AMPD1 Genotype Predicts
Survival in Patients with Heart Failure
|Evan Loh, M.D.
University of Pennsylvania Health System
Philadelphia, PA, USA
Loh reviewed studies in their laboratory
using a molecular epidemiology approach to determine whether
the inheritance of the adenosine monophosphate deaminase-1
mutant (AMPD1) was a predictor of survival in patients
with heart failure. A functional hypothesis for the roles
of adenosine and AMPD1 in heart failure and its therapeutic
implications were presented.
Understanding the natural history
of heart failure remains a challenge. Loh's view,
yet to be proven, is that symptomatic heart failure,
with its inevitable steep progression ending in
death, is preceded by a very long asymptomatic phase
in patients with left ventricular (LV) systolic
dysfunction. Supporting evidence from the Framingham
Heart Study was cited: after hospitalization with
symptomatic heart failure the median survival for
men is 1.7 years and for women 3.4 years. Comparing
the survival of patients with heart failure and
women with pre-menopausal breast carcinoma, a group
perceived to have a very high morbidity and mortality
rate, the 10-year survival of symptomatic heart
failure patients is about 25-30% and for breast
cancer patients about 50-55%.
Limitations of univariate predictors
Clinical, hemodynamic, neurohormonal,
and functional parameters have limited usefulness
in predicting survival in patients with heart failure,
because they are 1) univariate, 2) cross-sectional
and not followed serially, 3) the response to therapy
is not incorporated, and 4) there is little understanding
about how quickly longitudinal changes occur.
Several parameters that are known
to reflect worsening heart failure are not used
in clinical practice and provide little information
that is relevant in an individual patient. Increasing
levels of norepinephrine in patients with more advanced
disease is associated with worsening survival. Large
population-based studies show outcomes are worse
with lower left ventricular ejection fractions (LVEF).
But, LVEF is not an important predictor of exercise
capacity or long-term natural history. Data from
the post-infarction study by Moss clearly show that
one-year mortality increases as LVEF falls. However,
resting LVEF and exercise capacity have little correlation.
Data from the Ve-HeFT studies of nearly 1300 patients
who underwent metabolic stress testing showed virtually
no correlation between LVEF and VO2 maximum,
with r values less than 0.2.
Data from Mancini of patients referred
for heart transplantation evaluation showed that
patients with significantly reduced VO2 maximum
had a 50-60% one year mortality. The data further
show that exercise performance was the only predictor
of long-term survival. LVEF and norepinephrine levels
are not used carefully as predictors of survival,
and exercise capacity has been relied on heavily.
Thus, Loh thinks the development of
symptomatic heart failure dramatically changes the
slope of the natural history deterioration curve.
The SAVE trial showed that in post-infarct patients
randomized to either captopril or placebo, the captopril-treated
patients had an overall 19% reduction in overall
cardiovascular (CV) mortality (p=0.019). These asymptomatic
patients had a 4-year mortality rate of 20-25%.
Mortality following first hospitalization for heart
failure more than doubled in the captopril-treated
group to 50-55% at 4 years. In the placebo group
the mortality was nearly 70%.
epidemiology in heart failure
A molecular epidemiology approach
was used to determine predictors of survival in
patients with heart failure. The genetic background
of a person likely plays a role in determining the
response to the complex pathophysiology of heart
failure. To date, genetic determinants of prognosis
and survival in patients with heart failure have
not been established.
Molecular epidemiology seeks to determine
whether a biologically-plausible candidate gene
with an association to a well-defined disease affects
1) etiology, 2) risk stratification, 3) prognosis,
and 4) design of novel therapeutic interventions.
In contrast to identifying genes that cause a disease,
this approach seeks to identify genes that modify
the natural history.
