geisinger research talk incorporating addendum 090215

A Translational Survey of FactorsAffecting Renal & Cardiovascular

Outcomes in Chronic Kidney Disease

Kevin Ho, M.D.Associate Professor of Medicine &

Clinical and Translational Science

University of Pittsburgh School of Medicine & UPMC

Renal-Electrolyte DivisionClinical and Translational Science Institute

KHv0815

Channels Mediating RenalPotassium Secretion

to

Metabolic Sensors in the Regulation of Vascular Tone & Blood Pressure

Two Inward Rectifier (Kir) Channel Subfamilies

and Their Partners: ROMK and KATP

Clinical Diversity of Kir Channelopathies

KHv0815

Early Studies in Renal Potassium Transport

Clearance Studies in Whole Animals 1940’s – 1950’s(applied flame photometry, Hald PM 1947)

• Renal tubular K+ secretion can be dissociated from (and exceed) glomerular K+ filtration (McCance RA & Widdowson EM 1937)

• Reabsorption of K+ along nephron occurs proximal to K+ secretion• Na+ reabsorption is necessary for effective K+ secretion

(Berliner RW & Kennedy TJ 1948 – 1950; summarized in Berliner’s 1961 “Harvey Lecture”)

Micropuncture Studies of Single Tubules 1960’s - 1980’s(free-flow micropuncture, Walker AM et al. 1941; in vivo microperfusion, in vitro microperfusion)

Localized K+ handling processes to specific nephron segments

• K+ reabsorption -- proximal tubule, loop of Henle• K+ secretion -- distal tubule, cortical collecting duct• K+ secretion and reabsorption -- cortical/medullary collecting ducts

depending on dietary K+ intake

Giebisch GH, Wingo CS. Sem in Nephrol 2013;33:209-214KHv0815

Apical Secretory K+ Channels in TAL and CCD

• 30-pS and 70-pS K+ channels

• K+ recycling across apical

membrane– Required for NaCl reabsorption

– Generates lumen-positive potential

driving paracellular reabsorption of

Na+, Ca2+, Mg2+

– Hyperpolarize cell membrane

facilitating basolateral Cl- transport

• 35-pS K+ channel mediates bulk of

apical K+ secretion in CCD under

normal conditions

30-pS and 70-pS channels in TAL

and 35-pS channels in CCD

principal cells are absent in ROMK-

knockout mice

(+)

2 K+

3 Na+

K+

Cl-

Cl-

Na+

Mg2+Ca2+

Na+

K+

2 Cl-

K+30pS

K+70pS

K+146pS

TAL Cell

3 Na+

2 K+

(-)Na+

ALK+35pS

K+140pS

PrincipalCell

MR

AL

KHv0815

Functional Expression Cloning in Xenopus Laevis Oocytes

• Major breakthrough in K+ channel structure (Tempel BL, Timpe LC, Jan LY, Jan YN)

– Identification/cloning of Shaker gene in Drosophila melanogaster (1987)

– Validation by expression in Xenopus laevis oocytes (1988)

Kloc M. Methods 2010;51:1-2; Terhag J et al. Methods 2010;51:66-74

• At BWH, we used functional expression cloning in Xenopus laevis

oocytes to isolate a cDNA encoding an outer medullary K+ channel

(ROMK) starting with rat kidney (inner stripe, outer medulla) mRNA

KHv0815

KHv0815

Ho K…Hebert SC et al. Nature 1993;362:31-38

1992 ROMK1 Model & 21st Century Model

Ho K…Hebert SC et al. Nature 1993;362:31-38

Welling PA, Ho K. Am J Physiol 297:F849-F863 (2009)KHv0815

http://ajprenal.physiology.org/content/vol297/issue4/images/large/zh20080955960003.jpeg


ROMK1 Conserves Pore-Forming H5 & P Regions

Yellen G. Nature 2002;419:35-42

Original figure from K. Ho lab notebook, April, 1992

P Region

ROMK P segment

Welling PA, Ho K. Am J Physiol

297:F849-F863 (2009)KcsA Pore

KHv102313

Distribution & Function of ROMK Isoforms in Nephron

• Alternative splicing and promoter usage of KCNJ1 (11q24) result in three ROMK channel subunits characterized by different NH2-termini and differential expression but identical biophysical characteristics

