Estimated Glomerular Filtration Rate From a Panel of Filtration Markers-Hope for Increased Accuracy Beyond Measured Glomerular Filtration Rate?


Inker, L.A.; Levey, A.S.; Coresh, J.

Advances in Chronic Kidney Disease 25(1): 67-75

2018


The recent Kidney Disease Improving Global Outcomes 2012 CKD guidelines recommend estimating GFR from serum creatinine (eGFR<sub>cr</sub>) as a first-line test to assess kidney function and using cystatin C or measured glomerular filtration rate (GFR) as confirmatory tests. eGFR<sub>cr</sub> may be inaccurate in people with variation in muscle mass or diet, and eGFR<sub>cys</sub> is not more accurate than eGFR<sub>cr.</sub> eGFR<sub>crcys</sub> is more accurate than either, but it is not independent of eGFR<sub>cr</sub>. Measured GFR is not practical and is susceptible to error due to variation in clearance methods and in the behavior of exogenous filtration markers. Over the past few years, we have hypothesized, and begun to test the hypothesis, that a panel of filtration markers (panel eGFR) from a single blood draw would require fewer demographic or clinical variables and could estimate GFR as accurately as measured GFR. In this article, we describe the conceptual background and rationale for this hypothesis and summarize our work thus far including evaluation of novel low-molecular-weight proteins and metabolites and then outline how we envision that such a panel could be used in clinical practice, research, and public health.

AC
KD
Estimated
Glomerular
Filtration
Rate
From
a
Panel
of
Filtration
Markers—Hope
for
Increased
Accuracy
Beyond
Measured
Glomerular
Filtration
Rate?
Lesley
A.
Inker,
Andrew
S.
Levey,
and
Josef
Coresh
The
recent
Kidney
Disease
Improving
Global
Outcomes
2012
CKD
guidelines
recommend
estimating
GFR
from
serum
creatinine
(eGFR
c
„)
as
a
first-line
test
to
assess
kidney
function
and
using
cystatin
C
or
measured
glomerular
filtration
rate
(GFR)
as
confir-
matory
tests.
eGFR,
r
may
be
inaccurate
in
people
with
variation
in
muscle
mass
or
diet,
and
eGFR
zys
is
not
more
accurate
than
eGFR
cr
_
eGFR
c
„,,.
is
more
accurate
than
either,
but
it
is
not
independent
of
eGFli
c
„.
Measured
GFR
is
not
practical
and
is
suscep-
tible
to
error
due
to
variation
in
clearance
methods
and
in
the
behavior
of
exogenous
filtration
markers.
Over
the
past
few
years,
we
have
hypothesized,
and
begun
to
test
the
hypothesis,
that
a
panel
of
filtration
markers
(panel
eGFR)
from
a
single
blood
draw
would
require
fewer
demographic
or
clinical
variables
and
could
estimate
GFR
as
accurately
as
measured
GFR.
In
this
article,
we
describe
the
conceptual
background
and
rationale
for
this
hypothesis
and
summarize
our
work
thus
far
including
evaluation
of
novel
low-molecular-weight
proteins
and
metabolites
and
then
outline
how
we
envision
that
such
a
panel
could
be
used
in
clin-
ical
practice,
research,
and
public
health.
©
2017
by
the
National
Kidney
Foundation,
Inc.
All
rights
reserved.
Key
Words:
Glomerular
filtration
rate,
Creatinine,
Metabolomics,
Cystatin
C,
Estimated
GFR
INTRODUCTION
Assessing
kidney
function
is
an
integral
part
of
the
prac-
tice
of
medicine,
research,
and
public
health.
1
'
2
The
recent
Kidney
Disease
Improving
Global
Outcomes
(KDIGOs)
2012
CKD
guidelines
recommend
estimating
GFR
from
serum
creatinine
(eGFR,
)
as
a
first-line
test.
Indeed,
more
than
hundreds
of
million
measurements
of
serum
creatinine
are
performed
annually
by
clinical
laboratories
in
the
United
States,
with
more
than
90%
reporting
eGFR
automatically
when
serum
creatinine
is
measured."
eGFR,
may
be
inaccurate
in
people
with
variation
in
muscle
mass
or
diet,
and
there
is
great
interest
in
cystatin
C
as
an
alternative
filtration
marker
to
creatinine.
The
KDIGO
guidelines
recommend
using
eGFR
based
on
cystatin
C
(eGFR
cys
)
or
the
combination
of
the
two
(eGFR,_
cys
)
as
a
confirmatory
test
for
eGFRcr.
3
However,
there
are
limitations
of
this
approach
because
eGFR
cys
is
not
more
accurate
than
eGFR,
and
although
eGFR,_
cys
is
more
accurate
than
either
eGFR,
or
eGFR
cys
,
it
is
not
independent
of
eGFR,
Measured
GFR
(mGFR)
using
clearance
of
exogenous
filtration
markers
is
also
recommended
by
KDIGO
as
a
confirma-
tory
test.
While
mGFR
is
not
influenced
by
non-GFR
de-
terminants
of
endogenous
filtration
markers,
it
is
not
practical
and
is
susceptible
to
error
due
to
variation
in
clearance
methods
and
in
the
behavior
of
exogenous
filtration
markers.
Hence,
there
is
a
need
for
a
simple
but
more
accurate
estimate
of
GFR
to
guide
individual
decision-making.
Over
the
past
few
years,
we
have
hypothesized,
and
begun
to
test
the
hypothesis,
that
a
panel
of
filtration
markers
(panel
eGFR)
from
a
single
blood
draw
would
require
fewer
demographic
or
clinical
variables
and
could
estimate
GFR
as
accurately
as
mGFR.
The
overall
goal
of
this
review
is
to
describe
the
conceptual
background
and
rationale
for
this
hypothesis
and
summarize
our
work
thus
far.
To
do
so,
we
will
first
review
available
data
on
strengths
and
limitations
of
eGFR
based
on
creatinine
and
cystatin
C
as
well
as
those
of
mGFR.
We
will
then
re-
view
our
conceptual
framework
for
why
panel
of
filtration
markers
might
overcome
the
limitations
of
current
esti-
mating
equations
and
mGFR
and
discuss
our
exploration
of
candidate
filtration
markers
for
inclusion
in
the
panel.
Finally,
we
will
outline
how
we
envision
that
such
a
panel
could
be
used
in
clinical
practice,
research,
and
public
health.
From
the
Tufts
Medical
Center,
Boston,
MA;
and
Welch
Center
for
Prevention,
Epidemiology
and
Clinical
Research,
Johns
Hopkins
University,
Baltimore,
MD.
Financial
Disclosure:
L.I.
reports
funding
to
Tufts
Medical
Center
for
research
and
contracts
with
the
National
Institutes
of
Health,
National
Kidney
Foundation,
Pharmalink,
Gilead
Sciences,
Otsuka
and
has
a
provisional
patent
(Coresh,
Inker
and
Levey)
filed
8/15/2014
"Precise
estimation
of
glomerular
filtration
rate
from
multiple
biomarkers"
PCT/US2015/044567.
The
technology
is
not
licensed
in
whole
or
in
part
to
any
company.
Tufts
Medical
Center,
John
Hopkins
University,
and
Metabolon
Inc.
have
a
collaboration
agreement
to
develop
a
product
to
estimate
GFR
from
a
panel
of
markers.
A.L.
reports
funding
to
Tufts
Medical
Center
for
research
and
contracts
with
the
National
Institutes
of
Health,
National
Kidney
Foundation,
Amgen,
Pharmalink,
Gilead
Sciences,
Siemens
and
has
a
provisional
patent
(Coresh,
Inker
and
Levey)
filed
8/15/
2014
"Precise
estimation
of
glomerular
filtration
rate
from
multiple
bio-
markers"
PCT/US2015/044567.
The
technology
is
not
licensed
in
whole
or
in
part
to
any
company.