Establishing a causal relationship
To establish a causal relation between
a mutant allele and a clinical phenotype, an increased
frequency of the mutant allele in persons with phenotypic
characteristics of the disease is needed. This causal
relation will likely be useful to identify genes
that produce or modify symptoms later in life, and
hopefully to provide a rationale for investigating
the biochemical and/or physiologic basis of disease
This approach has proven useful in
identifying "disease-modifying" loci in a variety
of disease states: 1) apolipoprotein E e2 allele
and longevity, 2) apolipoprotein E e4 allele and
Alzheimer's disease, 3) CCR-5 chemokine receptor
gene mutation and resistance to HIV-1 infection,
4) NRAMP1 polymorphism and susceptibility to tuberculosis
in West Africans, and 5) CETP polymorphism and the
progression of coronary atherosclerosis.
The AMPD1 locus identifies and transcribes
the enzyme adenosine monophosphate deaminase (AMPD)
located on the human chromosome in the p13 region.
AMPD1 is putatively the most common single point
mutation in the Caucasian and African American populations,
with a heterozygous frequency of about 20%, and
a homozygote recessive frequency of about 3-4%.
It identifies a single non-lethal point mutation
in the second exon, resulting in a nonsense mutation.
Affected persons develop a skeletal muscle myopathy
characterized by impaired exercise capacity and
the potential to produce increased circulating levels
of adenosine_relating to the theme of exercise capacity
as a good predictor of survival in patients with
AMPD is driven by AMP levels in ATP
catabolism during times of exercise or stress. AMP
is catabolized to IMP in patients with two normal
copies of the gene, therefore providing fumarate
for the purine nucleotide cycle to allow regeneration
of ATP. In persons with a defective enzyme, AMP
is shuttled via 5'nucleotidase and de-phosphorylates
This increased potential capacity
to produce increased circulating levels of adenosine
with exercise is important because adenosine is
1) a putative modulator of the ischemic preconditioning
response, 2) a potent coronary and arteriolar vasodilator,
3) anti-adrenergic, 4) anti-inflammatory, and 5)
has anti-arrhythmic properties. These factors provide
an opportunity to intervene in the progression cascade
of heart failure at an early stage, perhaps in the
Plasma concentrations of hypoxanthine
and ammonia decrease and plasma concentrations of
adenosine increase in proportion to the severity
of heart failure. Their hypothesis is that the ability
to understand modulation of AMP levels in relation
to production of circulating adenosine may lead
to a permissive effect that improves the prognosis
of patients with symptomatic heart failure.
To determine if the inheritance of
the mutant AMPD1 allele was associated with increased
survival duration in patients with symptomatic heart
failure, a series of 132 patients referred for heart
failure transplantation evaluation were studied.
The mean age was 52.8 +/- 11.2 years, mean LVEF
19.7 +/- 6.7%, and VO2 max 13.9 +/- 4.9
ml/kg/min. Etiology of heart failure was coronary
artery disease in 69 patients, idiopathic cardiomyopathy
in 48, and other in 15. The baseline characteristics
are shown in Table
A simple AMPD1 allele-specific oligonucleotide
detection assay was used. For each patient, clinical
phenotypic data was collected and AMPD1 genotyping
1. Probability of survival without cardiac transplantation
in heart failure patients stratified across AMPD1genotype
by Kaplan-Meier analysis. The two curves are statistically
different (P=0.002). (Circulation 1999;99:1422-1425.)
Twenty-one of the 132 patients were
either heterozygous or homozygous recessive for the
mutation. These 21 patients had a significantly longer
duration of heart failure symptoms before presentation,
compared to those without the mutation (7.6 +/-6.5
yrs vs 3.2 +/- 3.6 yrs, respectively). There was virtually
no difference between these groups in terms of age,
max, LVEF, cardiac index, pulmonary
capillary wedge pressure, and pulmonary vascular resistance.
Figure 1 shows the probability of survival without
heart transplant based on AMPD1 gentotype.
The patients with either one or two
copies of the mutation had a significantly better
survival at 15 years of nearly 70-72%, compared
to patients without either one or two copies of
the mutation. This was based on a retrospective
review from the time of presentation of symptomatic
heart failure to time of presentation for evaluation
and/or death. The single most important multivariate
predictor was inheritance of the AMPD1 allele, with
a hazard ratio of 4.65 (95% CI 1.48-14.68; p=0.009).
