Dynamic carbon 13 breath tests for the study of liver function and gastric emptying

In gastroenterological practice, breath tests (BTs) are diagnostic tools used for indirect, non-invasive assessment of several pathophysiological metabolic processes, by monitoring the appearance in breath of a metabolite of a specific substrate. Labelled substrates originally employed radioactive carbon 14 (14C) and, more recently, the stable carbon 13 isotope (13C) has been introduced to label specific substrates. The ingested 13C-substrate is metabolized, and exhaled 13CO2 is measured by mass spectrometry or infrared spectroscopy. Some 13C-BTs evaluate specific (microsomal, cytosolic, and mitochondrial) hepatic metabolic pathways and can be employed in liver diseases (i.e. simple liver steatosis, non-alcoholic steato-hepatitis, liver fibrosis, cirrhosis, hepatocellular carcinoma, drug and alcohol effects). Another field of clinical application for 13C-BTs is the assessment of gastric emptying kinetics in response to liquids (13C-acetate) or solids (13C-octanoic acid in egg yolk or in a pre-packed muffin or the 13C-Spirulina platensis given with a meal or a biscuit). Studies have shown that 13C-BTs, used for gastric emptying studies, yield results that are comparable to scintigraphy and can be useful in detecting either delayed- (gastroparesis) or accelerated gastric emptying or changes of gastric kinetics due to pharmacological effects. Thus, 13C-BTs represent an indirect, cost-effective and easy method of evaluating dynamic liver function and gastric kinetics in health and disease, and several other potential applications are being studied.


Introduction
Breath tests (BTs) are diagnostic tools based on the ingestion of various substrates that are processed at different levels in the gastrointestinal tract [1]. The principle of BTs relies on the concept that the metabolized substrate leads to the production of gases (e.g. CO 2 , H 2 ) that pass into the blood, are excreted and quantified in expired air. The interest towards the use of BTs has increased since they are relatively simple, safe and noninvasive tools with potential applications in several clinical conditions including liver diseases, H. Pylori infection, gastrointestinal motility, small intestinal bacterial overgrowth, and sugar (fructose, lactose) malabsorption. BTs using specific substrates labelled with the stable (non-radioactive) isotope 13 C have been also employed for the assessment of hepatic functional reserve and to measure gastric emptying. The potential applications of such 13 C-BTs will be discussed in the present paper. 13 C breath tests for the study of liver function

General features
A major challenge in clinical hepatology is the accurate and non-invasive assessment of liver function in patients with chronic liver disease. The evaluation of liver status is currently based on serum parameters of synthesis (prothrombin, cholesterol, albumin), hepatocellular integrity (transaminases), detoxification (ammonium), excretion and cholestasis (bilirubin, alkaline phosphatase, cGT), associated with imaging techniques (ultrasonography, computerized tomography, magnetic resonance and elastography) or scores such as the Child-Turcotte-Pugh classification, which combines clinical (ascites and degree of encephalopathy) and serum (bilirubin, albumin, and prothrombin time) parameters [2]. Such 'static' tests of liver function may have limitations, especially when dealing with prediction of outcomes and in assessing liver dysfunction in critically ill patients [3]. In this respect, BTs represent novel indirect 'dynamic' tools that provide additional insights in functional diagnosis and follow-up of patients with liver diseases [4].
The principles of BTs in hepatology are based on both biochemical and pharmacological considerations. Mechanisms of liver damage often include dysfunction of subcellular organelles, such as microsomal hypertrophy, mitochondrial abnormalities, and activation of peroxisomal metabolism (i.e. long chain fatty acids). Thus, assessing specific functions of such organelles through BTs may provide useful information to clinicians. Also, BTs allow the study of specific time-dependent metabolic processes by assessing the hepatic clearance of metabolically active substances [5]. In this context, for a given exogenous substrate: (where HEPATIC EXTRACTION is the ratio of the difference between inflow and outflow concentration Ä by inflow concentration of the probe) [6].
Hepatic clearance is defined as flow-limited (range 0.7-1.0) or enzyme-limited (<0.3) [7]. The general characteristics of the 'ideal' substrate for the study of liver function are depicted in Table 1 [8].
The intrinsic complexity of liver metabolic pathways does not allow a single functional test to explore the whole liver function. Different substrates are therefore used to assess cytosolic, microsomal or mitochondrial function ( Figure 1). Such substrates are marked with the natural stable isotope of carbon, carbon 13 ( 13 C; currently the most widely used isotope). After intestinal absorption, the given substrate undergoes liver metabolism at different levels, which ultimately results in the production and appearance of 13 CO 2 in exhaled air, as a marker of specific liver metabolic functions [4,9] (Figure 2).

