- Email: [email protected]
Cirrhosis does not shift the circadian phase of plasma fibrinolysis
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2002 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.
Vol. 97, No. 6, 2002 ISSN 0002-9270/02/$22.00 PII S0002-9270(02)04155-2
Cirrhosis Does Not Shift the Circadian Phase of Plasma Fibrinolysis Fabio Piscaglia, M.D., Ph.D., Ramon C. Hermida, Ph.D., Sebastiano Siringo, M.D., Cristina Legnani, Ph.D., Giuliano Ramadori, and Luigi Bolondi, M.D. Division of Internal Medicine, Department of Internal Medicine and Gastroenterology, and Department of Angiology and Blood Coagulation, University of Bologna, Bologna, Italy; Bioengineering and Chronobiology Laboratories, ETSI Telecomunicacio`n, University of Vigo, Vigo, Spain; and Department of Internal Medicine, University of Go¨ttingen, Go¨ttingen, Germany
OBJECTIVES: The aim of the present investigation was to study the endogenous circadian clock phase rhythm in cirrhotic patients. METHODS: The study population comprised 13 patients with cirrhosis (seven in Child-Pugh class A and six in classes B/C) and nine healthy controls. Plasma melatonin, tissue plasminogen activator antigen, and plasminogen activator inhibitor 1 antigen were measured at 4-h intervals over a 24-h period. Multiple-components rhythmometry using population mean cosinor methods were employed to analyze the findings. RESULTS: All three variables were characterized in both patients and controls by a statistically significant circadian rhythm, with similar profiles. The peak times of tissue plasminogen activator and plasminogen activator inhibitor 1 antigens were practically identical in controls and cirrhotic patients, irrespective of Child-Pugh class (calculated peak at times 6:52, 6:56, and 7:20 for the inhibitor in controls and Child-Pugh class A and classes B/C patients, respectively; p ⫽ ns), whereas the peak of melatonin was delayed in classes B/C patients (at times 2:08, 1:56, and 4:00, respectively; p ⬍ 0.05). CONCLUSION: The similar circadian phases of plasminogen activator inhibitor antigen in controls and cirrhotic patients in the present investigation indicates that the output rhythm of the internal timekeeping system is not shifted in this pathological condition. (Am J Gastroenterol 2002;97: 1512–1517. © 2002 by Am. Coll. of Gastroenterology)
INTRODUCTION Although sleep disturbance is considered a classic sign of hepatic encephalopathy, it can frequently be revealed also in cirrhotic patients without encephalopathy (1). Abnormalities in the circadian timekeeping system have been suggested to be responsible for this disturbance. Indeed, higher S.S. is currently affiliated with the Division of Internal Medicine, Ospedale Garibaldi, Catania, Italy.
daytime levels of melatonin, considered as an indicator of the internal clock, and a phase delay of the onset of rise in nocturnal plasma melatonin have been reported in cirrhosis (2– 4). The liver participates in the degradation of melatonin to its metabolites, which are then excreted into urine and bile. Hence, reduced hepatic function in cirrhosis may explain decreased urinary metabolite excretion and increased levels of plasma melatonin (3, 4). In experimental studies high plasma levels of melatonin are able to shift the circadian phase, either advancing it or delaying it according to the day time of melatonin administration, in a direction opposite to that induced by light exposure (5, 6). Unfortunately, however, it is difficult to predict the possible effect of a chronic increase of plasma melatonin, like that observed in cirrhosis. Consequently, because the delayed peak observed in human cirrhosis might reflect a primary disturbance of the internal clock (2) and the impaired hepatic metabolism of melatonin (3, 4), it is not yet clear whether the phase of the circadian timekeeping system is primarily deranged or secondarily or not at all (2, 3, 7). In fact, other than studies with melatonin, little work has been done to investigate circadian parameters in chronic liver disease (8), and the function of the internal clock remains unclear in cirrhosis. The internal clock is hypothesized to be located in the hypothalamus in humans, from where it acts on neural and endocrine pathways to regulate individual circadian rhythms (9). To date there is no way to directly assess this clock in humans; its output has therefore to be investigated through its effects on downstream pathways, of which plasma melatonin is one of the main. Unfortunately, most of these pathways are theoretically disturbed by the metabolic and hemodynamic alterations present in cirrhosis, making it particularly difficult to ascertain whether the circadian clock phase is shifted or not in this pathological condition, as discussed for melatonin. Among various possible substances or circulatory parameters known to present diurnal variations and hence possibly suitable for use in a comparison with melatonin to investigate circadian rhythms in humans, we chose to assess some
AJG – June, 2002
components of the fibrinolytic system. In fact, these possess a strong and well-defined rhythm, largely independent from environmental factors and disease conditions (10 –12), at variance with circulatory parameters, such as blood pressure and heart rate, which usually suffer influence by external conditions, particularly in cirrhosis (13). The exact connection of the fibrinolytic system with the internal clock is still unclear, as it is for other circadian indicators, but it and the others are linked to the light/dark cycle, as demonstrated by a certain phase derangement in shift workers or under experimental conditions (14, 15). Total plasma fibrinolytic activity has been known for several decades to follow a circadian rhythm, with the peak in late afternoon. About a dozen years ago, the role of individual fibrinolytic factors was established (16). In particular, it was demonstrated that total plasma fibrinolytic activity is mainly due to the activity of tissue plasminogen activator (tPA) and governed by the interaction of the plasma antigen levels of the latter with the activity of its main inhibitor, the plasminogen activator inhibitor 1 (PAI-1) (14, 16, 17). The activity of PAI-1, in turn, strictly follows PAI-1 antigen levels, which ultimately determine the circadian rhythm of spontaneous fibrinolytic activity in blood (12, 18). The aim of the present study was to investigate the circadian patterns of melatonin, tPA, and PAI antigens in a cohort of cirrhotic patients to establish whether the circadian output phase of the endogenous timekeeping system is preserved or modified in this pathological condition.
