Reproducibility of the
Methionine Breath Test
The objective of this study was to assess variability
of the methionine breath test by administering the test
twice within 7 days to healthy volunteers and patients
with previously diagnosed liver disease. The study included
20 healthy volunteers, 11 females and 9 males, with a
mean age of 33 years (range 18-59). The 22 liver patients
included (10 females and 12 males) and the mean age was
51 years (range 26-71). The liver patients included 11
patients with stable cirrhosis with no alcohol intake
for > 1 year, 5 actively drinking alcoholic cirrhotic
patients, and 6 patients with cirrhosis and alcoholic
hepatitis. The methionine breath test was performed while
participants were at rest after an 8-hour fast. Reproducibility
of the tests was evaluated using a paired Student’s
t-test. A linear correlation coefficient was also calculated
for the two tests. The null hypothesis was that the difference
between the two tests was equal to zero. The null hypothesis
was tested at p < 0.05.
Reproducibility of the replicate tests was excellent
as the paired t-test accepted the hypothesis that no
difference was detectable between the tests (p=0.735).
There was a significant correlation (r = 0.966, p<0.001)
between methionine breath test 1 and test 2. The correlation
plot is given in figure 1. These results show the reproducibility
of the methionine breath test over a broad clinical
range of liver function.
Figure: Methionine Breath Test 1 vs Test 2 data for
healthy controls and liver patients.
We found that the reproducibility
of the methionine breath test was excellent in 20 volunteers
as well as in the 20 patients with liver disease. Short-term
reproducibility proved satisfactory, as differences
between test 1 and test 2 were not significantly different
from zero at 95% confidence with a paired t-test. Correlation
analysis yielded (r = 0.966, p<0.001) across the
entire clinical range.
Modulation of Mitochondrial Function
We explored how the methionine breath
test responded to agents that effect mitochondrial function.
Alcohol and aspirin (ASA) have a pronounced effect on
changing the NADH/NAD ratio in the mitochondria. How
the change of the NADH/NAD ratio affects methionine
metabolism is illustrated in figure 2.
Figure 2: Effect of alcohol and Aspirin on methionine
metabolism
When alcohol or aspirin change the
NAD/NADH ratio this either decreases or increases methionine
oxidation, respectively. When alcohol is metabolized
to acetaldehyde, NAD is reduced to NADH and thus NAD
is not available for the metabolism of 13C-methionine.
Therefore, 13CO2 production is
decreased. When ASA is metabolized, NADH is oxidized
to NAD making more NAD available for methionine metabolism.
Therefore, 13CO2 production is
increased.
Initially 20 healthy volunteers were evaluated with
the methionine breath test. Then, the next day, participants
were randomized to ingest either alcohol (n=9) or aspirin
(n=11) 15 minutes prior to the methionine breath test.
Subjects ingested 60 cc (2 oz) vodka (Smirnoff, 86 proof)
dissolved in 200 ml orange juice. The dose of aspirin
ingested was 30 mg/kg body weight (5-8 tablets) and
was swallowed with 100 ml water. A paired Student t-test
compared the mean of methionine breath test tests 1
and 2 to the methionine breath test performed during
the ingestion of alcohol or aspirin.
Data for the two treatment groups is shown in table
1. The mean maximum % oxidation per hour for all subjects
was 6.71 ± 2.55 (mean ± 1 SD). In subjects who ingested
alcohol, the methionine breath test decreased to 3.01
± 0.81 (mean change from baseline -59 ± 17%, p <
0.003) and in those who ingested aspirin, the methionine
breath test increased to 8.43 ± 1.16 (mean change from
baseline 36 ± 36%, p<0.001).
Table 1: Effect of alcohol and ASA ingestion on the
Methionine Breath Test in healthy controls. Results
are expressed as maximum percent dose per hour.