Tufts
Medical
Center,
John
Hopkins
University,
and
Me-
tabolon
Inc.
have
a
collaboration
agreement
to
develop
a
product
to
estimate
GFR
from
a
panel
of
markers.
J.C.
reports
funding
to
Johns
Hopkins
University
for
research
and
contracts
with
the
National
Institutes
of
Health
and
National
Kid-
ney
Foundation
and
has
a
provisional
patent
(Coresh,
Inker
and
Levey)
filed
8/
15/2014
"Precise
estimation
of
glomerular
filtration
rate
from
multiple
bio-
markers"
PCT/US2015/044567.
The
technology
is
not
licensed
in
whole
or
in
part
to
any
company.
Tufts
Medical
Center,
John
Hopkins
University
and
Me-
tabolon
Inc.
have
a
collaboration
agreement
to
develop
a
product
to
estimate
GFR
from
a
panel
of
markers.
Address
correspondence
to
Lesley
A
Inker,
Division
of
Nephrology,
Tufts
Medical
Center,
800
Washington
Street,
Box
#391,
Boston,
MA
02111.
E-mail:
©
2017
by
the
National
Kidney
Foundation,
Inc.
All
rights
reserved.
1548-5595/$36.00
https://doi.org/10.1053/j.ackd.2017.10.004
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
67
68
Inker
et
al
ASSESSMENT
OF
GFR—HOW
ACCURATE
ARE
WE?
GFR
cannot
be
measured
directly
in
humans;
thus,
"true"
or
physiologic
GFR
cannot
be
known
with
cer-
tainty.
Instead,
we
use
serum
levels
or
clearance
measure-
ments
of
filtration
markers,
exogenous,
or
endogenous
solutes
that
are
eliminated
mainly
by
glomerular
filtration
to
assess
GFR.
Current
methods
are
associated
with
sys-
tematic
or
random
error
(bias
and
imprecision,
respec-
tively)
in
their
determination,
which
may
limit
their
use
depending
upon
the
magnitude
of
the
error
and
require-
ment
for
accuracy
for
clinical
decision-making
(Table
1).
Measured
GFR
The
"gold
standard"
for
the
measurement
of
GFR
is
uri-
nary
clearance
of
an
ideal
filtration
marker,
defined
as
sub-
stance
that
is
freely
filtered
at
the
glomerulus,
neither
reabsorbed,
secreted,
synthesized,
or
metabolized
by
the
tubules
and
does
not
alter
the
function
of
the
kidney.
The
"classic"
method
of
Smith
used
urinary
clearance
of
inulin,
a
5200
Da
polymer
of
fructose,
during
a
continuous
intravenous
infusion.
Inulin
is
difficult
to
use
and
not
available
in
the
United
States,
and
urinary
clearance
mea-
surements
are
cumbersome,
so
alternative
filtration
markers
and
methods
have
been
proposed
but
all
deviate
from
the
"gold
standard"
leading
to
potential
sources
of
bias
as
compared
to
true
GFR.
For
example,
plasma
clearance
after
a
bolus
intra-
venous
infusion
is
simpler
to
perform
but
may
differ
from
urinary
clearance
due
to
nonequilibration
across
body
fluid
compartments
and
extra-renal
elimination
of
the
filtration
marker.
In
addition,
all
alternative
filtration
markers
deviate
from
ideal
behavior.
8
A
recent
systematic
review
evalu-
ated
alternative
methods
in
comparison
to
the
classic
pro-
cedure
of
Smith
and
noted
wide
variation
in
performance
even
within
the
same
method.
Clearance
measurements
are
difficult
to
perform
leading
to
imprecision
in
mGFR.
The
usual
method
to
quantify
imprecision
in
mGFR
is
through
repeated
measures.
The
within
person
coefficient
of
variation
for
repeated
measures
on
different
days
for
GFR
measurement
methods
varies
from
approximately
5%
to
15%,
with
higher
values
for
uri-
nary
clearance
than
plasma
clearances.
lu-16
True
GFR
may
vary
over
short
intervals,
so
observed
variation
in
mGFR
likely
reflects
normal
biological
variation
in
true
GFR
as
well
as
measurement
error.
By
contrast,
the
imprecision
in
measurements
of
serum
concentrations
of
endogenous
filtration
markers
can
be
less
than
mGFR,
in
part
because
fluctuations
in
true
GFR
affect
serum
concentrations
of
filtration
markers
more
slowly
than
clearances,
and
in
part,
because
it
is
simpler
to
measure
a
serum
concentration
than
to
perform
a
clearance
measurement.
Beyond
emphasizing
the
need
for
a
more
accurate
confir-
matory
test,
error
in
mGFR
has
important
implications
for
interpretation
of
error
in
eGFR.
Error
in
mGFR
does
not
affect
the
serum
levels
of
endogenous
filtration
markers.
However,
since
we
use
mGFR
as
the
reference
test
for
eval-
uating
the
accuracy
of
eGFR,
observed
errors
in
eGFR
may
in
part
be
due
to
error
in
mGFR
(Table
1).
With
advances
in
GFR
estimation,
accuracy
of
eGFR
will
improve,
and
the
relative
contribution
of
error
in
mGFR
to
the
observed
er-
ror
in
eGFR
will
increase.
To
demonstrate
the
impact
of
er-
ror
in
mGFR
on
the
observed
error
in
eGFR,
we
assessed
the
effect
of
variability
of
GFR
measurement
on
the
perfor-
mance
of
the
Modification
of
Diet
in
Renal
Diseases
(MDRDs)
Study
equation
and
Chronic
Kidney
Disease
Epidemiology
Collaboration
(CKD-EPI)
in
two
clinical
tri-
als
with
GFR
measured
using
urinary
clearance
of
iothala-
mate.
The
left
hand
panel
of
Figure
1
shows
the
difference
between
two
mGFRs
on
average
62
days
apart
in
the
African
American
Study
of
Kidney
Diseases
(AASK)
and
MDRD
Study.
A
total
of
12%
of
subjects
had
measures
that
were
discrepant,
as
defined
by
a
difference
of
more
than
25%.
The
right
hand
panel
of
Figure
1
shows
the
improvement
in
accuracy
when
more
precise
mGFR
was
used
as
the
reference
test
[reduction
in
large
errors
(1-P
30
)
from
17%
to
3.9%].
Estimating
GFR
Determinants
of
Endogenous
Filtration
Markers.
Serum
levels
of
an
endogenous
filtration
marker
are
determined
not
only
by
the
level
of
GFR,
but
also
by
physiolog-
ical
processes
other
than
GFR
(generation,
kidney
tubular
secretion
and
reab-
sorption,
and
extra-renal
elimination).
Collectively,
these
physiological
pro-
cesses
are
termed
non-GFR
_
.
j•
determinants,
and
their
steady-state
relationships
to
GFR
and
serum
concentra-
tions
are
shown
in
Figure
2.
8
These
physiological
processes
are
generally
not
measured,
so
estimating
equations
use
easily
measured
demographic
and
clinical
variables
as
sur-
rogates.
GFR
estimates
are
more
accurate
than
the
serum
level
of
the
marker
alone
but
have
two
principal
limita-
tions
which
are
possible
sources
of
error.
First,
surrogates
only
capture
the
average
relationships
between
the
marker
and
its
non-GFR
determinants.
Second,
the
relationship
between
the
marker
and
its
non-GFR
determinants
may
vary
across
populations.
The
non-GFR
determinants
may
vary
across
markers
even
though
the
serum
level
for
each
marker
is
correlated
to
GFR.
Estimating
GFR
Using
Creatinine
and
Cystatin
C.
Creatinine
is
the
most
commonly
used
endogenous
filtra-
tion
marker.
It
is
freely
filtered
by
the
glomerulus
but
un-
dergoes
extra-renal
elimination
by
the
gut,
is
secreted
by
the
tubules,
and
is
generated
by
muscle
mass
or
diet.
Creatinine-based
estimating
equations
include
age,
sex,
race,
or
weight
as
surrogates
for
creatinine
generation
from
muscle
mass
or
diet.