MATERIALS AND METHODS Patients The study was performed in 13 hospitalized patients (mean age ⫾ SD ⫽ 54.3 ⫾ 8.1 yr; 10 males and three females) with liver cirrhosis (Child-Pugh class: A [seven], B [five], and C [one]; 11 with a virus-related disease, hepatitis B virus or hepatitis C virus, and two with alcoholic cirrhosis) and in nine healthy subjects (mean age ⫽ 50.2 ⫾ 10.3 yr), recruited from the hospital staff and their relatives. Most subjects were initially enrolled for a separate study assessing the diurnal pattern of fibrinolytic activity in cirrhosis, and a few data for control subjects are also reported elsewhere (19). At the time of the study all subjects had abstained from alcohol for at least 1 month. All virus-related cirrhotic patients and most controls had been teetotalers for a long time. The two alcoholic patients had abstained from alcohol for at least 1 yr. The remaining control subjects had alcohol intakes below 20 g/day. Previous or present signs of hepatic encephalopathy (minimal or mild) were shown in two out of six (33%) cirrhotic patients in Child-Pugh classes B/C, whereas no such signs were present in patients in class A. All gave their informed consent to participate in the study, whose design was in accordance with the principles of the Declaration of Helsinki and was approved by the local ethical committee.
Circadian Clock in Cirrhosis
1513
Study Protocol Starting at 8:00 AM of the day of the study and continuing until 8:00 AM the next day, blood withdrawals were repeated at intervals of 4 h. The subjects lay supine for at least 15 min before each blood withdrawal. Artificial illumination was permitted, if needed, only from 7:00 AM to 10:30 PM, whereas in the remaining period a very dim artificial light (less than 15 lux) was turned on only to allow correct venipunctures at 12:00 AM and 4:00 AM; otherwise patients remained in the dark. Subjects were asked to lie in their beds from 10:30 PM to 7:00 AM, whereas in the remaining period they were allowed to move freely within the hospital. Smoking and consumption of caffeine-containing substances were prohibited from midnight of the day preceding the study. Breakfast, lunch, and dinner were administered just after the measurements of 8:00 AM, 12:00 PM, and 8:00 PM; in the intervals subjects were permitted to drink water freely but not to consume any caloric-containing solid or liquid. Plasma Assays Blood samples were drawn from an antecubital vein with minimal stasis and clean venipuncture, using a 21-gauge butterfly needle. Separate venipunctures were performed at each time point. Venous blood was immediately separated into different aliquots: 1) one anticoagulated with trisodium citrate (0.129 mol/L, 1/10) for tPA measurement (tPA:Ag, TintElize tPA, Biopool, Umea, Sweden); 2) a second put in tubes (Diatube Htm; sodium citrate, citric acid, theophylline, adenosine and dipyridamole, dipyridamole [Diagnostica Stago, Asnie`res, France]) to avoid platelet activation and secretion of PAI and to correctly measure it (PAI:Ag, TintElize PAI-1, Biopool); and 3) a third put in tubes containing ethylenediaminetetraacetic acid (2 mg/ml) and aprotinin (400 U/ml) for melatonin assay (Melatonin-RIA, IBL, Hamburg, Germany). Blood samples were kept in melting ice until subjected to 20 min of centrifugation at 3500 rpm and 4°C. Platelet-poor plasma was distributed in coded plastic tubes, snap frozen in liquid nitrogen, and stored at ⫺20°C until assayed. All measurements were done in duplicate. Statistical Analysis Data were analyzed on a Power Macintosh G3 microcomputer using Chronolab, a software package for biological time series analysis by linear least-squares estimation (20). Circadian characteristics of each variable were established by population multiple-components analysis (21), a method designed for analysis of nonsinusoidal hybrid data (time series of data collected from a group of subjects, as in clinical situations involving patients). The method provides estimates of the rhythm-adjusted mean or midline estimating statistic of rhythm, defined as the average value of the rhythmic function fitted to the data, as well as the amplitude (defined as half the extent of the rhythmic change in a cycle approximated by the fitted curve) and acrophase (lag from a defined reference time point— here, local midnight— of the
1514
Piscaglia et al.
AJG – Vol. 97, No. 6, 2002
Table 1. Results of Multiple-Components Rhythmometric Analysis Group, No. of Subjects Melatonin Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups tPA:Ag Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups PAI:Ag Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups
MESOR
p for the Rhythm
34.8 pg/ml 221.9 pg/ml 187.7 pg/ml 261.8 pg/ml
ORTHO
95% Cls
AMPLI
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.002
2:08 2:36 1:36 4:00 ⬍0.001
1:28–2:40 1:28–3:32 0:40–2:40 3:36–4:24
142% 163% 197% 159%
3.5 ng/ml 11.8 ng/ml 6.3 ng/ml 18.1 ng/ml
⬍0.05 ⬍0.05 0.165, ns 0.331, ns
9:04 8:43 8:40 8:44 0.852, ns
8:00–9:56 7:24–11:56 7:18–11:47 7:33–12:08
17% 12% 14% 10%
10.1 ng/ml 14.1 ng/ml 7.6 ng/ml 21.5 ng/ml
⬍0.02 ⬍0.005 ⬍0.02 ⬍0.05
6:52 7:16 6:56 7:20 0.690, ns
5:16–8:16 6:16–8:12 5:32–9:24 6:16–8:28
37% 25% 33% 19%
Results of the circadian analysis in the various subgroups. AMPLI ⫽ amplitude, half the distance between the maximum and the minimum of the fitted curve; MESOR ⫽ midline estimating statistics of rhythm, average value of the best fitted curve with two harmonic components of 24 and 12 h; ORTHO ⫽ orthophase, lag from a defined reference time point (here local midnight) of the crest time in the curve fitted to the data; p among groups ⫽ p from the comparison of the times of orthophase in the three groups of controls and Child-Pugh A and B/C patients; p for the rhythm ⫽ p from testing the zero amplitude assumption for the circadian rhythm.
crest time in the curve fitted to the data) for every fitted component. When all fitted components are harmonics from a fundamental period, the method of multiple components also provides three additional parameters: the overall amplitude, defined as half the difference between the maximum and the minimum of the best fitted curve in one fundamental period; the orthophase, defined as the lag from a defined reference time point of the crest time, within a fundamental period, in the curve of multiple components fitted to the data (21); and the bathyphase, defined as the lag from a defined reference time point within a fundamental period, of the time of the lowest value in the curve of multiple components fitted to the data. The orthophase (similarly to the acrophase of each component) and the bathyphase are usually expressed as negative angular degrees, with 360° ⫽ 1 fundamental period, or, preferably, in time units (here, h:min, with 24 h used consistently as the fundamental period). p for the rejection of the zero amplitude assumption was determined, with rhythm detection considered statistically significant if p ⱕ 0.05. Circadian parameters were also compared between patients and controls with a nonparametric test developed to compare parameters obtained from population multiple-components analysis (22). Because the main objective of this study was to compare the temporal relationship of the studied variables between patients and controls, interindividual differences were reduced by expressing the data in percentage of each individual mean.
