- Email: [email protected]
13C-Phenylalanine and 13C-Methacetin breath test to evaluate functional capacity of hepatocyte in chronic liver disease
LIVER,PANCREAS,AND BIllAllY TRACT
OIGESTLIUER OIS 2000;32:226-32
13C-Phenylalanine and 13GMethacetin breath test to evaluate functional capacity of hepatocyte in chronic liver disease S. Lara Baruaue M. Razquin I. Jimenezl A. Vazquezl J. P. Gisbert’ J. M. Pajaresl I
Background. To grade liver damage, Child-Pugh classification is used but these tests do not reflect the quantitative functional hepatic reserve. Aims. 13GPhenylalanine Breath Test and 13GMethacetin Breath Test are evaluated as possible tools, being both safe and easy to perform, to quantify functional hepatic reserve in chronic liver disease patients. Patients. Both tests were performed in 48 healthy volunteers and 48 chronic liver disease patients. Methods. Breath samples were collected after taking 13C-Phenylalanine (100 mg] and 13GMethacetin (75 mg]. 13COp enrichment was measured using mass spectrometry Results. Both tests discriminated the hepatic function, decreasing results of the 13C02 enrichment agreeing with the increasing severity of the hepatic patient (13GPhenylalanine Breath Test multiple correlation coefficient: 0.72, global pcO.001; Methacetin Breath Test: 0.73, p
from:
Departments of Clinical Analysis end I Gastroenterology, Hospital de la Princase, Universidad AutrSnome, Madrid, Spain. AlrkrrrfarE~endenw: Dr. S. Lara Baruque, Servicio An#sis Clinicos. Hospital de le Princasa, c/Diego de Leon 62, 28006 Madrid, Spain. Fex: -04-91-570267. E-mail: [email protected] Submitted September 7, 1999. Revised February 18, 2000. Accepted March 14, 2000.
226
Digest Liver Dis 2000;32:226-32 Key words: ITtion;
chronic
liver
Methacetin disease;
breath test; 13C-Phenylalanine breath liver tests; quantitative liver function
test;
Child-Pugh
classifica-
Introduction In patients with liver disease, several indicators are used to grade liver damage in accordance with the Child Pugh classification I: ascites, encephalopathy, bilirubin, albumin and prothrombin time. Albeit, these tests do not reflect the quantitative functional hepatic reserve, they only show the endpoint hepatocellular damage, the hepatic injury. Child’s classification has been modified in various ways (deletion of nutrition status, addition of age, blood ammonia, number of previous bleeds) with little, if any, improvement.
S. Lara Baruque et al.
Standard liver blood tests provide information as to hepatocyte injury [alanine transaminase (ALT), aspartate transaminase (AST)], and bile formation (alkaline phosphatase activity and bilirubin). With the exception of serum albumin and prothrombin time, these tests do not reveal information concerning the residual hepatic function 2. For example, in a patient with inactive cirrhosis, serum hepatic enzyme levels may be normal in spite of the decreased hepatic function. On the other hand, hepatic function may be preserved in the presence of an extensive hepatocellular necrosis shown by a marked increase in transaminases. Hepatic function failure is, finally, the main cause of mortality in patients with serious liver dysfunction. Several quantitative tests have been proposed to measure residual hepatic function 3 4. Different aspects of liver function may be measured on the basis of the substrate selected. Substrates such as D-sorbitol 5 and galactose 6, that the liver efficiently extracts from blood, are chosen to measure hepatic blood flow. Other substances with a low hepatic extraction ratio, such as phenacetin, methacetin, caffeine, erythromycin and aminopyrine ’ * , have been investigated extensively as a simple and reliable index of hepatic microsomal enzyme reserve. Other substances with an intermediate hepatic extraction rate such as indocyanine green 9 and Lidocaine I0 ” are used, which combine microsomal function study and the functional liver plasma flow. Compared with healthy controls, the clearance of these substances are decreased in individuals with chronic liver disease and cirrhosis. On the other hand, liver is the main organ involved in the metabolism of aromatic aminoacids including phenylalanine, tryptophan and tyrosine I2 13,thus in severe liver disease, the central plasma clearance of these aminoacids is decreased and it has been shown that the central plasma clearance rate of aminoacids by the liver is an accurate predictor of hepatocyte function and reserve. Application of the clearance of these substances to quantify liver function is limited by the need of qualified staff, complexity, and need to obtain blood or body fluids for analysis. The possibility of devising a non-invasive test able to provide information concerning the functional liver cell mass, not requiring repeated blood sampling, indeed a safe test, easy to perform, and cost-effective, encouraged us to carry out this study. The purpose of this study was, therefore, to investigate hepatic cytosolit enzyme activity through the ‘C-Phenylalanine breath test (PBT) and, on the other hand, apply the 13CMethacetin breath test (MBT) to study the hepatic microsomal function. In the breath tests, a tracer molecule (Phenylalanine or Methacetin) is labelled with carbons enriched with the stable isotope of mass 13 (‘“C), which is a naturally occurring isotope. The en--..
richment of 13C02 in the expired air is analysed by an isotope ratio mass spectrometer (IRMS), a piece of equipment which is easy to manage, and the use of which has been increasingly used over the last few years in many hospitals. The most obvious advantage of stable isotopes is that they are non-radioactive and do not present any risk to subjects, and, furthermore, need no blood sampling and time-consumption is limited (one hour). Both non-invasive tests have been evaluated as possible tools to establish the hepatic reserve in patients with liver diseases and, indirectly, also hepatocyte damage.