Regardless
of
the
specific
equation,
the
accuracy
of
eGFR,
is
limited
by
variation
in
muscle
mass
or
diet
Current
glomerular
filtration
rate
(GFR)
estimates
are
limited
in
their
accuracy.
Combining
filtration
markers
in
a
panel
from
a
single
blood
draw
could
require
fewer
demographic
or
clinical
variables
and
could
estimate
GFR
as
accurately
as
measured
GFR.
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
Estimated
GFR
From
a
Panel
of
Filtration
Markers
69
Table
1.
Factors
Affecting
Errors
in
Measured
and
Estimated
GFR
Factors
Bias
Imprecision
Error
in
measured
GFR
compared
to
true
(physiologic)
GFR
Biological
variation
in
GFR
Exogenous
filtration
markers
Protein
binding
Tubular
reabsorption
and
secretion
Extra-renal
elimination
Assay
Errors
in
urine
collection
for
urinary
clearances
Errors
in
plasma
clearances
Edematous
conditions
Insufficient
time
interval
for
low
GFR
Error
in
estimated
GFR
compared
to
measured
GFR
Differences
in
mean
level
of
GFR
in
population
in
which
estimating
equation
was
developed
and
to
which
it
is
applied
Differences
in
method
used
to
measure
GFR
in
which
estimating
equation
was
developed
and
to
which
it
is
applied
Imprecision
in
measured
GFR
Endogenous
markers
Generation
Tubular
reabsorption
and
secretion
Extra-renal
elimination
Assay
Abbreviation:
GFR,
glomerular
filtration
rate.
*Most
sources
of
bias
can
also
increase
imprecision
since
the
degree
of
bias
varies
across
individuals.
independent
of
age,
sex,
or
race.
People
with
extremes
of
muscle
mass
and
dietary
intake,
for
example,
those
who
are
malnourished
or
have
a
reduction
in
muscle
mass
from
illness
or
amputation,
are
likely
to
have
large
differ-
ences
between
eGFR,
and
mGFR,
leading
to
bias
in
these
populations.
-9
Even
in
populations
without
these
condi-
tions,
eGFR,
may
be
biased
because
of
systematic
differ-
ences
in
muscle
mass
or
diet
compared
to
the
population
in
which
the
equation
was
developed,
as
is
observed
across
racial
and
ethnic
groups
and
geographical
regions.
Finally,
even
in
populations
with
similar
average
levels
of
muscle
mass
and
dietary
intake,
eGFR,
is
imprecise
because
of
variability
in
the
relationship
of
muscle
mass
and
diet
to
age,
sex,
and
race.
Cystatin
C
is
an
alternative
endogenous
filtration
marker
that
is
less
influenced
by
muscle
and
diet
than
creatinine.
It
is
freely
filtered
at
the
glomerulus
but
has
some
degree
of
extra-renal
elimination
and
is
catabolized
in
the tubules
with
reabsorption
of
its
metabolites
and
is
generated
by
all
nucleated
cells.
20
'
21
Studies
show
that
eGFR
cys
is
not
more
accurate
than
eGFR,
but
eGFR,_
cys
is
more
precise
than
either
alone
22-24
Table
2
shows
the
coefficients
for
var-
iables
in
the
CKD-EPI
equations
using
creatinine,
cystatin
C,
or
both.
In
the
equation
using
both
markers,
the
coef-
ficient
for
each
marker
is
smaller
(closer
to
zero)
as
is
the
magnitude
of
the
coefficients
for
age,
sex,
and
race
(closer
to
1.0).
Thus,
deviations
from
the
average
values
for
the
relationship
of
these
variables
to
mGFR
affect
the
eGFR
less
than in
an
equation
which
includes
only
one
marker.
A
recent
large
study
of
non-GFR
determinants
of
cystatin
C
in
3
CKD
populations
showed
that
cystatin
C
was
more
highly
correlated
with
mGFR
than
creatinine
(-0.88
vs
—0.80)
(Table
3
upper
panel),
and
their
correlation
with
each
other
independent
of
mGFR
was
low
(0.38)
(Table
3
lower
panel).
2b
Similar
findings
were
shown
in
two
community-based
studies
in
participants
with
higher
GFR.
27
Consistent
with
these
findings,
numerous
factors
had
associations
with
serum
cystatin
C
independent
of
mGFR
which
differed
from
their
associations
with
serum
creatinine
(Fig.
3).
27
Male
sex
and
smoking
(each
associ-
ated
with
higher
serum
concentrations)
had
moderate
and
strong
associations,
respectively.
Other
associations
(height,
weight,
diabetes,
and
CRP)
were
weak.
In
partic-
ular,
race
was
not
significantly
associated
with
cystatin
C.
Findings
from
other
studies
vary
somewhat,
in
part
due
to
use
of
other
methods
and
absence
of
multivariable
adjustment.
They
suggest
that
higher
BMI
and
higher
levels
of
inflammation
as
indicated
by
disease
or
levels
of
CRP
may
also
be
important
factors
associated
with
higher
serum
concentrations
of
cystatin
C.
In
our
view,
there
are
important
limitations
to
eGFR,_
cys
.
First,
eGFR,_
cys
is
not
independent
from
eGFR,;
thus,
it
does
not
meet
the
requirement
for
a
true
confirmatory
test.
Second,
because
there
are
only
two
markers,
discrep-
ancies
between
eGFR,
and
eGFR
cys
may
be
difficult
to
interpret.
In
some
circumstances,
the
interpretation
is
clear:
For
example,
in
a
study
of
otherwise
healthy
ampu-
tees,
eGFR,
but
not
eGFR
cys
overestimated
mGFR.
Conversely,
in
a
patient
with
critical
illness
due
to
uncon-
trolled
HIV,
eGFR
cys
underestimated
mGFR
presumably
because
of
severe
inflammation.
However,
in
many
con-
ditions,
the
interpretation
is
less
clear,
in
part
because
fac-
tors
associated
with
non-GFR
determinants
of
cystatin
C
are
less
well
known.
Indeed,
data
on
children,
cystic
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
70
Inker
et
al
20
'
'CKD
EPicreagnine-cystatin
C
equation
total
CKD
EPicreatinine-cystatin
C
equation
limited
to
consistent
mGFRs
only
MDRD
Study
equation
total
- .-
MDRD
Study
equation
limited
to
consistent
mGFRs
only
o
.
0
o
"
13
Consistent
AASK
mGFR
°
Consistent
MDRD
Study
mGFR
Discrepant
AASK
mGFR
°
Discrepant
MDRD
Study
mGFR
15
-
10
0-
8
5
0
15
30
45
60
75
90
Reference
mGFR
Average
of
2
mGFRs
Average
of
3
mGFRs
Average
of
Pre-Randomization
mGFRs
(mUmin/1.73m
2
)
Gold
Standard
Figure
1.
Variability
in
measured
GFR
and
its
effect
on
performance
of
GFR
estimating
equations.
Left
hand:
Bland—Altman
plot
of
the
difference
of
the
two
prerandomization
glomerular
filtration
rate
measurements
(mGFRs)
by
the
average
of
pre-
randomization
mGFRs.
Symbols
indicate
consistent
African
American
Study
of
Kidney
Disease
(AASK)
mGFRs
(black
plus
signs),
discrepant
AASK
mGFRs
(gray
plus
signs),
consistent
Modification
of
Diet
in
Renal
Disease
(MDRD)
Study
mGFRs
(black
circles),
and
discrepant
MDRD
Study
mGFRs
(gray
circles).
Right
hand:
Estimated
accuracy
of
CKD-EPI
creatinine-cys-
tatin
C
equation
and
the
MDRD
Study
equation
in
estimating
six
different
gold
standard
measured
glomerular
filtration
rates
(mGFRs):
reference
mGFR,
average
of
two
mGFRs,
and
the
average
of
three
mGFRs
without
and
with
limiting
the
analysis
to
consistent
(within
25%
of
each
other)
mGFRs.
Boxes
indicate
the
CKD-EPI
creatinine-cystatin
C
equation,
and
circles
indicate
the
MDRD
Study
equation.