24 h and 12 h (Table 1 and Figs. 1 and 2). There was no statistically significant difference in the time of peak (orthophase) of melatonin between healthy subjects and patients with cirrhosis, when the latter were considered as one single group, being 2:08 and 2:36, respectively (p ⫽ 0.576) (Fig. 1, left). However, a statistically significant difference became evident (p ⬍ 0.001) (Table 1) when patients were grouped according to liver function and divided into the two subgroups of Child-Pugh classes A and B/C. Indeed, the circadian orthophase was delayed in classes B/C by 2 h 24 min, as compared to class A patients, (p ⫽ 0.004) (Fig. 1, right) and by 1 h 52 min, as compared to controls (p ⫽ 0.012) (Table 1). There was no significant difference between class A patients and controls (32 min) (p ⫽ 0.482). The calculated peak times of tPA:Ag and PAI:Ag in patients were practically identical to those in controls (differences of 24 and 20 min, respectively; ps ⫽ 0.735 and 0.786 for controls vs cirrhotics as one single group [Fig. 2] and 0.852 and 0.690 considering the three groups). However, unlike melatonin, the peak times were also comparable considering patients with mild and severe liver functional impairment separately: in fact, no intergroup difference could be found (orthophase at 8:44 and 8:40 [p ⫽ 0.999] for tPA:Ag in class A and B/C patients, respectively, and 6:56 and 7:20 [p ⫽ 0.558] for PAI:Ag, although the circadian curve did not reach statistical significance in the two latter smaller subgroups) (Table 1).
RESULTS
DISCUSSION
All three variables were characterized for both patients and controls by a statistically significant circadian rhythm that can be described by a model with two components, with periods of
The present findings indicate that the timekeeping system does not undergo significant phase shifts in cirrhosis, at least as far as fibrinolytic components are concerned.
AJG – June, 2002
Circadian Clock in Cirrhosis
1515
Figure 1. Circadian variation in melatonin in healthy subjects and patients with cirrhosis (left) and in cirrhotic patients with different degrees of disease severity (right). The curve represented for each group corresponds to the best fitted model obtained by population multiple-components analysis, whereas the lines represent raw data. Arrows from the upper horizontal axis indicate circadian orthophases for each group.
Figure 2. Circadian variation in tPA:Ag (left) and PAI:Ag (right) in healthy subjects and patients with cirrhosis. The curve represented for each group corresponds to the best fitted model obtained by population multiple-components analysis, whereas lines represent raw data. Arrows from the upper horizontal axis indicate circadian orthophases for each group.
1516
Piscaglia et al.
Among various circadian functions, we chose to study tPA:Ag and PAI:Ag for the assessment of the timekeeping system in comparison to melatonin, because these show stable circadian variations in both healthy and diseased subjects (12) and are not affected by body position, food intake, or exercise (10, 11, 23), and thus most environmental factors that could bias the study were a priori excluded. A more frequent sampling than we did would theoretically allow a more accurate description of circadian curves. However, it seems impractical. Each fibrinolytic assessment, in fact, requires a new venipuncture and an awakened subject, so that a very frequent sampling would also likely interfere with the physiological rhythm, especially during the night. A previous study reported the disappearance of the normal circadian variations of heart rate and blood pressure in cirrhosis (13), but these variables are considerably affected by food intake and in posture changes (24). As far as melatonin is concerned, our findings in cirrhotics with severe liver impairment confirm findings by others, who reported a delayed peak time of melatonin in cirrhosis (2). The mechanism of this delay remains to be fully elucidated (3, 7). Our data support the hypothesis that impaired catabolism of melatonin or of other substances, related to reduced hepatic function, is the main determinant of this delay and not a phase derangement in the rhythmic output of the internal clock. In fact, a significant delay was present in patients with more severe impairment of liver function— namely, in Child-Pugh classes B/C— but not in those with nearly normal hepatic function (i.e., in class A). On the contrary, almost identical acrophase times for PAI:Ag and tPA:Ag were observed not only in controls but also in patients with mild or severe impairment of liver function. Unlike melatonin, which is metabolized mainly by the liver (4, 25), PAI:Ag and tPA:Ag undergo only partial hepatic metabolism, and this might account for the preservation of normal phase rhythms. These findings might support the hypothesis that impaired peripheral catabolism of melatonin causes its delayed peak (3). A second possibility is that substances acting on the central nervous system and present in excess because of reduced hepatic clearance—as, for instance, compounds binding benzodiazepine receptors (26)—influence the drive of the internal clock and the effect of light/dark cycles on mesencephalic functions and melatonin production. Indeed, the suprachiasmatic nuclei of the anterior hypothalamus, site of the endogenous “biological clock,” have a complex array of connections to other structures within the central nervous system and to different effector and neurohormonal pathways (8). Therefore it might be hypothesized that exogenous or endogenous agents might differently affect the various pathways (i.e., those to the pineal gland for melatonin and, hypothetically, to vascular endothelium for fibrinolytic factors), making the final pattern and oscillation phase different. Had derangement of the internal clock output been the primary alteration for the delay of the melatonin peak, we might have expected a
AJG – Vol. 97, No. 6, 2002
delay also in the phase of the other substances tested, unless we postulate the presence of different and independent internal clocks, a hypothesis that differs from the currently accepted theory of a single clock in the suprachiasmatic nuclei (9). A potential criticism of this hypothesis might be the lack of known direct connections between the central nervous system and peripheral output of fibrinolytic components. However, this criticism intrinsically implies the currently nondemonstrated hypothesis of mechanisms located outside the central nervous system that are able to mantain strict circadian phase rhythms, but new discoveries in this field are possible. In addition, the present finding of preserved circadian phase of the central timekeeping system is corroborated by the results of a previous study showing that, in cirrhotic patients, the circadian pattern of plasma cortisol, another indicator of the circadian clock, is superimposable on that in controls (27). It is noteworthy that the amplitudes of the rhythms differed in controls and patients. In more detail, the amplitudes of the curves of tPA and PAI antigens were reduced in cirrhosis and, particularly, nearly halved in Child-Pugh B/C patients in comparison to controls, unlike melatonin (Table 1). A possible explanation is the increased 24-h basal level of fibrinolytic components in cirrhotics (19), upon which circadian variations are superimposed. This view would imply that the extent of the circadian increase is discrete, so that when it is applied upon a higher basal level it becomes relatively smaller. This hypothesis, however, does not explain why such a reduction of amplitude is not observed in melatonin, although the different extents of amplitudes in fibrinolytic parameters and melatonin (roughly 150% vs 25%) might play a role. In conclusion, the present pilot study suggests that the phase of the internal timekeeping system, at least as assessed by fibrinolytic components, is not modified in cirrhosis. It is also apparent that future studies evaluating circadian functions in cirrhosis should not include melatonin alone as an indicator of the rhythm. The exact cause of the phase shift in melatonin circadian rhythm in advanced cirrhosis and its connection with sleep disturbance remain to be elucidated.