Patients and methods Patients During a period of eight months, 48 patients (29 males and 19 females, mean age 58.2&13 years, with histologically confirmed liver diseases: 10 cirrhosis Child A, 15 cirrhosis Child B, 12 cirrhosis Child C, and 11 chronic hepatitis) were admitted consecutively to our hospital either as in-patients or out-patients. The viral aetiology (hepatitis B virus or hepatitis C virus) was the most frequent (n=30); alcoholic aetiology (n=16) and a histological diagnosed primary bile cirrhosis completed the hepatic dysfunction aetiology. In one patient, aetiology was unknown. Any history of smoking, alcohol and drug consumption was recorded. The drugs administered to the patients (spironolactone, furosemide, lactulose etc.) have no interactions with the oxidative metabolism of Methacetin by cytochrome PYC lA2 14. Patients with respiratory diseases, which could possibly alter breath function, were excluded, as were patients with any digestive disease which could cause malabsortion. Controls A total of 48 healthy volunteers (22 males and 26 females, mean age 41.7&15 years) were included as control subjects. None of these control subjects had a history of liver, breath or gastrointestinal malabsortion disease, neither were they receiving any medication at the time of the study. Biochemical hepatic function, prothrombin time and platelet count were normal in all subjects. Any history of smoking and alcohol abuse were recorded. Before the study, these subjects were on a normal diet and consumed no drugs for two weeks prior to the test which could affect hepatic enzyme activity. Written informed consent was obtained in healthy volunteers and patients, prior enrolment in this prospective study, in accordance with the clinical assay protocol approved by the Ethics Commission of Clinical Investigation of our hospital. 221
Breath tests for hepatic function
Liver blood tests Standard liver blood tests (serum values of prothrombin time, albumin, total bilirubin, AST, ALT and alkaline phosphatase) were collected before performing the breath test. The diagnosis of cirrhosis was based on histological findings and all patients were classified by the Child Pugh classification which comprises ascites, encephalopathy, serum albumin and bilirubin levels, as well as prothrombin time. Patients were classified according to their score into Class A (score 5-6), Class B (score 7-9) or Class C (score lo- 15) ’ . The mean values of total bilirubin (mg/dl) and the standard deviation (SD) were Child A (1.32+0.82), Child B (2.48&2.44), Child C (4.49k1.93) and chronic hepatitis was (l&0.86). Methacetin and Phenylalanine breath test protocol The two quantitative liver function tests were carried out on different days (within one week) both in the control and patient groups. Breath tests were made following an overnight fast with no control of prior dietary intake. Duplicate baseline breath samples were collected before administration of the 13C-Phenylalanine ‘* l3 and the 13C-Methacetin dose I5 consisting in 100 mg L- (l-13C) Phenylalanine (99% 13C, Cambridge Isotope Laboratories, Andover, MA, USA) and 75 mg (1 -13C) Methacetin (99% 13C, Cambridge Isotope Laboratories, Andover, MA, USA) dissolved in 50 ml of water. Thereafter, duplicate breath samples were collected every 10 minutes for 1 hour. The 13C02 enrichment, expressed as cumulative percent oxidation of dispensed dose was measured with a stable isotope mass spectrometer (Europe Scientific Tracermass). Analytical data were expressed as percentages of 13C02 recovery per hour using an area under the curve (AUC) method and assumed COZ production rate of 5 mmol/min/m2 body surface area, as described by Schneider et al. 16. Statistical analysis A 95% confidence interval (CI) was calculated. Multiple linear regression was performed. The dependent variables were the cumulative percent oxidation of Phenylalanine or Methacetin, and independent variables were age, sex, tobacco (smokers and non smokers), alcohol (less or more than 40 g ethanol/day) and hepatic function (classified as healthy volunteers, chronic hepatitis, cirrhosis Child A, Child B and Child C). To study the differences between each group, the hepatic function was coded as dummy binary variables according to the coding scheme that uses the following level as reference level (e.g., healthy volunteers vs chronic hepatitis, chronic hepatitis vs cirrhosis Child A . . .). This statistical analysis is particularly indicated for increasing or decreas228
ing stair-like results. We used a backward modelling strategy. The F partial was the statistic used for comparing models and the significant level was set at 0.05. Changes over 10% in the regression coefficient were considered as suggestive of “confusion”. Belsley criteria l7 were followed to diagnose “collinearity”, which was not demonstrated. Finally, the correlation between PBT and MBT was studied using linear regression. The areas under the ROC (receiver operating characteristic) curves from the various cut-off points for PBT and MBT were calculated; the value of the maximal possible (ideal) area is one ‘* 19. ROC curves were calculated considering all collecting times (1 hour), and also for each collecting time (every 10 minutes). Results
As can be seen in Figure 1 both the PBT and MBT test discriminated the hepatic function, the decreasing results of the 13C02 correlating with the increasing severity of the hepatic patient graded according to Child classification and histological findings. The correlation between the cumulative percent PBT or MBT oxidation and hepatic function by multiple linear regression is depicted in Figure 2. Results of multiple linear regression analysis for cumulative percent Phenylalanine oxidation are summarized in Table I, where it can be observed that the only variables which correlated with Phenylalanine oxidation were age and hepatic function. The PBT discriminated between all the different groups except chronic hepatitis vs Child A cirrhosis. The highest regression coefficient values were for healthy vs hepatitis, and Child B vs Child C cirrhosis. Results of multiple linear regression analysis for cu-
El
0 Healthy Volunteen (1~48) OChronic Hepatitis (~11) ~~Cirrhosis Child A (140) SCimhosis Child B (n=15) N3rrhcsis Child C (1142)
Fig. 1. Phenylalanine Breath Test and Methacetin lative percent oxidation in healthy volunteers tients. Data presented as mean f SD.