As shown in Table
2, independent multivariate predictors of survival
were identified to be: use of diuretics (HR 0.31,
95% CI 0.11-0.83, p=0.02), ischemic cardiomyopathy
(HR 2.93, 95% CI 1.39-6.19, p=0.005), a lower VO2
max (HR 0.80, 95% CI 0.72-0.89, p=0.001), and greater
age (HR 0.92, 95% CI 0.89-0.95, p=0.001).
In patients with advanced heart failure,
the inheritance of the mutant AMPD1 allele is associated
with a 1) milder clinical natural history, and 2)
prolonged survival after the onset of symptomatic
heart failure symptoms.
The limitations in this retrospective
association study include: 1) a selective patient
population of inference, 2) other potentially beneficial
effects of reduced AMPD activity can not be excluded,
and 3) another polymorphism in another gene on chromosome
1 may be in linkage disequilibrium with AMPD1, and
it alone or in conjunction with AMPD1 may be responsible
for the observed clinical benefit.
as a modulator of survival
In collaboration with Hisatomi, using
human tissue taken during heart transplantation,
they showed that the AMPD1 isoform was found only
in the right atrium.
Therefore, they believe that AMPD1
is important at either an autocrine or paracrine
level for cardiac myocyte production, directing
attention to the skeletal muscle to understand how
the modulation of circulating adenosine levels may
be based on changes in skeletal muscle physiology.
Preliminary data showed that patients
with more preserved exercise capacity had higher
levels of circulating adenosine. Patients with asymptomatic
heart failure were exercised. Baseline and peak
exercise levels of norepinephrine and adenosine
were measured. Group A was comprised of 26 patients
with relatively impaired VO2 max (cut-off
of 16 ml/kl/min) and Group B was comprised of 27
patients with better preserved exercise capacity.
Both groups had a significant increase in serum
levels of circulating norepinephrine. But, the patients
with better preserved exercise capacity (greater
mean level of VO2 max) had a much greater
increase in circulating levels of adenosine, (150
pg/ml at baseline vs 180 at peak VO2 in
Group A, 110 pg/ml at baseline vs 170 at peak VO2
in Group B, p<0.0001.
A link between skeletal muscle and
cardiac physiology is provided by the belief that
the skeletal muscle is the potential source of adenosine.
Persons with the AMPD1 mutation produce higher circulating
levels of adenosine, which can go to the heart and
modulate ventricular function, coronary vasodilator
function, and also potentially vascular function
by modulating and counteracting the deleterious
effects of increased levels of norepinephrine that
are well characterized in patients with heart failure.
Therefore, adenosine can work as a compensatory
mechanism to stabilize patients with heart failure,
and decrease the deterioration seen in patients
with symptomatic heart failure. This hypothesis
remains to be proven.
Inhibition of AMP deaminase and/or
regulation of adenosine production may present novel
therapeutic targets to change the natural history
of heart failure. Molecular epidemiology techniques
offer promise for identification of disease modifying
loci in common complex disorders such as heart failure.
A genetic and biochemical plausibility to the hypothesis
is required; at least one plausible hypothesis based
on their data has been developed. The clinical importance
of the observation is dependent upon an understanding
of the genetic makeup of the population of inference.
The gene of interest must cause relevant and important
changes in structure or function at the level of
the gene product; this remains the principle experiment
they must prove using genotyping.
Future work includes: 1) amplifying
the genotype-phenotype association based on whether
or not this is an inherited mutation, 2) to risk
stratify future populations based on these studies
in a prospective manner (a large prospective study
is ongoing), and 3) develop translational interventions
based on the combined understanding of the molecular
and biochemical bases of the disease process.
Index | Previous Report
| Next Report
Sessions | Activities
Copyright © 2000
Japanese Circulation Society
All Rights Reserved.