ACKNOWLEDGMENTS Supported by MURST (Ministry of University and Scientific and Technological Research) ex-40% funds 1998–2000. The authors are grateful to Prof. A. Blei for helpful suggestions, K. Illmer for excellent technical assistance and kindness, and Drs. G. Donati, M. Valgimigli, and N. Celli for cooperation in sample collection. Reprint requests and correspondence: Fabio Piscaglia, M.D., Divisione Medicina Interna—Bolondi, via Albertoni 15, 40138 Bologna, Italy. Received Mar. 9, 2001; accepted Nov. 6, 2001.
AJG – June, 2002
Circadian Clock in Cirrhosis
REFERENCES 1. Cordoba J, Cabrera J, Lataif L, et al. High prevalence of sleep disturbance in cirrhosis. Hepatology 1998;27:339 – 45. 2. Steindl P, Finn B, Bendok B, et al. Disruption of the diurnal rhythm of plasma melatonin in cirrhosis. Ann Intern Med 1995;123:274 –7. 3. Steindl P, Ferenci P, Marktl W. Impaired hepatic catabolism of melatonin in cirrhosis. Ann Intern Med 1997;127:494. 4. Iguchi H, Kato K, Ibayashi H. Melatonin serum levels and metabolic clearance rate in patients with liver cirrhosis. J Clin Endocrinol Metab 1982;54:1025–7. 5. Brainard G, Lewy A, Menaker M, et al. Dose-response relationship between light irradiance and the suppression of plasma melatonin in human volunteers. Brain Res 1988;454: 212– 8. 6. Lewy A, Ahmed S, Jackson J, et al. Melatonin shifts human circadian rhythms according to a phase-response curve. Chronobiol Int 1992;9:380 –92. 7. Garfinkel D, Zisapel N. Liver cirrhosis and circadian rhythm. Ann Intern Med 1996;125:154 (letter). 8. Blei A, Zee P. Abnormalities of circadian rhythmicity in liver disease. J Hepatol 1998;29:832–5. 9. Hastings M. The brain, circadian rhythms, and clock genes. Br Med J 1998;317:19 –26. 10. Kofoed K, Gleerup G, Hedman C, et al. The circadian variation in fibrinolytic activity is not related to posture. Thromb Res 1994;15:447–50. 11. Andreotti F, Kluft C. Circadian variation of fibrinolytic activity in blood. Chronobiol Int 1991;8:336 –51. 12. Andreotti F, Prevost R, Lupi A, et al. Circadian patterns of plasma t-PA and PAI-1 antigen levels in patients with severe peripheral atherosclerosis. Fibrinolysis 1996;10(suppl 2):159 – 60. 13. Møller S, Wiinberg N, Henriksen J. Noninvasive 24-hour ambulatory arterial blood pressure monitoring in cirrhosis. Hepatology 1995;22:88 –95. 14. Irokawa M, Nishinaga M, Funayama H, et al. Effect of a change in the sleep/wake cycle on the diurnal variation of fibrinolytic parameters. J Thromb Thrombolysis 1998;5: 165– 8. 15. Peternel P, Stegnar M, Salobir U, et al. Shift work and circa-
16.
17.
18. 19.
20.
21.
22.
23.
24.
25.
26.
27.
1517
dian rhythm of blood fibrinolytic parameters. Fibrinolysis 1990;4(suppl 2):113–5. Andreotti F, Davies G, Hacektt D, et al. Major circadian fluctuations in fibrinolytic factors and possible relevance to time of onset of myocardial infarction, sudden cardiac death and stroke. Am J Cardiol 1988;15:635–7. Andreotti A, Davies G, Hackett D, et al. Circadian variation of fibrinolytic factors in normal human plasma. Fibrinolysis 1988;2:90 –2. Grimaudo V, Hauert J, Bachmann F, et al. Diurnal variation of the fibrinolytic system. Thromb Haemost 1988;59:495–9. Piscaglia F, Siringo S, Hermida R, et al. Diurnal changes of fibrinolysis in patients with liver cirrhosis and esophageal varices. Hepatology 2000;31:349 –57. Mojon A, Fernandez JR, Hermida RC. Chronolab: An interactive software package for chronobiologic time series analysis written for the Macintosh computer. Chronobiol Int 1992; 9:403–12. Fernandez JR, Hermida RC. Inferential statistical method for analysis of nonsinusoidal hybrid time series with unequidistant observations. Chronobiol Int 1998;15:191–204. Ferna´ ndez J. Nonparametric method for comparison of parameters from rhythmometric models with multiple components. Biol Rhythm Res 1995;26:387. Chandler W, Schwartw R, Stratton J, et al. Effects of endurance training on the circadian rhythm of fibrinolysis in men and women. Med Sci Sports Exerc 1996;28:647–55. Bernardi M, Di Marco C, Trevisani F, et al. The hemodynamic status of preascitic cirrhosis: An evaluation under steady-state conditions after postural change. Hepatology 1992;16:341– 6. Lane E, Moss H. Pharmacokinetics of melatonin in man: First pass hepatic metabolism. J Clin Endocrinol Metab 1985;61: 1214 – 6. Lozeva V, Valjakka A, Lecklin A, et al. Effects of histamine H1 receptor blocker, pyrilamine, on spontaneous locomotor activity of rats with long-term portacaval anastomosis. Hepatology 2000;31:336 – 44. Bernardi M, Palma RD, Trevisani F, et al. Chronobiological study of factors affecting plasma aldosterone concentration in cirrhosis. Gastroenterology 1986;91:683–91.