Breath Test: Cumuand liver disease pe-
“--l__l
S. Lara Baruque et al.
IWe II. Correlation between the cumulative percent Methacetin oxidation and hepatic function by multiple linear regression. VPrialfle"
Repooien coef8cient
Healthy volunteers Chronic hepatitis Cirrhosis Child A Cirrhosis Chid 8 *Coded healthy Child A ficient:
Fig. 2. Correlation between the cumulative percent Phenylalanine or Methacetin oxidation and hepatic function by multiple linear regresion. The discontinuous lines indicate the 95% confidence interval.
I
I. Correlation between the cumulative percent Phenylalanine oxidation and hepatic function by multiple linear regression. Table
lbtriable 0.04 1.91 0.36 1.47 2.04
Age
Healthy volunteers” Chronic hepatitis” Cirrhosis Child A Cirrhosis Child El” ‘Coded healthy Child A ficient:
0.01 0.68 0.87 0.82 0.69
0.0106 0.0060 0.6856 0.0717 0.0038
as dummy va-iables, using as reference the following level (e.g., volunteers vs chronic hepatitis, chronic hepatitis vs cirrhosis e&l. Multiple correlation coefficient: 0.72 (determination coef52%); global p
3.98 0.78 4 5.3
Stsndsrd error
P
1.7 2.2 2.1 I.8
0.0196 0.7249 0.0603 0.0049
as dummy variables, using as reference the following level [e.g., volunteers vs chmnic hepatitis, chronic hepatitis vs cirrhosis etc.]. Mu/tip/e correlation coefficient: 0.73 ldatermination coef54%1 global pcO.001.
mulative percent Methacetin oxidation are summarized in Table II. MBT discriminated between all the different groups except chronic hepatitis vs Child A cirrhosis. The highest regression coefficient values were for healthy vs hepatitis, and for Child B vs Child C cirrhosis. The correlation coefficient between PBT and MBT was 0.63 (p
lhbla Ill. Evaluation of areas under ROC [receiver operating characteristic1 curves for Phenylalanine Breath Test and Methecetin Breath Test in the evaluation of hepatic function depending on collecting time.
P8T cellootitlg time
At-08
SE
IO min. 20 min. 30 min. 40 min. 50 min. 60 min. Total time
0.75 0.82 0.85 0.84 0.81 0.74 0.85
0.05 0.04 0.04 0.04 0.04 0.05 0.04
PST: WPhenylalanine
breath
tast;
MST: W-Methacetin
MST collecting
time
IO min.
20 min. 30 min. 40 min. 50 min. 60 min. Total time breath
test;
SE: Standard
Error;
Area
SE
0.86 0.86 0.82 0.81 0.80 0.78 0.84
0.84 0.04 0.04 0.04 0.05 0.05 0.04
Area: Area under ROC curve.
229
Breath tests for hepatic function
lb& WL Evaluation of ‘areas under ROC keceiver operating characteristic1 curvea and different cut-off points for Phenylalanine Breath Test in the evaluation of hepetic function.
Total
Healthy volunteers vs any hepatic dysfunction
0.85
0.04
5.5
0.75
0.84
4.5
Total
Healthy volunteers or chronic hepatitis vs cirrhosis
0.87
0.04
5
0.76
0.87
6
Total
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.90
Cl.04
4.92
0.89
0.86
6.5
30 min
Healthy volunteers vs any hepatic dysfunction
0.85
0.04
1.19
0.77
0.87
5.8
30 min
Healthy volunteersor chronic hepatitis vs ciwttosis
0.84
0.04
1.07
0.73
0.87
5.9
30 min
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.88
0.04
1.10
0.89
0.83
5.3
Tateh considering ttW ckmh%Him insteed cirrhossl:
b
all collecting times II hour]; Aree: Area under ROC curve; SE: Standerd Error; Cut-OR cut-off point for Phenytalenina Breath Test value in of hepetic function as one of the two categories [higher then cut-off point va{ua is indicetive of lass severe condition, e.g. healthy volunteers LR: Likabod ratio.
Table K Evaluation of areas under ROC kaceiver operating characteristic) curves and different cut-off points for Methacetin Breath Test in evaluation of hepatic function.
Total
Healthy volunteers vs any hepatic dysfunction
0.84
0.04
13.1
0.71
0.94
11.4
Total
Healthy volunteers or chronic hepatitis vs cirrhosis
0.86
0.04
11.7
0.78
0.95
15.3
Total
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.90
0.04
11.7
0.87
0.88
7.5
IO min
Healthy volunteers vs any hepatic dysfunction
0.86
0.04
0.93
0.79
0.87
6.3
IO min
Healthy volunteers or chronic hepatitis vs cirrhosis
0.89
0.04
0.90
0.86
0.83
5.2
IO min
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child El or Child C
0.89
0.04
0.65
0.84
0.83
4.8
Total: considering the classification instead cirrhosisl;
all collecting times I4 hour); Area: Area under HOC curve; SE: Standard Error; Cut-off: cut-off point for Methacatin Breath Test value in of hepetic function as one of the two categories [higher than cut-off point value is indicative of lass severe conditions e.g. healthy volunteers LR: Likelihood ratio.