Dotted
lines
include
all
participants.
Solid
lines
include
only
participants
with
consistent
(<25%)
mGFRs.
P30
is
calculated
as
the
percentage
of
estimated
GFR
within
30%
of
the
gold
standard
(the
average
of
three
mGFRs).
Reprinted
with
Permission
from
Kwong
YT,
Stevens
LA,
Selvin
E,
et
al.
Imprecision
of
urinary
iothalamate
clearance
as
a
gold
standard
measure
of
GFR
decreases
the
diagnostic
accuracy
of
kidney
function
estimating
equations.
American
Journal
of
Kidney
Diseases.
2010;
56(1):39-49.
Abbreviations:
CKD-EPI,
Chronic
Kidney
Disease
Epidemiology
Collaboration;
GFR,
glomerular
filtration
rate.
fibrosis,
and
muscle
wasting
diseases
show
variation
on
the
relative
performance
of
eGFR,
vs
eGFR
cy5
.
19
From
an
analytical
perspective,
having
at
least
three
markers
would
allow
triangulation,
or
convergence,
that
is
not
pro-
vided
by
two
markers.
Third,
eGFR,_
cys
requires
race
as
well
as
age
and
sex,
which
may
be
difficult
to
specify
in
racial-ethnic
groups
other
than
blacks
and
whites,
in
geographical
regions
outside
of
North
America,
Europe,
and
Australia,
in
people
whose
race
cannot
be
easily
cate-
G
(Ce
U
x
V
(kidney)
UxV=GFRxP-TR+TS
GFR
G-E=GFR
xP-TR
+TS
TR
TS
GFR
=
(G
+
TR
-
TS
-
E)
/
P
Figure
2.
Determinants
of
serum
levels
of
endogenous
filtration
markers.
The
serum
concentration
(S)
of
an
endogenous
filtration
marker
is
determined
by
its
genera-
tion
(G),
extra-renal
elimination
(E),
and
urinary
excretion
(UV).
Urinary
excretion
is
the
sum
of
filtered
load
(GFR
x
S)
plus
secretion
(TS)
minus
tubular
reabsorption
(TR).
In
the
steady
state,
urinary
excretion
is
equals
generation
minus
extra-renal
elimination.
By
substitution
and
rear-
rangement,
GFR
can
be
expressed
as
the
reciprocal
of
the
serum
concentration
multiplied
by
the
non-GFR
determi-
nants.
Modified
with
Permission
from
Stevens
LA
and
Levey
AS.
Measured
GFR
as
a
confirmatory
test
for
estimated
GFR.
J
Am
Soc
Nephrol.
2009;
20(11):2305-2313.
Abbreviation:
GFR,
glomerular
filtration
rate.
gorized,
and
in
people
in
whom
information
on
race
is
not
available.
Fourth,
although
a
standardized
reference
mate-
rial
for
cystatin
C
is
now
available,
considerable
variation
remains
among
cystatin
C
assays,
limiting
its
use.
34
For
these
reason,
we
have
explored
additional
markers
for
in-
clusion
in
a
panel.
PANEL
ESTIMATED
GFR
Conceptual
Framework
The
traditional
approach
to
improving
GFR
estimation
from
endogenous
filtration
markers
is
to
search
for
a
single
marker
with
minimal
non-GFR
determinants.
However,
as
described
above,
all
endogenous
markers
are
affected
by
non-GFR
determinants,
since
all
are
generated
within
the
body
and
most
are
eliminated
by
processes
in
addition
to
glomerular
filtration.
In
principle,
if
the
number
of
filtra-
tion
markers
included
in
an
estimating
equation
is
larger
and
the
markers
are
less
correlated
with
each
other
inde-
pendent
of
GFR,
then
the
coefficients
for
each
filtration
marker
will
be
smaller
and
the
coefficients
for
the
demo-
graphic
or
clinical
variables
will
be
closer
to
1.0.
Thus,
esti-
mating
GFR
from
multiple
noncorrelated
markers
would
minimize
the
impact
of
non-GFR
determinants
of
each
marker
and
lessen
the
need
for
demographics
and
clinical
characteristics
as
surrogates
for
the
non-GFR
determi-
nants,
with
increased
precision
as
the
number
of
markers
increases
(Fig.
4).
In
addition,
in
the
steady
state,
the
short-term
variability
in
serum
concentrations
of
endogenous
filtration
markers
should
be
less
than
the
short-term
variability
in
"true"
GFR
and
their
non-GFR
G
(Diet)
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
Estimated
GFR
From
a
Panel
of
Filtration
Markers
71
Table
2.
Coefficients
for
Variables
and
in
CKD-EPI
Equations
Equations
Developed
in
a
Diverse
Population
Equations
Developed
in
CKD
Populations
Variable
Creatinine
Cystatin
C
Creatinine
and
Cystatin
C
BTP
B2M
BTP-B2M
Creatinine
cr
-1.209
cr
-0.601
Cystatin
C
cys
-
1.328
cys
-0.711
BTP
BTP-°.
694
BTP
-0.278
B2M B2M
-0.851
B2
m
-0.586
Age
(linear)
ci.mage
omsage
0.995age
omaage
Female
sex
0.75
0.93 0.83
0.90
Black
race
1.159
1.08
Abbreviations:
B2M,
I3-2-microglobulin;
BTP,
p-trace
protein;
CKD-EPI,
Chronic
Kidney
Disease
Epidemiology
Collaboration;
GFR,
glomerular
filtration
rate;
Scr,
serum
creatinine;
Scys,
serum
cystatin.
For
creatinine
and
cystatin
C,
equation
development
in
5352
subjects
with
and
without
CKD,
regressing
logarithm
of
Scr,
Scys,
or
both
on
logarithm
of
measured
GFR
(two-slope
linear
splines
with
single
knot
for
Scr
[sex
specific]
and
Scys).
Coefficients
for
creatinine,
cystatin
C,
and
sex
are
for
the
regressions
above
the
knots.
22
For
BTP
and
B2M,
equation
development
in
2380
subjects
with
CKD.
determinants.
For
these
reasons,
we
hypothesized
that
a
panel
of
markers
to
estimate
GFR
(panel
eGFR)
could
be
as
accurate
as
or
even
more
accurate
than
mGFR
as
an
in-
dex
of
the
true
GFR.
Furthermore,
we
hypothesized
some
specific
characteristics
of
panel
eGFR
and
its
performance.
1.
Panel
eGFR
can
lead
to
substantial
improvement
in
GFR
estimation
even
if
the
individual
novel
marker
is
not
more
accurate
than
well-established
markers.
2.
Panel
eGFR
would
particularly
improve
GFR
estima-
tion
in
clinical
settings
where
the
estimate
based
on
any
single
filtration
marker
is
inaccurate—that
is,
the
incremental
value
of
a
novel
marker
above
an
estab-
lished
marker
is
likely
to
be
greater
in
study
popula-
tions
in
which
the
eGFR
based
on
the
established
marker
is
biased
or
imprecise;
3.
Panel
eGFR
might
allow
GFR
estimation
without
spec-
ification
of
demographic
and
clinical
characteristics.
Thus
far,
we
have
identified
two
groups
of
candidate
filtration
markers
for
inclusion
in
a
panel:
low-
molecular-weight
(LMW)
serum
proteins
and
metabolites.
Below,
we
discuss
our
work
with
each
group
thus
far.
As
describing
evidence
for
or
against
their
candidacy
for
filtration
markers
to
be
included
in
a
panel,
we
focus
on
the
following
five
parameters:
(1)
correlation
with
mGFR
and
with
other
markers
adjusting
for
mGFR;
(2)
knowl-
edge
of
their
non-GFR
determinants
and
demographic
and
clinical
factors
associated
with
them;
(3)
performance
eGFR
based
on
the
novel
markers
compared
to
eGFR,
or
eGFR,
ys
;
(4)
need
for
demographics
when
used
in
esti-
mating
equations;
and
(5)
practical
details
for
implementa-
tion,
such
as
standardization
of
assays
and
likelihood
that
they
can
be
run
on
single
platform
which
would
ultimately
minimize
cost
and
blood
draws.