– TAL: ROMK2/3 mediate K+ recycling enabling NaCl, Ca2+/Mg2+ reabsorption

– DCT/CCD: ROMK1 is the primary mediator of distal renal K+ secretionBoim MA, Ho K et al. Am J Physiol 1995;268:F1132-F1140

Welling PA, Ho K. Am J Physiol 2009;297:F849-F863KHv0815

Aldosterone & K+ Intake: ROMK1 Surface Density

K+ adaptation with increased K+

intake– post-translational

expression of active channels

• Expression (Aldo, K+ Intake)

– PKA/SGK1 phosphorylation

of N-term S44 (R39XRXXS44) overrides

C-terminal ER retention signal (YxNPxFV)

• Expression (dietary K+ restriction)

– WNK Kinases (4, L1): clathrin-

dependent endocytosis (via C-term

YxN373PxFV-motif binding to ARH)

– PTyrKase (cSrc, cYes) phosphorylation

of Y337 also leads to clathrin-dependent

endocytosis

– Channel endocytosis is negated by

Aldo / K+ intake

• SGK1: WNK4 inhibition

• KS-WNK1: L-WNK1 inhibitionWelling PA. Sem in Nephrol 2013;33:215-228KHv0815

Inward Rectifier (Kir) K+ Channel Family

• Kir channels have diverse functions:

Myocardial function, vascular tone, heart

rate, insulin release, neuronal signaling,

and electrolyte transport across epithelia

• Seven subfamilies Kir1.0 to Kir7.0:

Characterized by differences in degree of

rectification and regulation by specific

cellular signals

Bichet D et al. Nature Rev Neurosci 2003;4:957-967

Rapedius M et al. EMBO Reports 2006;7:611-616

Hibino H et al. Physiol Rev 2010:90:291-366

Tetrameric

ROMK

Channel

KHv0815

Human Kir Diseases (Kir Channelopathies)

Identified Kir channel mutations: growing list of inherited human

disorders characterized by cardiac arrhythmias, cardiomyopathy,

seizures, deafness, hyperinsulinemia, diabetes, hyperaldosteronism,

renal Na+ and Cl- wasting, hypomagnesemia

Pattnaik BR et al. Mol Genet Metab 2012;105:64-72KHv0815

*Excludes diseases by genetic association

(-) Loss of function, (+) Gain of function

KHv0815

Kir Channel Mutations & Human Genetic Diseases*Kidney (& Nervous System)• Bartter’s Syndrome Type 2 (Kir1.1/KCNJ1/-)

• SeSAME/EAST Syndrome (Kir4.1/KCNJ10/-)

Heart• Andersen-Tawil Syndrome (Kir2.1/KCNJ2/-)

• Short Q-T Syndrome (Kir2.1/KCNJ2/-)

• Brugada Syndrome, Early Repolarization Syndromes (Kir6.1/KCNJ8/+)

Endocrine• 1° Aldosteronism/Adrenal Hyperplasia, Aldosterone-Producing Adenomas (Kir3.4/KCNJ5/-)

• Hyperinsulinemic Hypoglycemia of Infancy/Congenital Hyperinsulinism (Kir6.2/KCNJ11/-)

• Neonatal Diabetes (Permanent & Transient Types) (Kir6.2/KCNJ11/+)

Eye• Snowflake Vitreoretinal Degeneration (Kir7.1/KCNJ13/+)

Co-Subunit ABCC8/ABCC9 Mutations of KATP Channels (Kir6.0)

ABCC8 (SUR1)• Hyperinsulinemic Hypoglycemia of Infancy/Congenital Hyperinsulinism (SUR1/ABCC8/-)

• Neonatal Diabetes (Permanent & Transient Types) (SUR1/ABCC8/+)

ABCC9 (SUR2)• Cantu Syndrome (SUR2/ABCC9/+)

• Adrenergic Atrial Fibrillation (SUR2/ABCC9/-)

• Idiopathic Dilated Cardiomyopathy (SUR2/ABCC9/-)