Breath tests for the study of liver cytosolic function
Phenylalanine is an aromatic amino acid, converted to tyrosine in the cytosol of the hepatocytes by the enzyme phenylalanine hydroxylase. Phenylalanine oxidation is affected when liver function is impaired and by using a 13 C-phenylalanine BT, a reduced 13 CO 2 release in the exhaled breath is recorded [10,11]. The 13 C-phenylalanine BT has been used to predict postoperative complications and to monitor liver regeneration after partial hepatectomy, and to predict the severity of liver cirrhosis, staged according to the Child-Turcotte-Pugh score [12]. However, in a study by our group that compared the hepatic functional mass in chronic liver disease classified according to Child-Turcotte-Pugh score and serum bile acid levels, the diagnostic power of 13 C-phenylalanine BT was less than that of 13 C-methacetin BT [13].
Galactose is a carbohydrate primarily metabolized by the liver. The metabolism of galactose investigates the activity of galactose kinase, which catalyses the ATP-dependent phosphorylation of galactose to galactose 1-phosphate. Originating from the radioactive carbon 14 ( 14 C)-galactose BT, the nonradioactive 13 C-galactose BT has been also developed [14]. The performance of the 13 C-galactose BT relates inversely with the severity of liver disease-and in particular with cirrhosisreaching 71% sensitivity, 85% specificity, and 84% accuracy [15,16]. Combination with another BT investigating the microsomal function (aminopyrine) further increases the diagnostic sensitivity and specificity [15]. The 13 C-Galactose BT has also been used to assess progressive decline in liver mass function in patients with chronic liver disease due to Hepatitis C virus (HCV) infection [17].

Breath tests for the study of liver microsomal function
Aminopyrine is known as an antipyretic analgesic drug. The 13 C-labelled methyl groups of aminopyrin are demethylated with transformation into formate, formaldehyde, and the produced bicarbonate releases 13 CO 2 which is exhaled in breath. After oral administration, aminopyrine is completely absorbed and has a low hepatic extraction (E ¼ 0.2), seen to be independent of liver blood flow and not disturbed by hepatic vascular shunts. Aminopyrine is metabolized by the hepatic microsomal cytochromes and provides an index of functional hepatic mass. Aminopyrine BTs have been used to provide prognostic information, to study, graft rejection, severity of paracetamol intoxication and to detect alcoholic liver injury. When used to discriminate the degree of liver fibrosis, aminopyrine BT displayed good specificity and sensitivity for advanced fibrosis or Table 1. General characteristics required of an ideal substrate for studying dynamic liver function Pharmacokinetic and metabolic aspects Rapidly and consistently absorbed when administered orally Primarily metabolized in the liver Low (20-30%) hepatic extraction ratio (i.e. metabolism independent from liver blood flow) Clear metabolic pathway; simple pharmacokinetic; short elimination half-life Figure 1. Sites where metabolic processes may be explored by breath test in hepatocytes. In particular, 13 C-a-ketoisocaproic acid, 13 C-methionine, and 13 C-octanoate are the three substrates more widely employed for the dynamic assessment of mitochondrial function (see text for details). cirrhosis but performed poorly for intermediate fibrosis stages [7,18,19]. Some limitations need to be considered in the interpretation of aminopyrine BT: (i) the age of the subject can affect the P450-dependent N-demethylation, (ii) the simultaneous use of drugs with enzyme inductive effect or inhibiting the microsomal enzymatic activity might alter the findings of the BT, and (iii) female sex hormones can produce a negative effect on aminopyrine metabolism [6,20].
Methacetin is a derivative of phenacetin which is metabolized rapidly by the hepatic microsomal enzyme systems CYP1A2 into acetaminophen and 13 CO 2 by a single O-dealkylation step. Since methacetin has a high extraction (E > 0.8) [21] and undergoes extensive first-pass clearance, its metabolism can be altered by hepatic blood flow alterations and by hepatic 'first-pass' effect. Capacity for metabolizing methacetin is lower in elderly people than in other adults [22]. Methacetin BT was shown to accurately assess the degree of liver damage in patients with histologically proven chronic liver diseases and to distinguish chronic aggressive hepatitis from liver cirrhosis; it also distinguished early cirrhosis (Child A) from non-cirrhotic patients. Methacetin BT was a useful predictive marker of clinical outcomes in chronic HCV patients [23][24][25]. Moreover, methacetin BT can better estimate the degree of fibrosis in patients with chronic HCV infection than biochemical parameters (i.e. aspartate aminotransferase-to-platelet ratio and aspartate aminotransferase-to-alanine aminotransferase ratio) or Fibroindex [26,27].