Vol. 97, No. 6, 2002 ISSN 0002-9270/02/$22.00 PII S0002-9270(02)04155-2
Cirrhosis Does Not Shift the Circadian Phase of Plasma Fibrinolysis Fabio Piscaglia, M.D., Ph.D., Ramon C. Hermida, Ph.D., Sebastiano Siringo, M.D., Cristina Legnani, Ph.D., Giuliano Ramadori, and Luigi Bolondi, M.D. Division of Internal Medicine, Department of Internal Medicine and Gastroenterology, and Department of Angiology and Blood Coagulation, University of Bologna, Bologna, Italy; Bioengineering and Chronobiology Laboratories, ETSI Telecomunicacio`n, University of Vigo, Vigo, Spain; and Department of Internal Medicine, University of Go¨ttingen, Go¨ttingen, Germany
OBJECTIVES: The aim of the present investigation was to study the endogenous circadian clock phase rhythm in cirrhotic patients. METHODS: The study population comprised 13 patients with cirrhosis (seven in Child-Pugh class A and six in classes B/C) and nine healthy controls. Plasma melatonin, tissue plasminogen activator antigen, and plasminogen activator inhibitor 1 antigen were measured at 4-h intervals over a 24-h period. Multiple-components rhythmometry using population mean cosinor methods were employed to analyze the findings. RESULTS: All three variables were characterized in both patients and controls by a statistically significant circadian rhythm, with similar profiles. The peak times of tissue plasminogen activator and plasminogen activator inhibitor 1 antigens were practically identical in controls and cirrhotic patients, irrespective of Child-Pugh class (calculated peak at times 6:52, 6:56, and 7:20 for the inhibitor in controls and Child-Pugh class A and classes B/C patients, respectively; p ⫽ ns), whereas the peak of melatonin was delayed in classes B/C patients (at times 2:08, 1:56, and 4:00, respectively; p ⬍ 0.05). CONCLUSION: The similar circadian phases of plasminogen activator inhibitor antigen in controls and cirrhotic patients in the present investigation indicates that the output rhythm of the internal timekeeping system is not shifted in this pathological condition. (Am J Gastroenterol 2002;97: 1512–1517. © 2002 by Am. Coll. of Gastroenterology)
INTRODUCTION Although sleep disturbance is considered a classic sign of hepatic encephalopathy, it can frequently be revealed also in cirrhotic patients without encephalopathy (1). Abnormalities in the circadian timekeeping system have been suggested to be responsible for this disturbance. Indeed, higher S.S. is currently affiliated with the Division of Internal Medicine, Ospedale Garibaldi, Catania, Italy.
daytime levels of melatonin, considered as an indicator of the internal clock, and a phase delay of the onset of rise in nocturnal plasma melatonin have been reported in cirrhosis (2– 4). The liver participates in the degradation of melatonin to its metabolites, which are then excreted into urine and bile. Hence, reduced hepatic function in cirrhosis may explain decreased urinary metabolite excretion and increased levels of plasma melatonin (3, 4). In experimental studies high plasma levels of melatonin are able to shift the circadian phase, either advancing it or delaying it according to the day time of melatonin administration, in a direction opposite to that induced by light exposure (5, 6). Unfortunately, however, it is difficult to predict the possible effect of a chronic increase of plasma melatonin, like that observed in cirrhosis. Consequently, because the delayed peak observed in human cirrhosis might reflect a primary disturbance of the internal clock (2) and the impaired hepatic metabolism of melatonin (3, 4), it is not yet clear whether the phase of the circadian timekeeping system is primarily deranged or secondarily or not at all (2, 3, 7). In fact, other than studies with melatonin, little work has been done to investigate circadian parameters in chronic liver disease (8), and the function of the internal clock remains unclear in cirrhosis. The internal clock is hypothesized to be located in the hypothalamus in humans, from where it acts on neural and endocrine pathways to regulate individual circadian rhythms (9). To date there is no way to directly assess this clock in humans; its output has therefore to be investigated through its effects on downstream pathways, of which plasma melatonin is one of the main. Unfortunately, most of these pathways are theoretically disturbed by the metabolic and hemodynamic alterations present in cirrhosis, making it particularly difficult to ascertain whether the circadian clock phase is shifted or not in this pathological condition, as discussed for melatonin. Among various possible substances or circulatory parameters known to present diurnal variations and hence possibly suitable for use in a comparison with melatonin to investigate circadian rhythms in humans, we chose to assess some
AJG – June, 2002
components of the fibrinolytic system. In fact, these possess a strong and well-defined rhythm, largely independent from environmental factors and disease conditions (10 –12), at variance with circulatory parameters, such as blood pressure and heart rate, which usually suffer influence by external conditions, particularly in cirrhosis (13). The exact connection of the fibrinolytic system with the internal clock is still unclear, as it is for other circadian indicators, but it and the others are linked to the light/dark cycle, as demonstrated by a certain phase derangement in shift workers or under experimental conditions (14, 15). Total plasma fibrinolytic activity has been known for several decades to follow a circadian rhythm, with the peak in late afternoon. About a dozen years ago, the role of individual fibrinolytic factors was established (16). In particular, it was demonstrated that total plasma fibrinolytic activity is mainly due to the activity of tissue plasminogen activator (tPA) and governed by the interaction of the plasma antigen levels of the latter with the activity of its main inhibitor, the plasminogen activator inhibitor 1 (PAI-1) (14, 16, 17). The activity of PAI-1, in turn, strictly follows PAI-1 antigen levels, which ultimately determine the circadian rhythm of spontaneous fibrinolytic activity in blood (12, 18). The aim of the present study was to investigate the circadian patterns of melatonin, tPA, and PAI antigens in a cohort of cirrhotic patients to establish whether the circadian output phase of the endogenous timekeeping system is preserved or modified in this pathological condition.