The area was over 0.84 in all cases, with a likelihood ratio constantly over 4.5. The evaluation of areas under ROC curves and different cut-off points for MBT both considering all the colletting times and also the 10 min collecting time (the best one, as previously described), are summarized in Table V. The area was, in all cases, over 0.84, with a likelihood ratio constantly over 4.8. 230
Discussion Results emerging from the present study show that both PBT and MBT are able to discriminate the hepatic functional capacity between the different groups studied. As liver function becomes more severely impaired, as defined by Child Pugh classification I, so does hepatocyte metabolism (Fig. 1): both Phenylalanine and Methacetin metabolism are decreased in liver
--
S. Lara Bsruqua et al.
dysfunction. Phenylalanine aminoacid metabolism is blocked in liver disease so that phenylalanine hydroxylase, in these patients, is estimated to be 20% of the normal value 2o 2’ Methacetin, a derivative of phenacetin, undergoes a rapid 0-dealkylation by hepatic microsomal enzyme systems 15. The fact that the highest discriminatory capacity is for Child B vs Child C, could be interesting for monitoring the progression of liver dysfunction, predicting risk of surgical interventions and for determining the appropriate timing for liver transplantation. The fact that none of the tests discriminate between chronic hepatitis patients and Child A cirrhosis is easily explained due to the fact that hepatic function and biochemical data can be similar and a histological study is often needed to differentiate between these two stages 22. These tests are not affected by diet, which may interfere in the biochemical and clinical data used for hepatic dysfunction diagnosis in the Child Pugh score; e.g. the prior intake of salt may artificially increase the degree of ascites without really affecting the hepatic function 23, or a protein diet may artificially increase the encephalopathy 2425. Comparing the PBT and MBT tests, it can be seen that no important differences exist as far as concerns the yield of hepatic function information between these two tests. The area under ROC curves in always >0.84 with a likelihood ratio >4.5 in both tests (Tables IV, V). The advantage of the MBT is that only one breath collection is needed at 10 minutes (however, the PBT best collection time is at 30 minutes), so that MBT is quicker to perform (Table III). On the other hand, MBT values, in contrast with those of PBT, are not influenced by age (Tables I, II). As far as concerns carrying out both tests in the same patient, it can be seen that if both tests are positive, the sensitivity in the diagnosis is very high (95%). Therefore, if we need to use these breath tests as a screening test, it would be useful to perform both tests, although specificity will then be decreased. Today, the Child Pugh classification remains the most widely accepted tool as a measure of disease severity in liver disease. However, in addition to being safe, non-invasive, inexpressive, easy to perform, and well correlated with the Child Pugh score, PBT and MBT could provide additional information in that direct hepatocyte functional capacity is measured, avoiding any subjectivity in hepatic dysfunction staging (the degree of ascites is not always easy to establish) and also avoiding other elements that could modify the real degree of hepatic damage (salt intake and degree of ascites, or protein intake and degree of encephalopathy). The function tests can provide different information from obtained by a liver biopsy, but not such detailed information on hepatic disease as that provided by histological examination. Furthermore, function tests can-
not indicate the extent of inflammation or the degree of fibrosis. In summary, the conclusion of our study is that both PBT and MBT tests, which are safe and easy to perform, are able to discriminate hepatic functional capacity not only between healthy individuals and patients with liver disease, but also between different stages of hepatic dysfunction. The possibility that PBT and MBT could be useful to predict the prognosis of liver disease and the rate of disease progression, compared with Child Pugh classification, must be tested in future prospective studies.
References I Child CG, Turcotte JG. Surgery and portal hypertension. In: Child CG. editors. The liver and mortal hvoertension. Philadelphia: W.B. Saunders Co.; 1964. p. 50-2.
Breath tests for hepatic function
stimulate the likelihood of metabolic pharmacokinetic interactions. Drug interactions. Clin Pharmacokinetic 1997;3:32. Is Matsumoto K, Suehiro M, Ilio M, Kawabe T, Shiratori Y, Okano K, et al. (‘C) Methacetin breath test for evaluation of liver damage Dig Dis Sci 1987;32:344-8. I6 Schneider JF, Schoeller DA, Nemchausky B, Boyer JL, Klein P. Validation of ‘CO2 breath analysis as a measurement of demethylation of stable isotope labelled aminopyrine in man. Clin Chim Acta 1978;84:153-62. I7 Belsley DA. Conditioning diagnostics: Collinearity and weak data in regression. New York: John Wiley & Sons; 1991. I8 Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839-43. I’) Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36. ?” Hehir DJ, Jenkins RL, Bistrian BR, Wagner D, Moldawer LL, Young
VR, et al. Abnormal Phenylalanine hydroxylation and tyrosine oxidation in a patient with acute fulminant liver disease with correction with liver transplantation. Gastroenterology 1985;89:659-63. 21 Hereber M, Talke H, Maier KP, Gerok W. Metabolism of phenylalanine in liver diseases. Klin Wochenschr 1980;58: 1189-96. 22 Desmet VJ, Gerber M, Moofnagle JH, Manns M, Seheur PJ. Classification of chronic hepatitis: diagnosis, grading and staging. Hepatology 1994;19:1513-20. Z3Porayko MK, Wiesner RH. Management of ascites in patients with cirrhosis. What to do when diuretics fail? Postgrad Med 1992;92:156-66. 24 Uribe M, Marquez A, Garcia Ramos G, Ramos-Uribe MH, Vargas F, Villalobos A, et al. Treatment of chronic portal-systemic encephalopathy with vegetable and animal protein diets. A controlled crossover study. Dig Dis Sci 1982;27: 1109- 16. 25 Weber FL Jr, Bagby BS, Licate L, Kelsen SG. Effects of branched chain aminoacids on nitrogen metabolism in patients with cirrhosis. Hepatology 1990; 11:942-50.
