Child C liver disease was not included in the calculation. The matched control had onset of increase at 7:30 p.m.

View larger version (14K): [in this window] [in a new window]   Figure 1. Time of onset of melatonin secretion and time of peak plasma melatonin levels. Top. Time of onset of melatonin secretion in patients (n = 6) was significantly displaced from 7:50 p.m. ± 26 min to 9:30 p.m. ± 13 min. Bottom. Time of peak plasma melatonin levels (n = 6) was significantly displaced from 12:36 a.m. ± 33 min to 5:36 a.m. ± 29 min. Horizontal bar represents the mean. * indicates P = 0.013; ** indicates P < 0.001.

View larger version (20K): [in this window] [in a new window]   Figure 2. Plasma melatonin levels (mean ±SE) expressed in pg/mL and measured at 30-minute intervals for 24 hours. Vertical dashed line indicates midnight. Levels were significantly higher in patients than in controls at every measurement point between 2:30 a.m. and 10:00 a.m. (P < 0.05).

Time of peak level Figure 1, bottom) was also significantly displaced from 12:36 a.m. ±33 min to 5:36 a.m. ±29 min (P < 0.001). The matched control had a time of peak level at 11:00 p.m. As seen in Figure 2, significant increases in absolute melatonin levels were seen during daytime and nighttime hours (P < 0.05 at every measurement between 2:30 a.m. and 10:00 a.m.).

Sleep Diaries

Analysis of the sleep diaries on study days 2 to 7 showed no significant difference in average bedtime (11:27 p.m. for patients compared with 11:06 p.m. for controls). The greater number of daytime naps taken by patients did not differ statistically from the number taken by controls (3.2 ±1.1 compared with 1 ±0.6; P = 0.075), but significantly more awakenings occurred during the night in patients with cirrhosis than in controls (4.5 ±1.4 compared with 0.2 ±0.2; P = 0.05) Interestingly, the patient who had no diurnal rhythmicity in the melatonin profile and who had elevated melatonin levels at all times also had the most naps and the longest duration of naps during the daytime (n = 8), as well as frequent awakenings during the night.

Discussion

Top Methods Statistical Analysis Results Discussion Author & Article Info References

We have shown that patients with cirrhosis and subclinical hepatic encephalopathy, studied in a strictly controlled environment, have an abnormal plasma melatonin pattern compared with healthy controls. The onset of the increase in plasma melatonin levels and the melatonin peak during the night were both displaced to later hours. Furthermore, plasma melatonin levels in patients with cirrhosis were significantly increased during daylight hours at a time when melatonin levels are normally low or absent.

To our knowledge, one previous study [6] has described elevated melatonin levels in patients with liver cirrhosis. Measuring at intervals of 4 to 6 hours, investigators found elevated serum melatonin levels in 22 patients with cirrhosis during the daytime and nighttime hours. Because measurements were done only at 10:00 p.m. and 4:00 p.m., no rigorous attempt was made to delineate onset or time of peak melatonin levels, which are critical measures of circadian clock output [7]. Because melatonin is eliminated by the liver, the authors suggested that a reduction in hepatic melatonin clearance explained the elevated daytime melatonin levels.

By measuring values every 30 minutes, we found a consistently delayed onset and time of peak melatonin levels in patients with cirrhosis. Because these variables may reflect circadian phase [8], our data suggest the existence of an alteration in the output of the suprachiasmatic nucleus, the "biological" clock. Two mechanisms could account for our findings.

First, the same metabolic disturbances that lead to hepatic encephalopathy could affect circadian function. In rats that have had portacaval anastomosis [9, 10], which is a model of complete portal-systemic shunting, the alterations in the day:night activity ratio can be improved by therapy with oral neomycin [11], a drug that lowers ammonia production in the gut and thus is used to treat hepatic encephalopathy. Ammonia, in addition to exerting potential toxic effects on the pineal gland [12], results in an increased entry of tryptophan into the brain, with alterations of serotonin (precursor of melatonin) and associated neurotransmission [13]. Other toxins, acting through peripheral benzodiazepine receptors, may have direct effects on the pineal gland [12], which is the site of melatonin synthesis.

Second, Lewy and colleagues [14] found that phase advances or phase delays occurred according to the time at which melatonin was administered to healthy volunteers. Phase delays could only be seen when the drug was given between 3:00 a.m. and 12:00 noon. In our study, this was a period during which melatonin levels in patients with cirrhosis were considerably elevated. Thus, high endogenous melatonin levels secondary to decreased hepatic elimination could alone cause a phase delay in the same way that exogenous melatonin does. Evidence from studies of human jet lag [15] and studies in blind patients [16, 17] also support the notion that melatonin has a phase-synchronizing effect.

Although analysis of the sleep diaries showed disrupted night sleep in patients with cirrhosis, we found no clear correlation between the sleep pattern and the melatonin rhythm. The literature on the relation between melatonin and sleep is controversial [18-20]. The subjective quality of sleep, however, was worse in patients than in controls, again indicating the clinical relevance of the problem. Further studies should examine other measures of circadian function. The effect of treatment of hepatic encephalopathy on the melatonin profile would also provide valuable information about the pathogenesis of this abnormality.