Breath tests for the study of liver mitochondrial function
Ketoisocaproate (KICA) is an intermediate in the metabolism of leucine. The decarboxylation of KICA and the generation of CO 2 reflects the mitochondrial branched-chain amino acid decarboxylation function [28]. This step is observed when transamination to leucine (the major competing pathway for KICA elimination) is suppressed by the simultaneous administration of fixed doses of leucine ( Figure 3a). This metabolic pathway of KICA has been tested in experimental models, in isolated mitochondria, in healthy subjects treated with acetyl salicylic acid or with low ethanol intake, and in patients with liver diseases [8,9,28,29]. 13 C-KICA decarboxylation is lower in alcoholic patients than in patients with non-alcoholic fatty liver disease (NAFLD) or controls [29,30]. We found that the mitochondrial decarboxylation capacity of KICA was lower in patients with advanced non-alcoholic steatohepatitis (NASH) than in healthy subjects and patients with simple liver steatosis. Notably, the 13 CO 2 cumulative recovery values following 13 C-KICA were inversely related to the extent of fibrosis, to serum hyaluronate, and to body size in NASH patients [31]. We extended the studies with 13 C-KICA BTs and found that KICA decarboxylation was significantly lower in cirrhotic patients with hepatocellular carcinoma (HCC) than in cirrhotic patients without HCC and identical Child-Pugh score. Moreover, KICA decarboxylation was deranged following radiofrequency ablation, but not after transarterial chemoembolization. Finally, the recurrence of HCC was associated with an early decrease of KICA decarboxylation [32]. In a different context, we recently found that a 13 C-KICA BT was abnormal (and therefore suggesting mitochondrial malfunction) in a female patient suffering from massive liver echinococcosis. Notably, mitochondrial liver function improved following pericystectomy and limited hepatectomy [33]. KICA BT is useful for the assessment of drug effects on liver mitochondrial function. Liver injury might occur following the use of such drugs, which accumulate in the mitochondria and interfere with respiratory complexes or electron transfer [34,35]. KICA BT may be helpful in ascertaining the integrity of these organelles before the administration of potentially toxic drugs and in detecting druginduced mitochondrial damage before the appearance of symptoms, in order to manage patients in a timely manner and prevent adverse effects. Examples are tacrolimus, aspirin, and ergot alkaloids. There are also potential applications with amiodarone, valproate, and retroviral drugs [36,37].
Methionine is an essential amino acid, principally metabolized in the liver with production of CO 2 [38]. 13   c) octanoic acid. CO2 is invariably produced at the end of the process. The use of 13 C-labelled substrates ultimately leads to production of 13 CO2 following mitochondrial metabolism. (Figure 3b). In human studies, 13 C-methionine BT has been used to assess mitochondrial function during acute intoxication and in patients with chronic liver diseases. For example, acute ethanol consumption impairs 13 C-methionine decarboxylation in healthy volunteers [39], whereas metabolism of methionine is decreased in patients with liver cirrhosis, and especially in those with ethanol aetiology, in patients with NAFLD, in those taking high-dose valproate or nucleoside analogues for the treatment of HIV, and in patients with Friedreich ataxia [37,[40][41][42][43][44]. A defective methionine metabolism has also been reported in hepatitis C-infected cells [45].
Octanoate is a medium chain fatty acid that enters mitochondria independently of the carnitine transport system. Within the mitochondria, octanoate undergoes b-oxidation, which generates acetyl co-enzyme A (AcCoA). AcCoA enters the Krebs cycle and is oxidized to CO 2 unless utilized for the synthesis of other energy-rich compounds (Figure 3c). The 13 C-octanoate BT has been employed for the study of gastric emptying (see below) but the test should also reflect hepatic mitochondrial function when gastric emptying and duodenal absorption are not severely deranged. Octanoate is therefore a potential substrate for non-invasive BT of hepatic mitochondrial b-oxidation. In fact, in animal models, the decarboxylation of octanoate was decreased in rats developing thioacetamideinduced acute hepatitis and liver cirrhosis [46]. In patients with NASH, the oxidation of octanoate was unchanged or increased [47,48], and unchanged in those with early stage and advanced cirrhosis with and without porto-systemic shunt [49]. 13 C breath tests for the study of gastric emptying