Bartter Syndrome Type 2: Kir1.1 MutationsKCNJ1 loss-of-function mutations

(autosomal recessive) result in

Bartter’s Type 2 (Antenatal Bartter’s)

– NaCl wasting, hypokalemia

– metabolic alkalosis, hypercalciuria

– Renin/Aldosterone

– Normotension/hypotension

– Neonatal hyperkalemia

• Mutations alter PKA phosphorylation,

• PIP2 binding, channel Po,

• K+ conduction, pH sensitivity, channel expression

• KCNJ1 knockout mice (similar phenotype) exhibit no 35pS or 70pS apical K+ channels

Welling PA, Ho K. Am J Physiol 297:F849-F863 (2009)

cAMPP450 20-HETE

Na+

K+

2 Cl-

K+

ROMK

NKCC2ClC-Kb

2 K+

Mg2+Ca2+

Barttin

KCCNa+

(+)

3 Na+

K+

Cl-

Cl-

(-)

(-)

CaSR

TAL

KHv0815



Primary Aldosteronism: Kir3.4 Mutations KCNJ5 loss-of-function mutations (autosomal

recessive)

– Aldosterone-producing adrenocortical adenomas (Conn’s Syndrome)

– Hereditary 1º aldosteronism/adrenal hyperplasia

– Kir3.4 subunits (KCNJ5) form homotetramers or heterotetramers with Kir3.1 subunits

– Mutations (P/H5 regions, M2 helix) reduce pore K+ selectivity resulting in Na+ conduction

• G151R: K+:Na+ permeability ratio = 1:1

• Adrenal glomerulosa cell depolarization, Ca2+

entry, aldosterone synthesis and cell proliferation

Choi M et al. Science 2011;331:768-772KHv102313

SESAME/EAST Syndrome: Kir4.1 Mutations

KCNJ10 loss-of-function mutations (autosomal recessive)

• Early-onset seizures (age 3-4 mo), ataxia, progressive axonal neuropathy (hypomyelination), developmental delay, mental retardation, sensorineural deafness (KCNJ10 knockout mice similar)

• Hypokalemia, metabolic alkalosis, severe hypomagnesemia, renin/aldosterone, renal Na+, K+, Mg2+ wasting

• Kir4.1 K+ channel subunits (CNS glial cells, inner ear, kidney) with Kir5.1 subunits form heteromultimeric Kir4.1/Kir5.1 channels

• Mediate basolateral K+ recycling across the DCT cell basolateral membrane• Mutations reduce basolateral K+ recycling and therefore Na+-K+-ATPase

function disrupting Na+ and Cl- transportSchollUI et al. PNAS106:5842-5847, 2009KHv0815

ABCC9 Gene PolymorphismsPredict Risk of Albuminuria in

Type 1 Diabetic Kidney Disease

Fotinou C et al. FEBS J 2013;280:1051-1063

Homology modelwithout Kir channel

SUR2B TetramerSingle-particle

electron microscopy

KATP Channels

Metabolic sensors & transducers

Couple cellular metabolism to membrane excitability by sensing changes in intracellular [ATP]/[ADP]

[ATP]i/[ADP]i KATP channel

inhibition membrane depolarization

• 3 Classes:

pancreatic β-cell/neuronal, cardiac myocyte, vascular/smooth muscle

Hetero-octameric complexes

Kir channel subunits 4x [Kir6.1 or 6.2]

+ Sulfonylurea receptors 4x [SUR1 or

SUR2A/2B]4

• Pancreatic β-cell:

Channel inhibition by hyperglycemia[ATP]i/[ADP]i or sulfonylureas depolarization Ca2+ channel activity insulin release

Fotinou C et al. FEBS J 2013;280:1051-1063KHv0815

K+

Top view

Sulfonylureareceptor (SUR)subunits (x4)

InwardrectifierK+ channelsubunits (x4)

Cardiac & VSM-KATP channels

• In heart & vasculatureCardiac KATP channels -- cardiac myocyte and nodal/conduction cellsVascular smooth muscle (VSM) KATP channels -- vascular smooth muscle/endothelial cells of arteries, arterioles