Because
our
work
is
in
development,
we
are
not
able
to
comment
upon
all
of
these
parameters
for
each
of
the
markers
discussed.
Candidate
Low-Molecular-Weight
Proteins
Like
cystatin
C,
(3-trace
protein
(BTP)
and
0-2-
microglobulin
(B2M)
are
LMW
serum
proteins
that
undergo
glomerular
filtration
and
are
reabsorbed
and
catabolized
by
the
proximal
tubule,
with
little
loss
in
the
urine,
and
whose
serum
levels
are
less
dependent
upon
muscle
mass
and
diet
than
creatinine.
36
BTP,
also
known
as
prostaglandin
D
synthase,
is
a
168
aminoacid
glycoprotein
with
heterogeneous
glycation
that
is
generated
in
the
choroid
plexus,
testes,
and
ovaries.
It
is
a
member
of
the
lipocalin
protein
superfamily
and
Table
3.
Correlations
Among
Filtration
Markers
in
CKD
Studies
Filtration
Marker
mGFR
Creatinine
Cystatin
C
Beta
Trace
Protein
Pearson
correlation
coefficients
Log
(creatinine)
-0.80
Log
(cystatin
C)
-0.88
0.81
Log
(p-trace
protein)
-0.81
0.75
0.82
Log
(I3-2-microglobulin)
-0.85
0.78
0.90
0.81
Partial
Pearson
correlation
coefficients,
controlling
for
log
(mGFR)
Log
(cystatin
C)
-
0.38
Log
(p-trace
protein)
0.30
0.41
Log
(I3-2-microglobulin)
0.31
0.62
0.42
Abbreviations:
AASK,
African
American
Study
of
Kidney
Disease;
B2M,
I3-2-microglobulin;
BTP,
p-trace
protein;
GFR,
glomerular
filtration
rate;
MDRD,
Modification
of
Diet
in
Renal
Disease;
mGFR,
measured
glomerular
filtration
rate.
Pearson
correlation
coefficients
and
partial
Pearson
correlation
coefficients
adjusted
for
mGFR
in
MDRD
Study,
AASK,
and
CRIC
(N=
3156).
All
p
<
0.001.
Serum
creatinine,
cystatin
C,
B2M,
and
BTP
were
transformed
to
the
GFR
scale
by
regressing
each
filtration
marker
on
mGFR
(after
log-transformation
of
markers
and
mGFR)
to
allow
for
comparisons
across
study-specific
single-marker
eGFRs
(eGFR
cr
,
eGFR
cys
,
eGFRB2M,
and
eGFR
B
-
r
p).
Other
factors
typically
used
in
estimation
equations
(eg,
age,
sex,
and
race)
were
not
incorporated.
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
BUN
Cr,
CysC,
BTP,
age
sex,
race
MDRD
&
AASK
Estimate
mGFR
by
closest
other
mGFR
in
AASK
&
MDRD
Creatinine
BTP
$
cysc
MDRD,
0
*
ti
tGFR
by
mGFR
AASK
&
MDRD
(042
weeks
apart)
Prec
is
ion,
R
MS
E
o
f
log
GFR
0.350
0.300
0.250
0.200
0.150
0.100
0.050
0.000
Estimate
mGFR
with
correlated
markers
(r=0.42)
Estimate
mGFR
with
uncorrelated
markers
(r=0)
Estimate
tGFR
with
correlated
markers
(r=0.42)
Estimate
TGFR
with
uncorrelated
markers
(r=0)
72
Inker
et
al
Factor
of
Interest
IQR
Geometric
Mean
Percent
Difference
in
Filtration
Ma
ker
(95%
Confidence
Interval)
Creatinine
Cystatin
C
BrMicroglobulin
13-Trace
Protein
Age
17.1
-6.6
(-7.5,
-5.6)
-1
(-2,
-0.1)
-0.5
(-1.8,
0.9)
-3.2
(-4.7,
-1.6)
Male
(vs
female)
23.1
(20.8,
25.5)
6.3
(4.6,
8.1)
2.7
(0.1,
5.2)
12.6
(9.3,
16)
Black
(vs
not
black)
-
ih
.
10.3
(7.6,
13)
-1.3
(-3.1,
0.6)
-5.8
(-8.6,
-3)
-9.2
(-11.9,
-6.4)
Height
(cm)
15.2
1.3
(-0.3,
2.8)
-1.3
(-2.6,
0)
-0.1
(-2.1,
2)
0.6
(-1.8,
3)
Weight
(kg)
26.3
2.3
(1.1,
3.4)
2.9
(1.9,
4)
1.2
(-0.3,
2.7)
-7.4
(-9.1,
-5.7)
Diabetes
(yes
vs
no)
-1.9
(-3.8,
0)
1.7
(0,
3.5)
2.1
(-0.3,
4.7)
2.2
(-0.8,
5.3)
Smoking
(current
vs
not
current)
-0.6
(-2.4,
1.2)
7.1
(5:37879)
- /
6.2
(3.6,
8.8)
-0.5
(-3.4,
2.5)
Serum
Albumin
(g/dL)
0.6
3.3
(2.2,
4.5)
1
(-2.2,
0.1)
-2.5
(-4,
-0.9)
-4.0
(-5.6,
-2.4)
Log
C-Reactive
protein
(mg/L)
1.7
-1
(-2,0)
2.8
(1.9,
3.9)
3.4
(1.8,
5.1)
0.3
(-1.1,
1.8)
Urine
creatinine
(mg/kg/d)
7.3
6.8
(5.4,
8.1)
-1
(-2.2,
0.1)
-1.2
(-2.7,
0.4)
0.6
(-1.4,
2.6)
Figure
3.
Non-GFR
determinants
of
LMW
serum
proteins
in
CKD
studies.
Percent
difference
in
levels
of
filtration
markers,
adjusted
for
measured
glomerular
filtration
rate
(mGFR),
mGFR
measurement
error,
study,
and
full
multivariable
adjustment
in
MDRD
Study,
AASK,
and
CRIC
(N=
3156).
All
predictors
listed
in
the
first
column
are
included
as
part
of
the
multivariable-
adjusted
model.
In
addition,
log-mGFR,
urine
urea
nitrogen,
and
log
urine
protein
were
included
in
the
model.
Percent
differ-
ence
in
levels
of
filtration
markers
for
an
interquartile
range
(IQR)
increase
in
continuous
variables, defined
as
the
difference
between
the
25th
and
75th
percentiles.
Strength
of
association
for
statistically
significant
results
is
indicated
by
color:
red,
strong
(absolute
change
>10%);
orange,
intermediate
(absolute
change
5%-10%
inclusive);
and
yellow,
weak
(absolute
change
<5%).
Modified
with
permission
from
Non-GFR
Determinants
of
Low-Molecular-Weight
Serum
Protein
Filtration
Markers
in
CKD.
26
Abbreviations:
AASK,
African
American
Study
of
Kidney
Disease;
CRIC,
chronic
renal
insufficiency
cohort;
GFR,
glomerular
filtration
rate;
LMW,
low
molecular
weight;
MDRD,
Modification
of
Diet
in
Renal
Disease.
(For
interpretation
of
the
references
to
color
in
this
figure
legend,
the
reader
is
referred
to
the
Web
version
of
this
article.)
biologically
acts
as
an
enzyme
promoting
the
conversion
of
prostaglandin
Fk
to
prostaglandin
D2.
B2M
is
a
100
aminoacid
protein."
B2M
is
the
beta-chain
of
major
histo-
compatibility
complex
class
I
molecules
that
is
generated
in
nucleated
cells
and
is
required
for
the
transport
of
major
histocompatibility
complex
I
heavy
chains
from
the
endo-
plasmic
reticulum
to
the
cell
surface.
Over
the
years,
there
has
been
interest
in
BTP
and
B2M
as
filtration
markers.