| METHIONINE BREATH
TEST Group 1 |
METHIONINE BREATH TEST Group
2 |
| Participant |
Baseline |
Alcohol |
Participant |
Baseline |
ASA |
| 1 |
6.10 |
3.36 |
10 |
8.28 |
9.47 |
| 2 |
13.05 |
4.54 |
11 |
5.92 |
10.08 |
| 3 |
5.72 |
2.76 |
12 |
4.42 |
8.88 |
| 4 |
7.04 |
2.22 |
13 |
6.32 |
8.73 |
| 5 |
4.89 |
2.97 |
14 |
7.30 |
7.77 |
| 6 |
4.32 |
3.39 |
15 |
6.02 |
7.34 |
| 7 |
5.37 |
1.86 |
16 |
8.72 |
9.24 |
| 8 |
13.66 |
3.52 |
17 |
5.92 |
6.47 |
| 9 |
6.17 |
2.45 |
18 |
4.55 |
8.61 |
| |
19 |
5.27 |
6.88 |
| 20 |
5.23 |
9.27 |
| Mean |
7.37 |
3.01 |
|
6.17 |
8.43 |
| SD |
3.48 |
0.81 |
1.41 |
1.16 |
The results indicate that the methionine
breath test significantly decreases with ingestion of
alcohol and increases with aspirin ingestion. Changes
in the methionine breath test, therefore, reflect changes
in hepatic mitochondrial function. This study suggests
that the methionine breath test is a sensitive and specific
measure of changes in hepatic mitochondrial function
induced by alcohol and aspirin and can be used as a
quantitative measure of mitochondrial function.
Ability of the Methionine Breath Test to detect Liver
Disease
The goal for this experiment was to determine if the methionine
breath test is capable of distinguishing healthy controls
from patients with liver disease. We studied 27 healthy
controls and 46 patients with well-characterized cirrhosis.
All patients were diagnosed with cirrhosis based on clinical
symptoms, standard serum liver tests, radiological testing,
and/or liver histology. All patients were clinically stable.
The current practice guideline is to determine liver disease
severity using the Child-Pugh (CP) classification that
uses standard liver blood test results and presence of
other symptoms. The CP score was determined on all liver
patients at the time of the first test. and were calculated
as shown in table 2. The CP score designations are
CP A, CP B, and CP C representing mild, moderate and severe
liver dysfunction, respectively.
Differences between healthy controls, CP A, CP B, and
CP C were compared using a One Way Analysis of Variance
with Bonferroni Group Mean Comparisons. Comparisons
were made between groups as follows; healthy controls
and CP A, CP A and CP B, and CP B and CP C.
Analysis of the data showed that the methionine breath
test distinguished healthy controls and patients with
different degrees of liver disease severity. Healthy
controls (9.16 ± 2.62, mean ± 1 SD) had methionine breath
test results that were significantly different (p<0.001)
from liver patients characterized as CP A (4.10 ± 4.53).
Liver patients classified as CP A (4.10 ± 4.53) were
significantly different (p=0.003) from CP B (2.57 ±
1.37) patients. CP B patients had methionine breath
test values that were less dramatic but still significantly
different (p=0.02) from CP C (1.33 ± 0.76) patients.
Using the threshold of 6% (maximum % dose oxidized /h)
patients were classified as either having normal or
abnormal mitochondrial function. Abnormal function was
classified as an methionine breath test < 6% and
normal function if > 6%. Utilizing the threshold
of 6%, sensitivity was 96%, specificity 100%, and accuracy
97% (see figure 3 and table 2).
Figure 3: Methionine Breath Test in controls and patients
showing the threshold level of 6% differentiates controls
from liver patients classified as CP A, CP B, or CPC.
|
Table 2: Sensitivity and Specificity of the METHIONINE
BREATH TEST for all participants.
METHIONINE BREATH TEST (%) |
Child Pugh A, B, C |
Controls |
| <6.0 |
44 Total Positive |
0 False Positive |
| >6.0 |
2 False Negative |
27 True Negative |
The sensitivity, specificity, and accuracy of the test
were calculated as follows:
Sensitivity = No. of True Positives/Total with Disorder
=
44/46 x 100 = 96%
Specificity = No of True Negatives/Total free of Disorder
=
27/27 x 100 = 100%
Accuracy = True Positives + True Negatives/ Total Investigated
=
[(44+27)/73] x 100 = 97%
The methionine breath test was able
to distinguish healthy controls, CP A, CP B, and CP
C patients with high specificity and sensitivity. Our
data suggests that mitochondrial function is compromised
as liver disease progresses either through the CP scoring
system or from stable cirrhosis to alcoholic hepatitis.
In addition, serial testing of patients over time may
be useful as a prognostic indicator of liver disease
progression and provide information for optimal timing
of therapeutic measures including liver transplantation.