• In kidneyVSM KATP channels -- vascular smooth muscle/endothelial cells of renal arteries/arterioles, glomeruli, podocytes, mesangial cells

• Vascular smooth muscle (VSM) KATP channelsActivation of channels by hypoxia, reduced intracellular [ATP]/[ADP], minoxidil or nicorandil results in myocyte relaxation and vasodilation

KHv0815

Hyperglycemia

K+

VSM KATPChannel

Ca2+

Glucose

GLUT1

L-Type Ca2+ Channel

[ATP]/[ADP]

(-)

(+)PKC

DAG

TGF-β +ΔVm

VSM KATP Channel

SUR2BKir6.1 K+ channel

Side view Top view

MikhailovMV et al. EMBO J 2005;24:4166-4175

Sulfonylurea Receptor Genes

ABCC8 encodes high-affinity sulfonylurea receptor SUR1– 39 exons, 84.0 Kb (tandem: Kir6.2/KCNJ11, 3.46 Kb)

ABCC9 encodes low-affinity sulfonylurea receptors SUR2A, SUR2B– 38 corresponding exons, 135.63 Kb (tandem: Kir6.1/KCNJ8, 9.86 Kb)

– alternative use of 3‘ terminal exons 38A or 38B yields SUR2A or SUR2B

ABCC8 (SUR1)

5’ 3’

Diazoxide sensitivity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

NBD1 NBD2

Human Chromosome 11p15.1

ABCC9 (SUR2)

Diazoxide sensitivity

Residues 1516-1522 (mouse)ADP-activation

SUR2BSUR2A

NBD1 NBD2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38A 38B

3’5’

Human Chromosome 12p12.1

KHv0815

ABCC9 Human Disease• Mutations of ABCC9 and KCNJ8 genes, encoding SUR2A/SUR2B and

Kir6.1 subnits which assemble to form Cardiac and VSM KATP

channels, lead to human genetic disease

• In family studies, ABCC9 mutations result in:o Dilated cardiomyopathy-ventricular tachycardia (OMIM 608569)

o Adrenergic atrial fibrillation (OMIM 614050),

o Cantu syndrome (hypertrichotic osteochondrodysplasia) (OMIM 239850)

characterized by cardiac hypertrophy, pulmonary hypertension,

lymphedema

• In human association studies, ABCC9 SNPs are associated with:(i) hypertension, (ii) response to ARB antihypertension therapy, (iii) metabolic

syndrome, (iv) sleep duration

KHv0815Fotinou C et al. FEBS J 2013;280:1051-1063

ABCC9 in Type 1 Diabetic Kidney Disease• Hypothesis

Single nucleotide polymorphisms (SNPs) in putative transcription factor

binding motifs of the ABCC9 promoter associate with risk of incident

albuminuria in Type 1 diabetes: rs17335932, rs35677639

• Study population

Pittsburgh Epidemiology of Diabetes Complications (EDC) Study

483 childhood-onset (age <17) T1DM subjects, initially seen 1950-1980,

baseline evaluation 1986-1988, biennial followup >20 years

• Renal outcomes

o Urine albumin excretion rate (AER)

Determined on ≥2 of 3 timed urines (baseline and follow-up visits)

o Albuminuria = albuminuria (AER ≥20 µg/min, Ualb:Ucr ≥0.03 mg/mg) or

advanced diabetic kidney disease (sCr >5 mg/dl, dialysis, renal transplant)

Incident = 1st occurrence of albuminuria

Persistent = 2 consecutive biennial visits with albuminuria

o Renal function decline = decrease in eGFR (MDRD) >30 ml/min/1.73 m2

from baseline eGFR

o Low eGFR = first occurrence of eGFR (MDRD) <60 ml/min/1.73 m2

(excluding subjects with baseline low eGFR)

KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246

Increased Risk of Incident Albuminuria

• Subjects without baseline albuminuria (n = 268)

At baseline, 268 of 483 genotyped subjects exhibited no albuminuria

(AER <20 µg/min, Ualb:Ucr <0.03 mg/mg)

• Genotyping of rs17335932 (173*) and rs35677639 (356*)