When
we
began
the
search
for
novel
filtration
markers,
data
on
their
performance
in
estimating
GFR
compared
to
creatinine
showed
potential
promise,
but
more
investi-
gation
was
needed.
Thus
far
we
have
developed
GFR
estimating
equations
based
on
B2M,
BTP,
and
the
combination
(eGFRB2m,
eGFR
BTh
and
eGFR
B2m
_
BTD
respectively)
in
CKD
popula-
tions
(Table
2).
As
expected,
equations
including
both
Estimated
Performance
of
a
"GFR
Panel"
2
3
4
5
6
7
9
10
Number
of
Markers
Figure
4.
Theoretical
effects
of
the
addition
of
filtration
markers
on
precision
(RMSE)
of
regressions
relating
filtration
markers
to
mGFR
or
true
GFR.
Simulations
show
the
effect
of
additional
markers
on
model
performance.
Lines
connect
simulations
based
on
data
from
MDRD
Study
and
AASK.
Two
assumptions
were
used
in
the
simulations:
first,
after
accounting
for
cor-
relation
of
the
markers
with
mGFR,
the
markers
have
a
mild-to-moderate
correlation
with
each
other
(r=
0.42);
and
second,
after
accounting
for
correlation
of
the
markers
with
mGFR,
the
markers
are
not
correlated
with
each
other
(r=
0).
Diamonds
indicate
RMSE
for
single
markers
on
the
left
and
another
mGFR
(mean
[SD]
62
[19]
days
apart)
on
the
right.
Bottom
2
lines
show
RMSE
is
lower
in
models
for
true
GFR
(tGFR).
Creatinine
and
cystatin
C
are
correlated
even
conditional
on
true
GFR,
which
diminishes
the
effect
of
their
combination.
The
2nd
and
4th
lines
demonstrate
the
additional
improvement
if
uncorre-
lated
markers
are
added
into
a
panel.
Abbreviations:
AASK,
African
American
Study
of
Kidney
Disease;
GFR,
glomerular
filtra-
tion
rate;
MDRD,
Modification
of
Diet
in
Renal
Disease;
mGFR,
measured
glomerular
filtration
rate;
SD,
standard
deviation.
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
Estimated
GFR
From
a
Panel
of
Filtration
Markers
73
BTP
and
B2M
had
smaller
coefficients
for
each
filtration
marker
than
the
equation
using
only
one
filtration
marker,
and
coefficients
for
age
and
sex
for
BTP
were
no
longer
sig-
nificant
when
B2M
was
included
in
the
equation.
Howev-
er,
we
found
that
BTP
and
B2M
do
not
improve
GFR
estimation
beyond
currently
available
equations
including
creatinine
and
cystatin
C,
although
the
combined
equation
is
less
reliant
on
demographic
factors,
including
race,
than
eGFR,_
cys
(Table
2).
25
Table
3
summarizes
the
correlations
among
the
filtration
markers
in
the
CKD
studies.
26
Correlations
of
B2M
and
BTP
with
mGFR
were
intermediate
between
those
of
cys-
tatin
C
and
creatinine.
After
accounting
for
mGFR,
B2M
and
cystatin
C
had
the
highest
correlation
with
each
other,
whereas
creatinine
had
the
lowest
correlations
to
the
other
markers.
These
findings
are
consistent
with
findings
in
two
community-based
studies
in
participants
with
higher
GFR.
27
In
Figure
3,
we
summarize
the
associations
of
de-
mographic
and
clinical
factors
with
the
B2M
and
BTP
inde-
pendent
of
mGFR
in
the
CKD
studies.
Like,
cystatin
C,
the
magnitude
of
associations
of
age,
sex,
and
race
with
B2M
and
BTP
was
less
strong
than
with
creatinine.
Findings
from
two
community-based
studies
at
higher
GFR
also
showed
smaller
associations
of
age
and
sex
with
B2M
and
BTP
than
with
creatinine,
but
no
significant
associa-
tion
of
race
with
B2M
and
BTP.
The
association
of
higher
CRP
with
higher
B2M
was
consistent
across
studies,
but
there
was
variation
in
other
associations.
Altogether,
these
results
indicate
that
much
remains
unknown
about
factors
associated
with
non-GFR
determinants
of
B2M
and
BTP.
Candidate
Metabolites
Even
as
we
began
the
evaluation
of
LMW
proteins,
we
anticipated
that
additional
markers
would
be
needed
to
fulfill
the
goals
of
a
panel
eGFR.
The
metabolites
we
had
identified
in
the
literature
had
cumbersome
assays
and
did
not
add
substantially
to
improved
performance.
We
then
turned
to
metabolomic
discovery
to
identify
new
markers.
We
selected
AASK
participants
with
consistent
mGFR
at
the
48-month
follow-up
visit
as
our
discovery
population
using
the
method
of
Metabolon
Inc,
a
triple
quadrupole
gas
chromatograph/mass
spectrometer
(GC/
MS)
and
an
updated
triple
quadrupole
liquid
chromato-
graph/mass
spectrometer)-based
metabolomics
platform
to
provide
semiquantifiable
results
on
600
metabolites.
We
found
18
metabolites
(13
known
and
5
unknown)
that
had
higher
correlations
with
mGFR
than
creatinine.
We
verified
these
results
in
an
ancillary
study
of
the
Multi-Ethnic
Study
of
Atherosclerosis
(MESA-Kidney)
and
found
similar
results.
Metabolites
with
stronger
corre-
lations
in
AASK
were
more
likely
to
have
stronger
correla-
tions
in
MESA-Kidney.
For
the
15
most
promising
metabolites
with
available
pure
standards
identified
in
AASK
and
other
ongoing
studies,
Metabolon
Inc
developed
targeted
assays
for
ab-
solute
quantification.
Using
the
targeted
assays
in
both
AASK
and
MESA-Kidney,
we
developed
GFR
estimating
equations
using
these
metabolites
with
or
without
creati-
nine
and
demographics
in
each
study
alone
and
in
the
combined
data
set.
Compared
to
eGFRcr
using
the
CKD-EPI
equation,
the
panels
both
with
and
without
de-
mographics
were
more
accurate
in
both
AASK
and
MESA-Kidney.
In
our
view,
these
results
provide
proof
of
the
concept
that
a
panel
of
novel
metabolites
that
does
not
include
creatinine
or
demographics
can
accurately
estimate
mGFR.
Given
their
high
correlations
with
mGFR,
we
do
not
anticipate
that
clinical
factors
will
be
needed
to
opti-
mize
the
performance
of
the
panel
eGFR.
However,
more
work
is
required
—in
particular
these
results
need
to
be
further
verified
in
other
populations,
we
need
to
under-
stand
more
about
their
correlations
of
these
novel
markers
with
each
other
and
factors
associated
with
their
non-GFR
determinants
as
we
have
done
with
the
LMW
proteins,
and
we
need
to
compare
its
performance
to
eGFR
cys
and
eGFR,
-
cys
In
order
to
use
panel
eGFR
in
clinical
practice,
the
test
will
need
to
be
validated
for
CLIA
approval,
and
this
process
is
underway.
Subsequent
testing
can
then
lead
to
submission
to
the
FDA
for
its
approval.
We
antici-
pate
that
the
advantage
of
panel
eGFR
will
be
greatest
in
populations
in
which
creatinine
generation
is
altered
or
less
well
characterized,
for
example,
extremes
of
muscle
mass
and
diet,
races
other
than
blacks
or
whites,
or
diffi-
culty
in
specification
of
race,
in
which
eGFR,
and
eGFR,-
cys
are
biased.
This
will
require
extensive
evalua-
tion.
HOW
DO
WE
ENVISION
THAT
PANEL
ESTIMATED
GFR
WOULD
BE
USED?
GFR
is
used
for
clinical
practice,
research,
and
public
health
policy.
We
anticipate
that
a
panel
eGFR
would
have
wide
spread
implications.
Clinical
We
anticipate
that
eGFR,
would
continue
to
be
the
first
test
used
to
assess
kidney
function.