MATERIALS AND METHODS Patients The study was performed in 13 hospitalized patients (mean age ⫾ SD ⫽ 54.3 ⫾ 8.1 yr; 10 males and three females) with liver cirrhosis (Child-Pugh class: A [seven], B [five], and C [one]; 11 with a virus-related disease, hepatitis B virus or hepatitis C virus, and two with alcoholic cirrhosis) and in nine healthy subjects (mean age ⫽ 50.2 ⫾ 10.3 yr), recruited from the hospital staff and their relatives. Most subjects were initially enrolled for a separate study assessing the diurnal pattern of fibrinolytic activity in cirrhosis, and a few data for control subjects are also reported elsewhere (19). At the time of the study all subjects had abstained from alcohol for at least 1 month. All virus-related cirrhotic patients and most controls had been teetotalers for a long time. The two alcoholic patients had abstained from alcohol for at least 1 yr. The remaining control subjects had alcohol intakes below 20 g/day. Previous or present signs of hepatic encephalopathy (minimal or mild) were shown in two out of six (33%) cirrhotic patients in Child-Pugh classes B/C, whereas no such signs were present in patients in class A. All gave their informed consent to participate in the study, whose design was in accordance with the principles of the Declaration of Helsinki and was approved by the local ethical committee.
Circadian Clock in Cirrhosis
1513
Study Protocol Starting at 8:00 AM of the day of the study and continuing until 8:00 AM the next day, blood withdrawals were repeated at intervals of 4 h. The subjects lay supine for at least 15 min before each blood withdrawal. Artificial illumination was permitted, if needed, only from 7:00 AM to 10:30 PM, whereas in the remaining period a very dim artificial light (less than 15 lux) was turned on only to allow correct venipunctures at 12:00 AM and 4:00 AM; otherwise patients remained in the dark. Subjects were asked to lie in their beds from 10:30 PM to 7:00 AM, whereas in the remaining period they were allowed to move freely within the hospital. Smoking and consumption of caffeine-containing substances were prohibited from midnight of the day preceding the study. Breakfast, lunch, and dinner were administered just after the measurements of 8:00 AM, 12:00 PM, and 8:00 PM; in the intervals subjects were permitted to drink water freely but not to consume any caloric-containing solid or liquid. Plasma Assays Blood samples were drawn from an antecubital vein with minimal stasis and clean venipuncture, using a 21-gauge butterfly needle. Separate venipunctures were performed at each time point. Venous blood was immediately separated into different aliquots: 1) one anticoagulated with trisodium citrate (0.129 mol/L, 1/10) for tPA measurement (tPA:Ag, TintElize tPA, Biopool, Umea, Sweden); 2) a second put in tubes (Diatube Htm; sodium citrate, citric acid, theophylline, adenosine and dipyridamole, dipyridamole [Diagnostica Stago, Asnie`res, France]) to avoid platelet activation and secretion of PAI and to correctly measure it (PAI:Ag, TintElize PAI-1, Biopool); and 3) a third put in tubes containing ethylenediaminetetraacetic acid (2 mg/ml) and aprotinin (400 U/ml) for melatonin assay (Melatonin-RIA, IBL, Hamburg, Germany). Blood samples were kept in melting ice until subjected to 20 min of centrifugation at 3500 rpm and 4°C. Platelet-poor plasma was distributed in coded plastic tubes, snap frozen in liquid nitrogen, and stored at ⫺20°C until assayed. All measurements were done in duplicate. Statistical Analysis Data were analyzed on a Power Macintosh G3 microcomputer using Chronolab, a software package for biological time series analysis by linear least-squares estimation (20). Circadian characteristics of each variable were established by population multiple-components analysis (21), a method designed for analysis of nonsinusoidal hybrid data (time series of data collected from a group of subjects, as in clinical situations involving patients). The method provides estimates of the rhythm-adjusted mean or midline estimating statistic of rhythm, defined as the average value of the rhythmic function fitted to the data, as well as the amplitude (defined as half the extent of the rhythmic change in a cycle approximated by the fitted curve) and acrophase (lag from a defined reference time point— here, local midnight— of the
1514
Piscaglia et al.
AJG – Vol. 97, No. 6, 2002
Table 1. Results of Multiple-Components Rhythmometric Analysis Group, No. of Subjects Melatonin Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups tPA:Ag Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups PAI:Ag Control, n ⫽ 9 Cirrhosis (all), n ⫽ 13 Child A cirrhosis Child B/C cirrhosis p among groups
MESOR
p for the Rhythm
34.8 pg/ml 221.9 pg/ml 187.7 pg/ml 261.8 pg/ml
ORTHO
95% Cls
AMPLI
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.002
2:08 2:36 1:36 4:00 ⬍0.001
1:28–2:40 1:28–3:32 0:40–2:40 3:36–4:24
142% 163% 197% 159%
3.5 ng/ml 11.8 ng/ml 6.3 ng/ml 18.1 ng/ml
⬍0.05 ⬍0.05 0.165, ns 0.331, ns
9:04 8:43 8:40 8:44 0.852, ns
8:00–9:56 7:24–11:56 7:18–11:47 7:33–12:08
17% 12% 14% 10%
10.1 ng/ml 14.1 ng/ml 7.6 ng/ml 21.5 ng/ml
⬍0.02 ⬍0.005 ⬍0.02 ⬍0.05
6:52 7:16 6:56 7:20 0.690, ns
5:16–8:16 6:16–8:12 5:32–9:24 6:16–8:28
37% 25% 33% 19%
Results of the circadian analysis in the various subgroups. AMPLI ⫽ amplitude, half the distance between the maximum and the minimum of the fitted curve; MESOR ⫽ midline estimating statistics of rhythm, average value of the best fitted curve with two harmonic components of 24 and 12 h; ORTHO ⫽ orthophase, lag from a defined reference time point (here local midnight) of the crest time in the curve fitted to the data; p among groups ⫽ p from the comparison of the times of orthophase in the three groups of controls and Child-Pugh A and B/C patients; p for the rhythm ⫽ p from testing the zero amplitude assumption for the circadian rhythm.
crest time in the curve fitted to the data) for every fitted component. When all fitted components are harmonics from a fundamental period, the method of multiple components also provides three additional parameters: the overall amplitude, defined as half the difference between the maximum and the minimum of the best fitted curve in one fundamental period; the orthophase, defined as the lag from a defined reference time point of the crest time, within a fundamental period, in the curve of multiple components fitted to the data (21); and the bathyphase, defined as the lag from a defined reference time point within a fundamental period, of the time of the lowest value in the curve of multiple components fitted to the data. The orthophase (similarly to the acrophase of each component) and the bathyphase are usually expressed as negative angular degrees, with 360° ⫽ 1 fundamental period, or, preferably, in time units (here, h:min, with 24 h used consistently as the fundamental period). p for the rejection of the zero amplitude assumption was determined, with rhythm detection considered statistically significant if p ⱕ 0.05. Circadian parameters were also compared between patients and controls with a nonparametric test developed to compare parameters obtained from population multiple-components analysis (22). Because the main objective of this study was to compare the temporal relationship of the studied variables between patients and controls, interindividual differences were reduced by expressing the data in percentage of each individual mean.