24th National Congress of the Associazione Italiana Studio Pancreas (AISP) Capri (Naples, Italy) - September 27-29, 2000
qfh Joint Meeting of Italian-Hungarian
Pancreatologists
Capri (Naples, Italy)
Organization: G. Uomo, PG. Rabitti, Pancreas Unit, Cardarelli Hospital, Naples, Italy. Phone: +39-081-7472208 - Fax: +39-081-5090707 - E-mail: [email protected]
232
OIGESTLIUER OIS 2000;32:226-32
13C-Phenylalanine and 13GMethacetin breath test to evaluate functional capacity of hepatocyte in chronic liver disease S. Lara Baruaue M. Razquin I. Jimenezl A. Vazquezl J. P. Gisbert’ J. M. Pajaresl I
Background. To grade liver damage, Child-Pugh classification is used but these tests do not reflect the quantitative functional hepatic reserve. Aims. 13GPhenylalanine Breath Test and 13GMethacetin Breath Test are evaluated as possible tools, being both safe and easy to perform, to quantify functional hepatic reserve in chronic liver disease patients. Patients. Both tests were performed in 48 healthy volunteers and 48 chronic liver disease patients. Methods. Breath samples were collected after taking 13C-Phenylalanine (100 mg] and 13GMethacetin (75 mg]. 13COp enrichment was measured using mass spectrometry Results. Both tests discriminated the hepatic function, decreasing results of the 13C02 enrichment agreeing with the increasing severity of the hepatic patient (13GPhenylalanine Breath Test multiple correlation coefficient: 0.72, global pcO.001; Methacetin Breath Test: 0.73, p
from:
Departments of Clinical Analysis end I Gastroenterology, Hospital de la Princase, Universidad AutrSnome, Madrid, Spain. AlrkrrrfarE~endenw: Dr. S. Lara Baruque, Servicio An#sis Clinicos. Hospital de le Princasa, c/Diego de Leon 62, 28006 Madrid, Spain. Fex: -04-91-570267. E-mail: [email protected] Submitted September 7, 1999. Revised February 18, 2000. Accepted March 14, 2000.
226
Digest Liver Dis 2000;32:226-32 Key words: ITtion;
chronic
liver
Methacetin disease;
breath test; 13C-Phenylalanine breath liver tests; quantitative liver function
test;
Child-Pugh
classifica-
Introduction In patients with liver disease, several indicators are used to grade liver damage in accordance with the Child Pugh classification I: ascites, encephalopathy, bilirubin, albumin and prothrombin time. Albeit, these tests do not reflect the quantitative functional hepatic reserve, they only show the endpoint hepatocellular damage, the hepatic injury. Child’s classification has been modified in various ways (deletion of nutrition status, addition of age, blood ammonia, number of previous bleeds) with little, if any, improvement.
S. Lara Baruque et al.
Standard liver blood tests provide information as to hepatocyte injury [alanine transaminase (ALT), aspartate transaminase (AST)], and bile formation (alkaline phosphatase activity and bilirubin). With the exception of serum albumin and prothrombin time, these tests do not reveal information concerning the residual hepatic function 2. For example, in a patient with inactive cirrhosis, serum hepatic enzyme levels may be normal in spite of the decreased hepatic function. On the other hand, hepatic function may be preserved in the presence of an extensive hepatocellular necrosis shown by a marked increase in transaminases. Hepatic function failure is, finally, the main cause of mortality in patients with serious liver dysfunction. Several quantitative tests have been proposed to measure residual hepatic function 3 4. Different aspects of liver function may be measured on the basis of the substrate selected. Substrates such as D-sorbitol 5 and galactose 6, that the liver efficiently extracts from blood, are chosen to measure hepatic blood flow. Other substances with a low hepatic extraction ratio, such as phenacetin, methacetin, caffeine, erythromycin and aminopyrine ’ * , have been investigated extensively as a simple and reliable index of hepatic microsomal enzyme reserve. Other substances with an intermediate hepatic extraction rate such as indocyanine green 9 and Lidocaine I0 ” are used, which combine microsomal function study and the functional liver plasma flow. Compared with healthy controls, the clearance of these substances are decreased in individuals with chronic liver disease and cirrhosis. On the other hand, liver is the main organ involved in the metabolism of aromatic aminoacids including phenylalanine, tryptophan and tyrosine I2 13,thus in severe liver disease, the central plasma clearance of these aminoacids is decreased and it has been shown that the central plasma clearance rate of aminoacids by the liver is an accurate predictor of hepatocyte function and reserve. Application of the clearance of these substances to quantify liver function is limited by the need of qualified staff, complexity, and need to obtain blood or body fluids for analysis. The possibility of devising a non-invasive test able to provide information concerning the functional liver cell mass, not requiring repeated blood sampling, indeed a safe test, easy to perform, and cost-effective, encouraged us to carry out this study. The purpose of this study was, therefore, to investigate hepatic cytosolit enzyme activity through the ‘C-Phenylalanine breath test (PBT) and, on the other hand, apply the 13CMethacetin breath test (MBT) to study the hepatic microsomal function. In the breath tests, a tracer molecule (Phenylalanine or Methacetin) is labelled with carbons enriched with the stable isotope of mass 13 (‘“C), which is a naturally occurring isotope. The en--..
richment of 13C02 in the expired air is analysed by an isotope ratio mass spectrometer (IRMS), a piece of equipment which is easy to manage, and the use of which has been increasingly used over the last few years in many hospitals. The most obvious advantage of stable isotopes is that they are non-radioactive and do not present any risk to subjects, and, furthermore, need no blood sampling and time-consumption is limited (one hour). Both non-invasive tests have been evaluated as possible tools to establish the hepatic reserve in patients with liver diseases and, indirectly, also hepatocyte damage.