Panel
eGFR
would
be
used
for
confirmation,
as
is
recommended
by
the
KDIGO
guidelines
for
the
use
of
eGFR
cys
,
eGFR,_
cys
and
clearance
measurements.
Circumstances
when
confirma-
tion
is
suggested
include
identification
of
CKD
or
when
clinical
decisions
need
to
be
made
that
would
alter
treat-
ment
plans,
such
as
use
of
radiological
tests
with
contrast
agents,
preemptive
transplants,
or
dosages
of
toxic
medi-
cations
(such
as
chemotherapy
or
antibiotics).
In
addition,
in
conditions
in
which
creatinine
generation
is
altered
or
not
well
characterized,
eGFR
cys
or
panel
eGFR
could
be
a
first
test.
Ultimately,
recommendations
for
when
and
how
to
implement
panel
eGFR
in
practice
will
need
to
also
consider
cost
of
the
test
and
reimbursement
by
insur-
ance
companies.
Research
More
accurate
GFR
estimates
will
be
helpful
for
more
effi-
cient
trials
of
CKD
progression.
With
more
precise
end
points,
trials
can
be
potentially
smaller
and
shorter,
and
therefore
less-expensive,
which
will
ultimately
lead
to
faster
development
of
drugs
to
slow
CKD
progression.
A
panel
eGFR
that
does
not
include
creatinine
will
also
be
useful
for
drugs
that
affect
muscle
mass
or
diet
in
the
eval-
uation
of
benefit
or
harm.
For
example,
drugs
used
for
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
weight
loss
cannot
be
evaluated
for
slowing
progression
or
nephrotoxicity
using
eGFR,
as
weight
loss
is
associated
with
decreased
creatinine
generation
and
increased
eGFR,
independent
of
effects
on
mGFR.
A
panel
eGFR
not
based
on
creatinine
may
also
improve
estimates
of
prognosis.
As
we
and
others
have
shown,
while
low
eGFR,
is
associated
with
increased
risk
of
end-stage
kidney
disease
and
mortality,
very
high
eGFR,
is
also
associated
with
increased
risk
for
mortality.
Lower
eGFR
cys
,
eGFR
B2m
,
and
eGFR
BTP
are
associated
with
12.
higher
risk
for
mortality
than
eGFR,
but
not
an
increased
risk
for
mortality
at
very
high
eGFR
cs
,
eGFR
B2m
,
or
eGFR
BTD
Our
interpretation
is
that
the
higher
mortality
risk
associated
with
very
high
eGFR,
but
not
with
other
markers
reflects
increased
risk
in
the
frail
elderly
who
have
reduced
musde
mass
and
creatinine
generation
but
not
very
high
true
GFR.
4143
Possibly,
the
higher
mortality
risk
for
eGFR
cys
,
eGFR
B2m
and
eGFR
BTP
than
eGFR,
at
lower
eGFR
also
represents
confounding
by
reduced
musde
mass
and
creatinine
generation
affecting
eGFR,
15.
but
not
other
markers.
Alternatively,
the
higher
mortality
risk
at
lower
eGFR
might
represent
confounding
by
factors
associated
with
non-GFR
determinants
of
cystatin
C,
B2M,
and
BTP,
such
as
smoking,
inflammation,
and
obesity.
In
principal,
a
panel
of
eGFR
would
more
likely
reflect
true
GFR,
and
therefore,
prognosis
based
on
panel
eGFR
would
more
likely
reflect
the
contribution
of
kidney
disease
to
the
risk
of
adverse
outcomes.
18.
8.
9.
10.
11.
13.
14.
16.
17.
74
Inker
et
al
Public
Health
More
accurate
estimates
will
allow
for
more
refined
prev-
alence
estimates.
Higher
CKD
stages
(lower
GFR)
is
asso-
ciated
with
higher
costs."
For
example,
in
some
health
systems,
provider
payment
increases
with
CKD
Stage"
but
is
primarily
defined
by
eGFRcr.
Panel
eGFR
could
lead
to
more
accurate
classification
of
those
patients
who
have
low
true
GFR
and
better
use
of
resource
planning
by
government,
insurance
companies,
and
provider
groups.
REFERENCES
1.
Stevens
LA,
Coresh
J,
Greene
T,
Levey
AS.
Assessing
kidney
func-
tion-measured
and
estimated
glomerular
filtration
rate.
N
Engl
I
Med.
2006;354(23):2473-2483.
2.
Levey
AS,
Inker
LA,
Coresh
J.
GFR
Estimation:
from
Physiology
to
public
health.
Am
I
Kidney
Dis.
2014;63(5):820-834.
3.
Kidney
Disease:
Improving
Global
outcomes
(ICDIGO).
KDIGO
2012
clinical
practice
guideline
for
the
evaluation
and
Management
of
chronic
kidney
disease.
Kidney
Int
Suppl.
2013;3(1):1-150.
4.
Miller
WG.
Estimating
glomerular
filtration
rate.
Clin
Chem
Lab
Med.
2009;47(9):1017-1019.
5.
Steindel
SJ,
Rauch
WJ,
Simon
MK,
Handsfield
J.
National
Inventory
of
clinical
laboratory
testing
Services
(NICLTS).
Development
and
test
distribution
for
1996.
Arch
Pathol
Lab
Med.
2000;124(8):1201-
1208.
6.
Miller
WG.
Reporting
estimated
GFR:
a
laboratory
perspective.
Am
I
Kidney
Dis.
2008;52(4):645-648.
7.
Smith
HW.
Measurement
of
the
filtration
rate.
In:
The
Kidney:
Struc-
ture
and
Function
in
Health
and
Disease.
NY,
NY:
Oxford
University
Press;
1951:39-62.
Stevens
LA,
Levey
AS.
Measured
GFR
as
a
confirmatory
test
for
esti-
mated
GFR.
I
Am
Soc
Nephrol.
2009;20(11):2305-2313.
Soveri
I,
Berg
UB,
Bjork
J,
et
al.
Measuring
GFR:
a
systematic
review.
Am
I
Kidney
Dis.
2014;64(3):411-424.
Davies
DF,
Shock
NW.
The
variability
of
measurement
of
insulin
and
diodrast
tests
of
kidney
function.
I
Clin
Invest.
1950;29(5):491-
495.
Florijn
KW,
Barendregt
JN,
Lentjes
EG,
et
al.
Glomerular
filtration
rate
measurement
by
"single-shot"
injection
of
inulin.
Kidney
Int.
1994;46(1):252-259.
Levey
AS,
Greene
T,
Schluchter
MD,
et
al.
Glomerular
filtration
rate
measurements
in
clinical
trials.
Modification
of
diet
in
renal
disease
study
group
and
the
diabetes
control
and
Complications
trial
research
group.
I
Am
Soc
Nephrol.
1993;4(5):1159-1171.
Agarwal
R.
Ambulatory
GFR
measurement
with
cold
iothalamate
in
adults
with
chronic
kidney
disease.
Am
I
Kidney
Dis.
2003;41(4):752-
759.
Chowdhury
TA,
Dyer
PH,
Bartlett
WA,
et
al.
Glomerular
filtration
rate
determination
in
diabetic
patients
using
iohexol
clearance-com-
parison
of
single
and
multiple
plasma
sampling
methods.
Clin
Chim
Acta.
1998;277(2):153-158.
Tan
GD,
Lewis
AV,
James
TJ,
et
al.
Clinical
usefulness
of
cystatin
C
for
the
estimation
of
glomerular
filtration
rate
in
type
1
diabetes:
reproducibility
and
accuracy
compared
with
standard
measures
and
iohexol
clearance.
Diabetes
Care.
2002;25(11):2004-2009.
Gaspari
F,
Perico
N,
Matalone
M,
et
al.
Precision
of
plasma
clearance
of
iohexol
for
estimation
of
GFR
in
patients
with
renal
disease.
I
Am
Soc
Nephrol.
1998;9(2):310-313.