Genotype frequencies: 173* AA (93.4%), 173* AG (6.6%), 173* GG (0%)

356* AA (43.3%), 356* AG (47.4%), 356* GG (9.3%)

Allelic frequencies: 173* A (96.7%), 173* G (3.3%)

356* A (67.0%), 356* G (33.0%)

• Genotypes associated with incident albuminuria

After 22 years of followup, rs17335932 AA and rs35677639 GG genotypes were

each associated with increased risk of incident albuminuria by comparison to

alternative genotypes


Combining Polymorphisms

• Combining genotypes of 2 polymorphisms

By combining 173* and 356* genotypes, all subjects (A1) were divided into 3

combined genotype groups:

– 173* AG + 356* AA/AG (Group 1) low + low risk– 173* AA + 356* AA/AG (Group 2) high + low risk

– 173* AA + 356* GG (Group 3) high + high risk

• Incident albuminuria by combined genotypes

– In combined low + low, high + low, and high + high risk genotype groups,

the incidence of albuminuria was 11.1%, 41.2%, and 58.3%, respectively

over 22 years of followup (p = 0.004)

– Persistent albuminuria exhibited a non-significant trend


Baseline Characteristics

Combined genotype groups

Subjects in combined genotype groups exhibited no differences in baseline

characteristics


Graded Probability of Incident Albuminuria

Kaplan-Meier survival analysis

• Significant differences by Log Rank test between combined genotype groups

• Group 1 (173* AG + 356* AA/AG), Group 2 (173* AA + 356* AA/AG), and

Group 3 (173* AA + 356* GG) showed high, intermediate, and low relative

probabilities for albuminuria-free survival, respectively (p = 0.004)


Multivariate Analyses of Incident Albuminuria

• Cox proportional hazard models were used to identify covariates

significantly associated with risk of incident albuminuria

• Adjusting for known covariates, combined rs17335932 and rs35677639

genotypes predicted the incidence of albuminuria

• Combining high risk, 173* AA and 356* GG, genotypes (Group 3)

conferred greatest risk (HR = 7.63, p = 0.008), while combining high

risk, 173* AA, and low risk, 356* AA/AG, genotypes (Group 2)

conferred an intermediate risk (HR = 4.71, p = 0.03)


Predicting Renal Function Decline

• The combined genotype group 173* AA/356* GG exhibiting greatest

risk also predicted renal function decline (decrease in eGFR (MDRD)

>30 ml/min/1.73 m2 from baseline eGFR) in Cox proportional hazard

models

• Did not predict low eGFR (<60 ml/min/1.73 m2 following baseline)


SNP 173* Affects Transcriptional Activity

• Rs17335932 G-allele (low risk) disrupts

a putative transcription factor binding

motif– Similar to LIM-only 2 GATA box -

regulating angiogenesis, tumor neo-

vascularization, hematopoiesis

• To test effect of each rs17335932 allele

on transcription efficiency

– Co-transfected firefly luciferase-reporter

vectors incorporating human ABCC9

5’UTR with A-(high-risk/WT) or G-(low-

risk/Mut) alleles and Renilla-luciferase

co-reporter (phRL-CMV) vector into

immortalized human mesangial cells

– Inversion of the A-allele genomic region

(REV-1.1kb) served as a negative control

• The G-allele (Mut1.1kb) reduced

transcriptional activity by 55.6% by

comparison to the A-allele (p < 0.00005)

– In a separate experiment, double site-

directed mutagenesis of the GATA site

also decreased transcriptional activity

KHv0815 Ho K…Orchard TJ. J Am Soc Nephrol 2005;16;589A

ABCC9 in Type 1 Diabetic Kidney Disease

Conclusions

• Combining genotypes for 2 ABCC9 gene promoter SNPs, rs17335932

(173*) and rs35677639 (356*), predict incident albuminuria in T1DM

subjects free of albuminuria at baseline after 22 years of followup

• In Cox proportional hazard models, combining high risk, 173* AA and 356*

GG, genotypes (Group 3) conferred greatest risk, while combining high risk,

173* AA, and low risk, 356* AA/AG, genotypes (Group 2) conferred

intermediate risk. Group 3 further predicted renal function decline

• Using immortalized human mesangial cells transfected with reporter

constructs incorporating the human ABCC9 promoter region, we

demonstrated that the rs17335932 G-allele, associated with low risk of

incident albuminuria, decreased gene transcription compared to the A-allele

(potentially by disrupting the GATA motif of a putative LIM-only protein

transcription binding site)