Patients and methods Patients During a period of eight months, 48 patients (29 males and 19 females, mean age 58.2&13 years, with histologically confirmed liver diseases: 10 cirrhosis Child A, 15 cirrhosis Child B, 12 cirrhosis Child C, and 11 chronic hepatitis) were admitted consecutively to our hospital either as in-patients or out-patients. The viral aetiology (hepatitis B virus or hepatitis C virus) was the most frequent (n=30); alcoholic aetiology (n=16) and a histological diagnosed primary bile cirrhosis completed the hepatic dysfunction aetiology. In one patient, aetiology was unknown. Any history of smoking, alcohol and drug consumption was recorded. The drugs administered to the patients (spironolactone, furosemide, lactulose etc.) have no interactions with the oxidative metabolism of Methacetin by cytochrome PYC lA2 14. Patients with respiratory diseases, which could possibly alter breath function, were excluded, as were patients with any digestive disease which could cause malabsortion. Controls A total of 48 healthy volunteers (22 males and 26 females, mean age 41.7&15 years) were included as control subjects. None of these control subjects had a history of liver, breath or gastrointestinal malabsortion disease, neither were they receiving any medication at the time of the study. Biochemical hepatic function, prothrombin time and platelet count were normal in all subjects. Any history of smoking and alcohol abuse were recorded. Before the study, these subjects were on a normal diet and consumed no drugs for two weeks prior to the test which could affect hepatic enzyme activity. Written informed consent was obtained in healthy volunteers and patients, prior enrolment in this prospective study, in accordance with the clinical assay protocol approved by the Ethics Commission of Clinical Investigation of our hospital. 221
Breath tests for hepatic function
Liver blood tests Standard liver blood tests (serum values of prothrombin time, albumin, total bilirubin, AST, ALT and alkaline phosphatase) were collected before performing the breath test. The diagnosis of cirrhosis was based on histological findings and all patients were classified by the Child Pugh classification which comprises ascites, encephalopathy, serum albumin and bilirubin levels, as well as prothrombin time. Patients were classified according to their score into Class A (score 5-6), Class B (score 7-9) or Class C (score lo- 15) ’ . The mean values of total bilirubin (mg/dl) and the standard deviation (SD) were Child A (1.32+0.82), Child B (2.48&2.44), Child C (4.49k1.93) and chronic hepatitis was (l&0.86). Methacetin and Phenylalanine breath test protocol The two quantitative liver function tests were carried out on different days (within one week) both in the control and patient groups. Breath tests were made following an overnight fast with no control of prior dietary intake. Duplicate baseline breath samples were collected before administration of the 13C-Phenylalanine ‘* l3 and the 13C-Methacetin dose I5 consisting in 100 mg L- (l-13C) Phenylalanine (99% 13C, Cambridge Isotope Laboratories, Andover, MA, USA) and 75 mg (1 -13C) Methacetin (99% 13C, Cambridge Isotope Laboratories, Andover, MA, USA) dissolved in 50 ml of water. Thereafter, duplicate breath samples were collected every 10 minutes for 1 hour. The 13C02 enrichment, expressed as cumulative percent oxidation of dispensed dose was measured with a stable isotope mass spectrometer (Europe Scientific Tracermass). Analytical data were expressed as percentages of 13C02 recovery per hour using an area under the curve (AUC) method and assumed COZ production rate of 5 mmol/min/m2 body surface area, as described by Schneider et al. 16. Statistical analysis A 95% confidence interval (CI) was calculated. Multiple linear regression was performed. The dependent variables were the cumulative percent oxidation of Phenylalanine or Methacetin, and independent variables were age, sex, tobacco (smokers and non smokers), alcohol (less or more than 40 g ethanol/day) and hepatic function (classified as healthy volunteers, chronic hepatitis, cirrhosis Child A, Child B and Child C). To study the differences between each group, the hepatic function was coded as dummy binary variables according to the coding scheme that uses the following level as reference level (e.g., healthy volunteers vs chronic hepatitis, chronic hepatitis vs cirrhosis Child A . . .). This statistical analysis is particularly indicated for increasing or decreas228
ing stair-like results. We used a backward modelling strategy. The F partial was the statistic used for comparing models and the significant level was set at 0.05. Changes over 10% in the regression coefficient were considered as suggestive of “confusion”. Belsley criteria l7 were followed to diagnose “collinearity”, which was not demonstrated. Finally, the correlation between PBT and MBT was studied using linear regression. The areas under the ROC (receiver operating characteristic) curves from the various cut-off points for PBT and MBT were calculated; the value of the maximal possible (ideal) area is one ‘* 19. ROC curves were calculated considering all collecting times (1 hour), and also for each collecting time (every 10 minutes). Results
As can be seen in Figure 1 both the PBT and MBT test discriminated the hepatic function, the decreasing results of the 13C02 correlating with the increasing severity of the hepatic patient graded according to Child classification and histological findings. The correlation between the cumulative percent PBT or MBT oxidation and hepatic function by multiple linear regression is depicted in Figure 2. Results of multiple linear regression analysis for cumulative percent Phenylalanine oxidation are summarized in Table I, where it can be observed that the only variables which correlated with Phenylalanine oxidation were age and hepatic function. The PBT discriminated between all the different groups except chronic hepatitis vs Child A cirrhosis. The highest regression coefficient values were for healthy vs hepatitis, and Child B vs Child C cirrhosis. Results of multiple linear regression analysis for cu-
El
0 Healthy Volunteen (1~48) OChronic Hepatitis (~11) ~~Cirrhosis Child A (140) SCimhosis Child B (n=15) N3rrhcsis Child C (1142)
Fig. 1. Phenylalanine Breath Test and Methacetin lative percent oxidation in healthy volunteers tients. Data presented as mean f SD.