24 h and 12 h (Table 1 and Figs. 1 and 2). There was no statistically significant difference in the time of peak (orthophase) of melatonin between healthy subjects and patients with cirrhosis, when the latter were considered as one single group, being 2:08 and 2:36, respectively (p ⫽ 0.576) (Fig. 1, left). However, a statistically significant difference became evident (p ⬍ 0.001) (Table 1) when patients were grouped according to liver function and divided into the two subgroups of Child-Pugh classes A and B/C. Indeed, the circadian orthophase was delayed in classes B/C by 2 h 24 min, as compared to class A patients, (p ⫽ 0.004) (Fig. 1, right) and by 1 h 52 min, as compared to controls (p ⫽ 0.012) (Table 1). There was no significant difference between class A patients and controls (32 min) (p ⫽ 0.482). The calculated peak times of tPA:Ag and PAI:Ag in patients were practically identical to those in controls (differences of 24 and 20 min, respectively; ps ⫽ 0.735 and 0.786 for controls vs cirrhotics as one single group [Fig. 2] and 0.852 and 0.690 considering the three groups). However, unlike melatonin, the peak times were also comparable considering patients with mild and severe liver functional impairment separately: in fact, no intergroup difference could be found (orthophase at 8:44 and 8:40 [p ⫽ 0.999] for tPA:Ag in class A and B/C patients, respectively, and 6:56 and 7:20 [p ⫽ 0.558] for PAI:Ag, although the circadian curve did not reach statistical significance in the two latter smaller subgroups) (Table 1).
RESULTS
DISCUSSION
All three variables were characterized for both patients and controls by a statistically significant circadian rhythm that can be described by a model with two components, with periods of
The present findings indicate that the timekeeping system does not undergo significant phase shifts in cirrhosis, at least as far as fibrinolytic components are concerned.
AJG – June, 2002
Circadian Clock in Cirrhosis
1515
Figure 1. Circadian variation in melatonin in healthy subjects and patients with cirrhosis (left) and in cirrhotic patients with different degrees of disease severity (right). The curve represented for each group corresponds to the best fitted model obtained by population multiple-components analysis, whereas the lines represent raw data. Arrows from the upper horizontal axis indicate circadian orthophases for each group.
Figure 2. Circadian variation in tPA:Ag (left) and PAI:Ag (right) in healthy subjects and patients with cirrhosis. The curve represented for each group corresponds to the best fitted model obtained by population multiple-components analysis, whereas lines represent raw data. Arrows from the upper horizontal axis indicate circadian orthophases for each group.
1516
Piscaglia et al.
Among various circadian functions, we chose to study tPA:Ag and PAI:Ag for the assessment of the timekeeping system in comparison to melatonin, because these show stable circadian variations in both healthy and diseased subjects (12) and are not affected by body position, food intake, or exercise (10, 11, 23), and thus most environmental factors that could bias the study were a priori excluded. A more frequent sampling than we did would theoretically allow a more accurate description of circadian curves. However, it seems impractical. Each fibrinolytic assessment, in fact, requires a new venipuncture and an awakened subject, so that a very frequent sampling would also likely interfere with the physiological rhythm, especially during the night. A previous study reported the disappearance of the normal circadian variations of heart rate and blood pressure in cirrhosis (13), but these variables are considerably affected by food intake and in posture changes (24). As far as melatonin is concerned, our findings in cirrhotics with severe liver impairment confirm findings by others, who reported a delayed peak time of melatonin in cirrhosis (2). The mechanism of this delay remains to be fully elucidated (3, 7). Our data support the hypothesis that impaired catabolism of melatonin or of other substances, related to reduced hepatic function, is the main determinant of this delay and not a phase derangement in the rhythmic output of the internal clock. In fact, a significant delay was present in patients with more severe impairment of liver function— namely, in Child-Pugh classes B/C— but not in those with nearly normal hepatic function (i.e., in class A). On the contrary, almost identical acrophase times for PAI:Ag and tPA:Ag were observed not only in controls but also in patients with mild or severe impairment of liver function. Unlike melatonin, which is metabolized mainly by the liver (4, 25), PAI:Ag and tPA:Ag undergo only partial hepatic metabolism, and this might account for the preservation of normal phase rhythms. These findings might support the hypothesis that impaired peripheral catabolism of melatonin causes its delayed peak (3). A second possibility is that substances acting on the central nervous system and present in excess because of reduced hepatic clearance—as, for instance, compounds binding benzodiazepine receptors (26)—influence the drive of the internal clock and the effect of light/dark cycles on mesencephalic functions and melatonin production. Indeed, the suprachiasmatic nuclei of the anterior hypothalamus, site of the endogenous “biological clock,” have a complex array of connections to other structures within the central nervous system and to different effector and neurohormonal pathways (8). Therefore it might be hypothesized that exogenous or endogenous agents might differently affect the various pathways (i.e., those to the pineal gland for melatonin and, hypothetically, to vascular endothelium for fibrinolytic factors), making the final pattern and oscillation phase different. Had derangement of the internal clock output been the primary alteration for the delay of the melatonin peak, we might have expected a
AJG – Vol. 97, No. 6, 2002
delay also in the phase of the other substances tested, unless we postulate the presence of different and independent internal clocks, a hypothesis that differs from the currently accepted theory of a single clock in the suprachiasmatic nuclei (9). A potential criticism of this hypothesis might be the lack of known direct connections between the central nervous system and peripheral output of fibrinolytic components. However, this criticism intrinsically implies the currently nondemonstrated hypothesis of mechanisms located outside the central nervous system that are able to mantain strict circadian phase rhythms, but new discoveries in this field are possible. In addition, the present finding of preserved circadian phase of the central timekeeping system is corroborated by the results of a previous study showing that, in cirrhotic patients, the circadian pattern of plasma cortisol, another indicator of the circadian clock, is superimposable on that in controls (27). It is noteworthy that the amplitudes of the rhythms differed in controls and patients. In more detail, the amplitudes of the curves of tPA and PAI antigens were reduced in cirrhosis and, particularly, nearly halved in Child-Pugh B/C patients in comparison to controls, unlike melatonin (Table 1). A possible explanation is the increased 24-h basal level of fibrinolytic components in cirrhotics (19), upon which circadian variations are superimposed. This view would imply that the extent of the circadian increase is discrete, so that when it is applied upon a higher basal level it becomes relatively smaller. This hypothesis, however, does not explain why such a reduction of amplitude is not observed in melatonin, although the different extents of amplitudes in fibrinolytic parameters and melatonin (roughly 150% vs 25%) might play a role. In conclusion, the present pilot study suggests that the phase of the internal timekeeping system, at least as assessed by fibrinolytic components, is not modified in cirrhosis. It is also apparent that future studies evaluating circadian functions in cirrhosis should not include melatonin alone as an indicator of the rhythm. The exact cause of the phase shift in melatonin circadian rhythm in advanced cirrhosis and its connection with sleep disturbance remain to be elucidated.