General features
A subgroup of patients may develop disturbed gastric motility disorders, namely accelerated ('dumping syndrome') or delayed gastric emptying, which is the most frequently investigated condition. Gastroparesis implies a delayed gastric emptying without mechanical obstruction and is a condition frequently associated with symptoms such as nausea, vomiting, bloating, early satiety, and or upper abdominal pain [50]. After excluding mechanical obstruction by means of upper endoscopy, computed tomography, barium follow-through examination or magnetic resonance enterography, the diagnosis of gastroparesis is based on functional studies to assess gastric motility in response to liquids and/or solid test meals. Investigations include the traditional scintigraphic studies [50], functional ultrasonography [51,52], magnetic resonance imaging and, recently, wireless motility capsule [53] and BTs (see below). Major limitations of such techniques consist of radiation hazards (scintigraphy), duration of the exam and experience of the operator (ultrasonography), and costs (magnetic resonance imaging and wireless capsule). To establish the aetiology of gastroparesis a further step should take into account the idiopatic form (affecting about half of the patients), diabetes mellitus, and post-surgical, viral, neurological, autoimmune, viral causes, scleroderma, as well as effect of medications (e.g. calcium channel blockers, dopamine antagonists, octreotide, etc.). To establish the aetiology, laboratory tests, gastroduodenal manometry, autonomic testing and additional studies-including single photon emission computed tomography and full thickness biopsies of the stomach and small intestine-might be necessary from case to case.