KHv0815

Major Modifiable Mortality Risk Factor

in Chronic Kidney Disease: Vascular Access

FistulaFirst: FFBI Vascular Access flyer

Conti CR. Acquired Heart Disease.

In: Netter collection of Medical Illustrations. 2nd ed.

Philadelphia, PA: Elsevier/Saunders; 2014

KHv0815

Tunneled DialysisCatheter

Risk of Death is Highest During 1st 120 Days

• One-half of deaths (46%) occur within the 1st 120 days of starting HD

• Dialysis catheters account for largest attributable fraction ofmortality risk during Year 1 (≤120 d and >120 d) among modifiable risk factors for ESRD patients on hemodialysis

• Primary cause of death = Cardiac death

Bradbury BD et al. Clin J Am Soc Nephrol 2007; 2: 89-99KHv0815

Dialysis Outcomes & Practice Patterns Study (DOPPS), 1996-2004

n = 4,802 incident HD patients (U.S.)

27.5 (≤120 d) vs 21.9 deaths (>120 d) /100 pers-yrs

CARD INF

120 d

Death

Rate

per

10

0 p

ers

on

-yrs

0 dDays at Risk

330 d

Causes of Death in Incident Dialysis Patients

USRDS Annual Data Report 2012KHv0815

2008-2010

Rates of sudden cardiac

death following dialysis

initiation in incident

dialysis patients

*

Tunneled Dialysis Catheters (TDCs)

• TDCs, prevalence = 20% (2012), are associated with increased

complications compared to AV fistulas: infection, inadequate

clearance, thrombosis, hospitalization

• Most significant: TDCs are associated with higher mortality in

both incident & prevalent hemodialysis patients

• Over 80% of incident HD patients use a TDC at initiation of

HD in U.S.

• Only 16% of incident HD patients use an AVF for 1st

outpatient HD treatment in the U.S.

– Without prior nephrology care, 89% of incident HD patients

start dialysis with TDCs (overall, 42% patients without pre-

ESRD care)

– With prior nephrology care (>12 mo), fewer patients (50%)

start with a TDC

Fistula First Data

http://www.fistulafirst.org

USRDS Annual Data Report 2013KHv0815

Ethier J et al. Nephrol Dial Transplant 2008;23:3219-3226KHv0815

Vascular Access Use Worldwide

Incident Patients Prevalent Patients

ANZ, Australia-New Zealand; BE,

Belgium; CA, Canada; FR, France;

GE, Germany; IT, Italy; Jpn, Japan;

SP, Spain; SW, Sweden; UK, United

Kingdom; US, United States

Dialysis Outcomes and Practice

Patterns Study (DOPPS) I, II, III

1996 to 2007

GEGE

US

US

VA Type Affects Mortality in HD Patients

• In 79,545 prevalent HD

patients (mean vintage 3.6 yrs)

o Compared with AVFs,

mortality risk is 34% greater

with TDCs

• Effect is more pronounced in the 4,741 incident HD patients (vintage <90 d)

o Compared with AVFs,mortality risk is 119% greater with TDCs

Lacson Jr E et al. AJKD 2009;54:912-921

Prevalent

Incident

AVF

TDC

TDC

AVG

AVG AVF

HR

fo

r M

ort

ali

tyH

R f

or

Mo

rta

lity

KHv0815

Conversion of Vascular Access Reduces Mortality

Conversion of vascular access type during a prior 4-month period changed mortality risk during following 8 months