ACKNOWLEDGMENTS Supported by MURST (Ministry of University and Scientific and Technological Research) ex-40% funds 1998–2000. The authors are grateful to Prof. A. Blei for helpful suggestions, K. Illmer for excellent technical assistance and kindness, and Drs. G. Donati, M. Valgimigli, and N. Celli for cooperation in sample collection. Reprint requests and correspondence: Fabio Piscaglia, M.D., Divisione Medicina Interna—Bolondi, via Albertoni 15, 40138 Bologna, Italy. Received Mar. 9, 2001; accepted Nov. 6, 2001.
AJG – June, 2002
Circadian Clock in Cirrhosis
REFERENCES 1. Cordoba J, Cabrera J, Lataif L, et al. High prevalence of sleep disturbance in cirrhosis. Hepatology 1998;27:339 – 45. 2. Steindl P, Finn B, Bendok B, et al. Disruption of the diurnal rhythm of plasma melatonin in cirrhosis. Ann Intern Med 1995;123:274 –7. 3. Steindl P, Ferenci P, Marktl W. Impaired hepatic catabolism of melatonin in cirrhosis. Ann Intern Med 1997;127:494. 4. Iguchi H, Kato K, Ibayashi H. Melatonin serum levels and metabolic clearance rate in patients with liver cirrhosis. J Clin Endocrinol Metab 1982;54:1025–7. 5. Brainard G, Lewy A, Menaker M, et al. Dose-response relationship between light irradiance and the suppression of plasma melatonin in human volunteers. Brain Res 1988;454: 212– 8. 6. Lewy A, Ahmed S, Jackson J, et al. Melatonin shifts human circadian rhythms according to a phase-response curve. Chronobiol Int 1992;9:380 –92. 7. Garfinkel D, Zisapel N. Liver cirrhosis and circadian rhythm. Ann Intern Med 1996;125:154 (letter). 8. Blei A, Zee P. Abnormalities of circadian rhythmicity in liver disease. J Hepatol 1998;29:832–5. 9. Hastings M. The brain, circadian rhythms, and clock genes. Br Med J 1998;317:19 –26. 10. Kofoed K, Gleerup G, Hedman C, et al. The circadian variation in fibrinolytic activity is not related to posture. Thromb Res 1994;15:447–50. 11. Andreotti F, Kluft C. Circadian variation of fibrinolytic activity in blood. Chronobiol Int 1991;8:336 –51. 12. Andreotti F, Prevost R, Lupi A, et al. Circadian patterns of plasma t-PA and PAI-1 antigen levels in patients with severe peripheral atherosclerosis. Fibrinolysis 1996;10(suppl 2):159 – 60. 13. Møller S, Wiinberg N, Henriksen J. Noninvasive 24-hour ambulatory arterial blood pressure monitoring in cirrhosis. Hepatology 1995;22:88 –95. 14. Irokawa M, Nishinaga M, Funayama H, et al. Effect of a change in the sleep/wake cycle on the diurnal variation of fibrinolytic parameters. J Thromb Thrombolysis 1998;5: 165– 8. 15. Peternel P, Stegnar M, Salobir U, et al. Shift work and circa-
16.
17.
18. 19.
20.
21.
22.
23.
24.
25.
26.
27.
1517
dian rhythm of blood fibrinolytic parameters. Fibrinolysis 1990;4(suppl 2):113–5. Andreotti F, Davies G, Hacektt D, et al. Major circadian fluctuations in fibrinolytic factors and possible relevance to time of onset of myocardial infarction, sudden cardiac death and stroke. Am J Cardiol 1988;15:635–7. Andreotti A, Davies G, Hackett D, et al. Circadian variation of fibrinolytic factors in normal human plasma. Fibrinolysis 1988;2:90 –2. Grimaudo V, Hauert J, Bachmann F, et al. Diurnal variation of the fibrinolytic system. Thromb Haemost 1988;59:495–9. Piscaglia F, Siringo S, Hermida R, et al. Diurnal changes of fibrinolysis in patients with liver cirrhosis and esophageal varices. Hepatology 2000;31:349 –57. Mojon A, Fernandez JR, Hermida RC. Chronolab: An interactive software package for chronobiologic time series analysis written for the Macintosh computer. Chronobiol Int 1992; 9:403–12. Fernandez JR, Hermida RC. Inferential statistical method for analysis of nonsinusoidal hybrid time series with unequidistant observations. Chronobiol Int 1998;15:191–204. Ferna´ ndez J. Nonparametric method for comparison of parameters from rhythmometric models with multiple components. Biol Rhythm Res 1995;26:387. Chandler W, Schwartw R, Stratton J, et al. Effects of endurance training on the circadian rhythm of fibrinolysis in men and women. Med Sci Sports Exerc 1996;28:647–55. Bernardi M, Di Marco C, Trevisani F, et al. The hemodynamic status of preascitic cirrhosis: An evaluation under steady-state conditions after postural change. Hepatology 1992;16:341– 6. Lane E, Moss H. Pharmacokinetics of melatonin in man: First pass hepatic metabolism. J Clin Endocrinol Metab 1985;61: 1214 – 6. Lozeva V, Valjakka A, Lecklin A, et al. Effects of histamine H1 receptor blocker, pyrilamine, on spontaneous locomotor activity of rats with long-term portacaval anastomosis. Hepatology 2000;31:336 – 44. Bernardi M, Palma RD, Trevisani F, et al. Chronobiological study of factors affecting plasma aldosterone concentration in cirrhosis. Gastroenterology 1986;91:683–91.