Kwong
YT,
Stevens
LA,
Selvin
E,
et
al.
Imprecision
of
urinary
iotha-
lamate
clearance
as
a
gold-standard
measure
of
GFR
decreases
the
diagnostic
accuracy
of
kidney
function
estimating
equations.
Am
I
Kidney
Dis.
2010;56(1):39-49.
Earley
A,
Miskulin
D,
Lamb
EJ,
Levey
AS,
Uhlig
K.
Estimating
equations
for
glomerular
filtration
rate
in
the
era
of
creatinine
standardization:
a
systematic
review.
Ann
Intern
Med.
2012;156(11):785-795.
19.
Inker
LA,
Okparavero
A.
Cystatin
C
as
a
marker
of
glomerular
filtration
rate:
prospects
and
limitations.
Curr
Opin
Nephrol
Hyper-
tens.
2011;20(6):631-639.
20.
Tenstad
0,
Roald
AB,
Grubb
A,
Aukland
K.
Renal
handling
of
ra-
diolabelled
human
cystatin
C
in
the
rat.
Scand
I
Clin
Lab
Invest.
1996;56(5):409-414.
21.
Baran
D,
Tenstad
0,
Aukland
K.
Localization
of
tubular
uptake
segment
of
filtered
Cystatin
C
and
Aprotinin
in
the
rat
kidney.
Acta
Physiol
(oxft.
2006;186(3):209-221.
22.
Inker
LA,
Schmid
CH,
Tighiouart
H,
et
al.
Estimating
glomerular
filtration
rate
from
serum
creatinine
and
cystatin
C.
N
Engl
I
Med.
2012;367(1):20-29.
23.
Fan
L,
Inker
LA,
Rossert
J,
et
al.
Glomerular
filtration
rate
estimation
using
cystatin
C
alone
or
combined
with
creatinine
as
a
confirma-
tory
test.
Nephrol
Dial
Transpl.
2014;29(6):1195-1203.
24.
Anderson
AH,
Yang
W,
Hsu
CY,
et
al.
Estimating
GFR
among
par-
ticipants
in
the
chronic
renal
Insufficiency
Cohort
(CRIC)
study.
Am
I
Kidney
Dis.
2012;60(2):250-261.
25.
Inker
LA,
Tighiouart
H,
Coresh
J,
et
al.
GFR
estimation
using
beta-
trace
protein
and
beta2-microglobulin
in
CICD.
Am
I
Kidney
Dis.
2016;67(1):40-48.
26.
Liu
X,
Foster
MC,
Tighiouart
H,
et
al.
Non-gfr
determinants
of
low-
molecular-weight
serum
protein
filtration
markers
in
CKD.
Am
I
Kidney
Dis.
2016;68(6):892-900.
27.
Foster
MC,
Levey
AS,
Inker
LA,
et
al.
Non-GFR
Determinants
of
Low
Molecular
Weight
Serum
Protein
Filtration
Markers
in
the
Elderly:
AGES-Kidney
and
MESA-Kidney.
In
Press;
28.
Knight
EL,
Verhave
JC,
Spiegelman
D,
et
al.
Factors
influencing
serum
cystatin
C
levels
other
than
renal
function
and
the
impact
on
renal
function
measurement.
Kidney
Int.
2004;65(4):1416-1421.
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75
Estimated
GFR
From
a
Panel
of
Filtration
Markers
75
29.
Macdonald
J,
Marcora
S,
Jibani
M,
et
al.
A
GFR
estimation
using
cys-
37.
tatin
C
is
not
independent
of
body
composition.
Am
I
Kidney
Dis.
2006;45(5):712-719.
30.
Mathisen
UD,
Melsom
T,
Ingebretsen
OC,
et
al.
Estimated
GFR
as-
38.
sociates
with
cardiovascular
risk
factors
independently
of
measured
GFR.
I
Am
Soc
Nephrol.
2011;22(5):927-937.
31.
Rule
AD,
Bailey
KR,
Lieske
JC,
Peyser
PA,
Turner
ST.
Estimating
the
39.
glomerular
filtration
rate
from
serum
creatinine
is
better
than
from
cystatin
C
for
evaluating
risk
factors
associated
with
chronic
kidney
disease.
Kidney
Int.
2013;83(6):1169-1176.
40.
32.
Thurlow
JS,
Abbott
KC,
Linberg A,
et
al.
SCr
and
SCysC
concentra-
tions
before
and
after
traumatic
amputation
in
male
soldiers:
a
case-
control
study.
Am
I
Kidney
Dis.
2014;63(1):167-170.
33.
Brown
CS,
Kashani
KB,
Clain
JM,
Frazee
EN.
Cystatin
C
falsely
41.
underestimated
GFR
in
a
critically
ill
patient
with
a
new
diagnosis
of
AIDS.
Case
Rep
Nephrol.
2016;2016:9349280.
34.
Eckfeldt
JH,
Karger
AB,
Miller
WG,
Rynders
GP,
Inker
LA.
Perfor-
42.
mance
in
measurement
of
serum
cystatin
C
by
laboratories
Partici-
pating
in
the
College
of
American
Pathologists
2014
CYS
Survey.
43.
Arch
Pathol
Lab
Med.
2015;139(7):888-893.
35.
Juraschek
SP,
Coresh
J,
Inker
LA,
et
al.
Comparison
of
SET=
concentrations
of
beta-trace
protein,
beta2-microglobulin,
cystatin
C,
44.
and
creatinine
in
the
US
population.
Clin
I
Am
Soc
Nephrol.
2013;8(4)584-592.
36.
Filler
G,
Kusserow
C,
Lopes
L,
Kobrzynski
M.
Beta-trace
protein
as
a
45.
marker
of
GFR-history,
indications,
and
future
research.
Clin
Bio-
chem.
2014;47(13-14):1188-1194.
Hoffmann
A,
Nimtz
M,
Conradt
HS.
Molecular
characterization
of
beta-trace
protein
in
human
serum
and
urine:
a
potential
diagnostic
marker
for
renal
diseases.
Glycobiology.
1997;7(4):499-506.
Melegos
DN,
Grass
L,
Pierratos
A,
Diamandis
ER.
Highly
elevated
levels
of
prostaglandin
D
synthase
in
the
serum
of
patients
with
renal
failure.
Urology.
1999;53(1):32-37.
Viberti
GC,
Keen
H,
Mackintosh
D.
Beta
2-microglobulinaemia:
a
sensitive
index
of
diminishing
renal
function
in
diabetics.
Br
Med
I
(din
Res
Ed.
1981;282(6258):95-98.
Coresh
J,
Inker
LA,
Sang
Y,
et
al.
Multiple
metabolites
correlate
more
strongly
with
measured
glomerular
filtration
rate
than
creatinine:
a
Verification
study
[Abstract
#1713].
I
Am
Soc
Nephrol.
2016.
Shlipak
MG,
Matsushita
K,
Arnlov
J,
et
al.
Cystatin
C
versus
creat-
inine
in
determining
risk
based
on
kidney
function.
N
Engl
I
Med.
2013;369(10):932-943.
Stevens
LA,
Levey
AS.
Chronic
kidney
disease
in
the
elderly-how
to
assess
risk.
N
Engl
I
Med.
2005;352(20):2122-2124.
Inker
LA,
Coresh
J,
Sang
Y,
et
al.
Filtration
markers
as
predictors
of
end-stage
renal
disease
and
mortality
-individual
participant
data
meta-analysis.
Clin
I
Am
Soc
Nephrol.
2017;12(1):69-78.
United
States
Renal
Data
System
Annual
Data
Report
2017.
Chapter
6:
Healthcare
Expenditures
for
Persons
with
CKD.
Available
at
www.usrds.org
.
Accessed
November
16,
2017.
National
Institute
for
Health
and
Care
Excellence
NICE.
Chronic
kidney
disease
in
adults,
https://www.nice.org.uk/guidance/qs5
.
Accessed
April
17,
2017.
Adv
Chronic
Kidney
Dis.
2018;25(1):67-75