Breath tests for the study of gastric emptying
Gastric emptying has also been investigated by BTs using the radioactive isotope 14 C and more recently, the non-radioactive stable isotope 13 C, both labelling substrates dissolved in liquids (acetate in water or fruit juice) or solid test meals (octanoic acid in egg yolk). Octanoic acid, a medium-chain fatty acid fully retained in egg yolk during mixing and grinding in the stomach, is rapidly liberated in the duodenum and quickly transported to the liver via the vena portae. The 13 C-octanoic acid rapidly enters the mitochondria without requiring carnitine (see also above: liver BTs) and is rapidly oxidized and metabolized to 13 CO 2 . The assumption of the 13 C-BT is that small intestinal absorption of the substrate, liver metabolism, and pulmonary excretion of 13 CO 2 must be preserved and rapid. Thus, a delay in the appearance of 13 CO 2 in breath samples is exclusively due to a reduced solid transit pace between the stomach and the duodenum, i.e. gastric emptying becomes the rate-limiting step when the substrate is rapidly absorbed. An important diagnostic parameter is the half-emptying time, i.e. the time at which half of the solid meal has moved from the stomach to the duodenum [54][55][56][57][58][59]. Breath samples are taken at baseline and every 30 min over a period of 240 min, using a non-linear regression formula [55] or, less frequently (e.g. at 45, 150 and 180 mins) during a period of 180 min using the generalized linear regression model [60] of the 13 CO 2 excretion-time curve. Measurement of exhaled 13 CO 2 is performed by mass spectrometry or infrared spectroscopy. Smoking, physical exercise, drinking and eating are prohibited during the test, and gastric motility-altering drugs are not allowed. The retention of 13 C-octanoic acid in the solid meal should be confirmed by in vitro incubation studies [57,61].
Several researchers have employed the *C-BTs (where * denotes the variable isotope number) for the study of gastric emptying, as summarized in Table 2. Results vary according to the number of healthy subjects and type of patients enrolled, methodology and type of *C isotope, test meal characteristics (liquid, solid, calorie content or preparation), and interpretation of results. Two mathematical models-the non-linear regression model [55] and the general linear regression model [62]have been proposed for the calculation of the BT half-emptying time, starting from the time-dependent 13 CO 2 excretion curve. Several studies have compared BT with scintigraphy and a few have performed studies investigating the absorption kinetics of *C-octanoic after duodenal instillation or the retention of octanoic in specific meals.
The results depicted in Table 2 show a wide range of subjects/ patients ranging in number from 6 to more than 130 and different outcomes in both health and disease. For example, for a liquid test meal, 13 C-acetate yielded a mean half-emptying time of 78 6 14 min in a study with 20 healthy subjects (coefficient of variation 10-36%) and showed a highly reproducible positive linear correlation with technetium-99 ( 99m Tc)-albumin colloid scintigraphy [64]. Data were confirmed in subsequent studies [63,65].
In patients with functional dyspepsia, the mean half-emptying time was significantly delayed ( Table 2) and showed a greater inter-individual variability [63,64]. The authors concluded that the 13 C-acetate BT is a reliable, widely applicable diagnostic tool to assess the gastric emptying of liquids and liquid phases in semi-solid test meals.
Starting from a seminal study in healthy subjects and dyspeptic patients, Ghoos et al. used 14 C-and 13 C-OA in egg yolk to validate the BTs according to the non-linear regression model, providing three kinetic parameters: half-emptying time, lag phase and gastric emptying coefficient [55]. Inter-individual and day-to-day variabilities were acceptable. The normal range of gastric emptying in healthy subjects was 72 6 22SD min. Further studies have used BTs during pharmacologically-induced accelerated or delayed gastric emptying [72,73].
When validating the 13 C-Spirulina platensis solid test meal in 30 healthy volunteers, the Mayo Clinic group found optimal correlation vs. scintigraphy, even when restricting the measurements to a few time points (i.e. baseline, 75, 90 and 180 min, i.e. the Mayo 'reduced model') [60]. The half-emptying time was 100 6 20SD min the solid meal and 91 6 15 min with the S. platensis biscuit. In a subsequent study in 57 healthy subjects, simulated disturbances of gastric emptying were induced by either accelerating (using i.v. erythromycin) or delaying (using i.v. atropin) gastric emptying. As compared with scintigraphy and normal half-emptying values (95 6 24 min) the sensitivity and specificity of 13 C-Spirulina platensis BT using the solid meal were 86% and 80%, respectively, in detecting either accelerated (71 6 14 min) or delayed emptying (207 6 44 min), using a normal range for half-emptying time of 70-150 min [68].
In the multicentre Italian study employing the pre-packed solid meal EXPIROGer V R , we recently found that the reference range of half-emptying time in 131 healthy subjects was 88 6 29SD min with a normal upper cut-off value of 146 min. There was no significant difference between subjects sorted by sex or age. The within-subject variability of half-emptying time was 17%. By contrast, the half-emptying values were significantly delayed, showing 179 6 50 min in diabetic patients with gastroparesis and 151 6 20 min in untreated celiac disease patients [57]. In the initial series of healthy subjects using the novel three-muffins test, the 13 C-octanoic BT yielded a half-emptying time of 108 6 18SD (Portincasa et al.: unpublished observations) [70].
Using the 13 C stable isotope represents a simple, non-invasive, office-or field-based technique [60,74]. The accuracy of BTs for the study of gastric emptying appears to be comparable with scintigraphic studies [60,64,69,73], but the technique, although easy to perform, fully non-invasive, and extendable also to children and pregnant women, represents an indirect method of assessing gastric emptying. In other words, variability in the rate of absorption, metabolism, and excretion of the marker between individuals must be considered in the interpretation of data. Imprecision with both stable isotope and scintigraphy in measuring gastric emptying, however, does reflect pathophysiological variations. The use of BTs for the study of gastric emptying appears promising for intra-individual comparisons. Further studies, however, might further improve standardization of the methodology in terms of statistical analysis, time-test, and sampling frequency [57,60,66,75,76].

Conclusions and future perspectives
BTs represent valuable diagnostic non-invasive tools for in vivo assessment of various enzyme activities, transport processes, or organ functions. By monitoring the metabolization of several 13 C-labelled substrates and the consequent appearance of 13 CO 2 in breath, it is possible to study the liver function dynamically (focussing on different intracellular pathways) as well as gastric emptying kinetics (similarly to scintigraphy, regarded as the 'gold standard'). BTs currently use the non-radioactive stable isotope carbon 13 C as tracer, which is safe in children, during pregnancy and while breast-feeding. BTs are simple, nonradioactive office-or field-based tools which, in epidemiological and pharmachodynamic studies, can better define pathophysiologically relevant abnormalities in the fields of hepatology and gastroenterology. Interpretation of BTs requires a knowledge of methodological limitations and potential pitfalls before routinely extending such studies in the clinical setting.
Conflict of interest statement: none declared.