Breath Test: Cumuand liver disease pe-
“--l__l
S. Lara Baruque et al.
IWe II. Correlation between the cumulative percent Methacetin oxidation and hepatic function by multiple linear regression. VPrialfle"
Repooien coef8cient
Healthy volunteers Chronic hepatitis Cirrhosis Child A Cirrhosis Chid 8 *Coded healthy Child A ficient:
Fig. 2. Correlation between the cumulative percent Phenylalanine or Methacetin oxidation and hepatic function by multiple linear regresion. The discontinuous lines indicate the 95% confidence interval.
I
I. Correlation between the cumulative percent Phenylalanine oxidation and hepatic function by multiple linear regression. Table
lbtriable 0.04 1.91 0.36 1.47 2.04
Age
Healthy volunteers” Chronic hepatitis” Cirrhosis Child A Cirrhosis Child El” ‘Coded healthy Child A ficient:
0.01 0.68 0.87 0.82 0.69
0.0106 0.0060 0.6856 0.0717 0.0038
as dummy va-iables, using as reference the following level (e.g., volunteers vs chronic hepatitis, chronic hepatitis vs cirrhosis e&l. Multiple correlation coefficient: 0.72 (determination coef52%); global p
3.98 0.78 4 5.3
Stsndsrd error
P
1.7 2.2 2.1 I.8
0.0196 0.7249 0.0603 0.0049
as dummy variables, using as reference the following level [e.g., volunteers vs chmnic hepatitis, chronic hepatitis vs cirrhosis etc.]. Mu/tip/e correlation coefficient: 0.73 ldatermination coef54%1 global pcO.001.
mulative percent Methacetin oxidation are summarized in Table II. MBT discriminated between all the different groups except chronic hepatitis vs Child A cirrhosis. The highest regression coefficient values were for healthy vs hepatitis, and for Child B vs Child C cirrhosis. The correlation coefficient between PBT and MBT was 0.63 (p
lhbla Ill. Evaluation of areas under ROC [receiver operating characteristic1 curves for Phenylalanine Breath Test and Methecetin Breath Test in the evaluation of hepatic function depending on collecting time.
P8T cellootitlg time
At-08
SE
IO min. 20 min. 30 min. 40 min. 50 min. 60 min. Total time
0.75 0.82 0.85 0.84 0.81 0.74 0.85
0.05 0.04 0.04 0.04 0.04 0.05 0.04
PST: WPhenylalanine
breath
tast;
MST: W-Methacetin
MST collecting
time
IO min.
20 min. 30 min. 40 min. 50 min. 60 min. Total time breath
test;
SE: Standard
Error;
Area
SE
0.86 0.86 0.82 0.81 0.80 0.78 0.84
0.84 0.04 0.04 0.04 0.05 0.05 0.04
Area: Area under ROC curve.
229
Breath tests for hepatic function
lb& WL Evaluation of ‘areas under ROC keceiver operating characteristic1 curvea and different cut-off points for Phenylalanine Breath Test in the evaluation of hepetic function.
Total
Healthy volunteers vs any hepatic dysfunction
0.85
0.04
5.5
0.75
0.84
4.5
Total
Healthy volunteers or chronic hepatitis vs cirrhosis
0.87
0.04
5
0.76
0.87
6
Total
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.90
Cl.04
4.92
0.89
0.86
6.5
30 min
Healthy volunteers vs any hepatic dysfunction
0.85
0.04
1.19
0.77
0.87
5.8
30 min
Healthy volunteersor chronic hepatitis vs ciwttosis
0.84
0.04
1.07
0.73
0.87
5.9
30 min
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.88
0.04
1.10
0.89
0.83
5.3
Tateh considering ttW ckmh%Him insteed cirrhossl:
b
all collecting times II hour]; Aree: Area under ROC curve; SE: Standerd Error; Cut-OR cut-off point for Phenytalenina Breath Test value in of hepetic function as one of the two categories [higher then cut-off point va{ua is indicetive of lass severe condition, e.g. healthy volunteers LR: Likabod ratio.
Table K Evaluation of areas under ROC kaceiver operating characteristic) curves and different cut-off points for Methacetin Breath Test in evaluation of hepatic function.
Total
Healthy volunteers vs any hepatic dysfunction
0.84
0.04
13.1
0.71
0.94
11.4
Total
Healthy volunteers or chronic hepatitis vs cirrhosis
0.86
0.04
11.7
0.78
0.95
15.3
Total
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child 8 or Child C
0.90
0.04
11.7
0.87
0.88
7.5
IO min
Healthy volunteers vs any hepatic dysfunction
0.86
0.04
0.93
0.79
0.87
6.3
IO min
Healthy volunteers or chronic hepatitis vs cirrhosis
0.89
0.04
0.90
0.86
0.83
5.2
IO min
Healthy volunteers or chronic hepatitis or cirrhosis Child A vs cirrhosis Child El or Child C
0.89
0.04
0.65
0.84
0.83
4.8
Total: considering the classification instead cirrhosisl;
all collecting times I4 hour); Area: Area under HOC curve; SE: Standard Error; Cut-off: cut-off point for Methacatin Breath Test value in of hepetic function as one of the two categories [higher than cut-off point value is indicative of lass severe conditions e.g. healthy volunteers LR: Likelihood ratio.