Serial testing may indicate a response to treatment
and therefore allow the clinician to employ a cost effective
treatment strategy by either continuing useful therapy
or discontinuing ineffective therapy.
It was not surprising that some overlap of the methionine
breath test results occurred for each of the Child’s
classifications. This is predicted by existing literature
that reported CP A patients have a wide range of hepatic
functional impairment ranging from nearly normal to
severely abnormal. The Child’s score is only for
classification purposes and is not a true measure of
liver function. The Child’s classification should
not be considered as a reference test of liver function.
In fact, the methionine breath test shows some patients
with CP A and CP B have normal function. The clinical
occurrence of normal liver function in a CP A or CP
B occurs frequently. This is the reason that the Child’s
scoring does not offer the clinician insight into quantitative
liver function reserves.
Determining Normal Range of the Methionine Breath
Test
One hundred and fifty (150) individuals without liver
disease were administered the methionine breath test to
understand the inter-subject variation and normal range
of the test. Subjects were determined to be free of liver
disease if they had normal liver chemistries, no blood
alcohol levels prior to testing, no Hepatitis C antigens,
and no substance abuse in the urine.
The mean and median rate of methionine oxidation was 4.00%
per hour in this larger group of subjects. We felt it
was necessary to revise the scoring of the test because
the differences in the numbers were very small. We have
converted the methionine oxidation rate to a “Methionine
Breath Test Score” by multiplying all results by
25. This made the results easy to interpret and easier
to inspect true differences between healthy subjects and
those with liver disease. Therefore, the mean methionine
breath test Score is now 100. A histogram of the results
revealed a normal distribution as shown below.
Establishment for a cutoff of Cirrhosis by the Methionine
Breath Test
A retrospective analysis of our database
was used to define a methionine breath test cutoff value
that would predict the presence of cirrhosis. In the
methionine breath test analysis, 165 subjects without
any known liver disease and 22 biopsy-proven cirrhotics
were used in the analysis. Using ROC analysis, a methionine
breath test cutoff score of less than 38 was indicative
of cirrhosis. The sensitivity of the methionine breath
test was 91% (20/22 correct) and the specificity was
98% (161/165). These results showed an overall accuracy
of the methionine breath test to predict cirrhosis at
97%. This level of accuracy is compelling to evaluate
further whether the methionine breath test can replace
or reduce the use of liver biopsies.
A reduction in liver biopsies will not only reduce healthcare
costs but also improve quality of life of patients.
Liver biopsies are associated with pain in 30% of patients,
severe complications in 0.3%, and death in 0.03%. It
has also been reported that the duration of the pain
after biopsy extends beyond the day of the biopsy in
40% of patients and extends for over 1 week in a small
number.7 In fact, when questioned after the
biopsy, 15% of the patients said that they would not
have agreed to the procedure if they knew ahead how
they would feel during and after the procedure.
Reduction of the Number of Breath Samples
We recently developed a way to reduce the number of breath
samples taken with the methionine breath test, which is
currently every 10 minutes for one hour. Using a database
of over 400 subjects, we have found an excellent correlation
between the 40 minute breath collection and the area-under-the-curve
for the entire hour. The correlation was found to be r2
= 0.986, p < 0.0001. There was no statistical difference
in the final methionine breath test score calculated with
the one point (40 minute time point) method or the six
point (every 10 minutes for 1 hour) method. Since the
database is so large, we are convinced that the methionine
breath test can be performed with just one breath collection
after administering the methionine dose.
Summary of experiments:
1)Reproducibility of the methionine breath test was excellent
in all participants with a correlation of (r = 0.97).
2)The statistically significant changes in the methionine
breath test after alcohol and aspirin ingestion, agents
that are known to inhibit or induce mitochondrial metabolism
respectively, indicate that the methionine breath test
monitors mitochondrial function.
3)Methionine breath test results were significantly different
for healthy controls and patients classified as Child-Pugh
A, B, and C. The methionine breath test was also able
to identify patients with normal and abnormal function
with a high degree of sensitivity, specificity, and accuracy.
4)An average methionine breath test Score is 100 and we
have established a cutoff for cirrhosis of 38.
5)The methionine breath test had an accuracy of 97% of
predicting cirrhosis suggesting it was useful to reduce
the number of liver biopsies performed.
6) We have established that the methionine breath test
can be performed with just two breath collections, a pre-breath
sample and a 40-minute post-methionine dose sample. |