The area was over 0.84 in all cases, with a likelihood ratio constantly over 4.5. The evaluation of areas under ROC curves and different cut-off points for MBT both considering all the colletting times and also the 10 min collecting time (the best one, as previously described), are summarized in Table V. The area was, in all cases, over 0.84, with a likelihood ratio constantly over 4.8. 230
Discussion Results emerging from the present study show that both PBT and MBT are able to discriminate the hepatic functional capacity between the different groups studied. As liver function becomes more severely impaired, as defined by Child Pugh classification I, so does hepatocyte metabolism (Fig. 1): both Phenylalanine and Methacetin metabolism are decreased in liver
--
S. Lara Bsruqua et al.
dysfunction. Phenylalanine aminoacid metabolism is blocked in liver disease so that phenylalanine hydroxylase, in these patients, is estimated to be 20% of the normal value 2o 2’ Methacetin, a derivative of phenacetin, undergoes a rapid 0-dealkylation by hepatic microsomal enzyme systems 15. The fact that the highest discriminatory capacity is for Child B vs Child C, could be interesting for monitoring the progression of liver dysfunction, predicting risk of surgical interventions and for determining the appropriate timing for liver transplantation. The fact that none of the tests discriminate between chronic hepatitis patients and Child A cirrhosis is easily explained due to the fact that hepatic function and biochemical data can be similar and a histological study is often needed to differentiate between these two stages 22. These tests are not affected by diet, which may interfere in the biochemical and clinical data used for hepatic dysfunction diagnosis in the Child Pugh score; e.g. the prior intake of salt may artificially increase the degree of ascites without really affecting the hepatic function 23, or a protein diet may artificially increase the encephalopathy 2425. Comparing the PBT and MBT tests, it can be seen that no important differences exist as far as concerns the yield of hepatic function information between these two tests. The area under ROC curves in always >0.84 with a likelihood ratio >4.5 in both tests (Tables IV, V). The advantage of the MBT is that only one breath collection is needed at 10 minutes (however, the PBT best collection time is at 30 minutes), so that MBT is quicker to perform (Table III). On the other hand, MBT values, in contrast with those of PBT, are not influenced by age (Tables I, II). As far as concerns carrying out both tests in the same patient, it can be seen that if both tests are positive, the sensitivity in the diagnosis is very high (95%). Therefore, if we need to use these breath tests as a screening test, it would be useful to perform both tests, although specificity will then be decreased. Today, the Child Pugh classification remains the most widely accepted tool as a measure of disease severity in liver disease. However, in addition to being safe, non-invasive, inexpressive, easy to perform, and well correlated with the Child Pugh score, PBT and MBT could provide additional information in that direct hepatocyte functional capacity is measured, avoiding any subjectivity in hepatic dysfunction staging (the degree of ascites is not always easy to establish) and also avoiding other elements that could modify the real degree of hepatic damage (salt intake and degree of ascites, or protein intake and degree of encephalopathy). The function tests can provide different information from obtained by a liver biopsy, but not such detailed information on hepatic disease as that provided by histological examination. Furthermore, function tests can-
not indicate the extent of inflammation or the degree of fibrosis. In summary, the conclusion of our study is that both PBT and MBT tests, which are safe and easy to perform, are able to discriminate hepatic functional capacity not only between healthy individuals and patients with liver disease, but also between different stages of hepatic dysfunction. The possibility that PBT and MBT could be useful to predict the prognosis of liver disease and the rate of disease progression, compared with Child Pugh classification, must be tested in future prospective studies.
References I Child CG, Turcotte JG. Surgery and portal hypertension. In: Child CG. editors. The liver and mortal hvoertension. Philadelphia: W.B. Saunders Co.; 1964. p. 50-2.
Breath tests for hepatic function
stimulate the likelihood of metabolic pharmacokinetic interactions. Drug interactions. Clin Pharmacokinetic 1997;3:32. Is Matsumoto K, Suehiro M, Ilio M, Kawabe T, Shiratori Y, Okano K, et al. (‘C) Methacetin breath test for evaluation of liver damage Dig Dis Sci 1987;32:344-8. I6 Schneider JF, Schoeller DA, Nemchausky B, Boyer JL, Klein P. Validation of ‘CO2 breath analysis as a measurement of demethylation of stable isotope labelled aminopyrine in man. Clin Chim Acta 1978;84:153-62. I7 Belsley DA. Conditioning diagnostics: Collinearity and weak data in regression. New York: John Wiley & Sons; 1991. I8 Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839-43. I’) Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36. ?” Hehir DJ, Jenkins RL, Bistrian BR, Wagner D, Moldawer LL, Young
VR, et al. Abnormal Phenylalanine hydroxylation and tyrosine oxidation in a patient with acute fulminant liver disease with correction with liver transplantation. Gastroenterology 1985;89:659-63. 21 Hereber M, Talke H, Maier KP, Gerok W. Metabolism of phenylalanine in liver diseases. Klin Wochenschr 1980;58: 1189-96. 22 Desmet VJ, Gerber M, Moofnagle JH, Manns M, Seheur PJ. Classification of chronic hepatitis: diagnosis, grading and staging. Hepatology 1994;19:1513-20. Z3Porayko MK, Wiesner RH. Management of ascites in patients with cirrhosis. What to do when diuretics fail? Postgrad Med 1992;92:156-66. 24 Uribe M, Marquez A, Garcia Ramos G, Ramos-Uribe MH, Vargas F, Villalobos A, et al. Treatment of chronic portal-systemic encephalopathy with vegetable and animal protein diets. A controlled crossover study. Dig Dis Sci 1982;27: 1109- 16. 25 Weber FL Jr, Bagby BS, Licate L, Kelsen SG. Effects of branched chain aminoacids on nitrogen metabolism in patients with cirrhosis. Hepatology 1990; 11:942-50.
24th National Congress of the Associazione Italiana Studio Pancreas (AISP) Capri (Naples, Italy) - September 27-29, 2000
qfh Joint Meeting of Italian-Hungarian
Pancreatologists
Capri (Naples, Italy)
Organization: G. Uomo, PG. Rabitti, Pancreas Unit, Cardarelli Hospital, Naples, Italy. Phone: +39-081-7472208 - Fax: +39-081-5090707 - E-mail: [email protected]
232