• In 70,852 prevalent HD patients

o Converting TDC to AVF/AVG:29% decrease in risk

• In 3,904 incident HD patients

o Conversion of TDC to AVF: 50% decrease in risk

o TDC to either AVF or AVG:15% decrease in risk

Lacson Jr E et al. AJKD 2009;54:912-921

Prevalent

Incident

HR

fo

r M

ort

ali

tyH

R f

or

Mo

rta

lity

TDCAVG AVFAVF/G

TDCAVG AVFAVF/G

KHv0815

Dialysis Catheter to Permanent Vascular Access in Incident HD Inpatients who Transition to Outpatient Hemodialysis

Interesting observation

• At UPMC-Presbyterian campus: 175 consecutive incident

hemodialysis patients (inpatient setting/UPMC-

Presbyterian Hospital or outpatient setting/nephrology

faculty clinic practice) were referred to outpatient dialysis

units over 9 months (4/08-12/08)

• Of these, 85% initiated hemodialysis acutely as

inpatients (149 of 175)

Why is the observation important?

• National vascular access guidelines (Fistula First/NKF-

KDOQI 2006) address permanent vascular access planning

in only CKD Stage 4 (eGFR 15 to 29 ml/min/1.73m2)

transitioning to ESRD (e.g. outpatients) initiating

hemodialysis

• One explanation for low observed rate of pre-ESRD

nephrology care in patients starting dialysis

Andersen S…Ho K. Amer J Kidney Dis 2011;57(4):A21,P-19KHv0815

Inpatient Hemodialysis Initiation826 consecutive adult inpatients* initiated on HD over 26 months; 466 dialysis-dependent at hospital discharge.

[*includes pre-existing ESRD (renal transplant, peritoneal dialysis) not on HD during preceding 12 mo.]

Andersen S…Ho K. Amer J Kidney Dis 2011; 57(4):A21,P-19

Subset of 84 incident HD inpatientsdischarged on HD to 8 dialysis units

Inpatient Subset (n=84)

Incident HD Inpatients n (%)

Non-ESRD 54 (64%)

ESRD:Peritoneal Dialysis

13 (16%)

ESRD:Prior Renal Transplant

17 (20%)

26 Months

(4/08 – 6/10)

Initiate

Inpatient HD

n=826

Inpatient

Death

n=214 (25.9%)

Discharge

To Outpatient Dialysis

n=466 (56.4%)

Inpatient

Renal Recovery

n=146 (17.7%)

Outpatient HD

Other Dialysis Units

n=375

Outpatient HD

Affiliated DCI Units

n=84

Outpatient PD

n=4

Lost to followup

n=3

KHv0815

Dialysis Catheter to Permanent Vascular Access in Incident HD Inpatients who Transition to Outpatient Hemodialysis

Inpatient

Pre-Discharge

Outpatient

Post-Discharge

Initial Vascular AccessUsed for Inpatient HD

90.5% TDC 9.5% AVF

87.8% Ptsmean 92.4 dmedian 67.5 d

DISCHARGE: Outpatient Units

UltrasoundVein Mapping

AV Fistula (AV Graft)Surgery/Creation

AV Fistula (AV Graft)Surgery/Creation

UltrasoundVein Mapping

826 InpatientsInitiate HD

459 InpatientsDischarged on HD

Subset ofIncident HDInpatients(n=84)

12.2% Ptsmean 5 dmedian 7 d

11.6% Ptsmean 2.8 dmedian 2 d

88.4% Ptsmean 112 dmedian 105.4 d

Inpatient

Hemodialysis Initiation

13.1% PtsPre-existingAVFs

54% PtsInitiating HD with TDCReceived AVF/AVG

Andersen S…Ho K. Amer J Kidney Dis 2011;57:A21,P-19

• Functional measure of outpatient TDC use,

Defined: HDTDC = [(No. outpatient HD treatments using TDC) / (No. of total

outpatient HD treatments)] during time interval (HDTDC is inverse to AVF/G use)

• EARLY AVF/G group exhibited lower TDC use with mean HDTDC values of 57%

and 33% for months 3 to 6 months and months 6 to 12, respectively, in comparison

to 91% and 74% for the LATE AVF/G group

Andersen S…Ho K. J Am Soc Nephrol 2011; 22: 787A, P-2910

Early AVF placement

≤ 90 days (n=29)

Late AVF placement

> 90 days (n=38)

p=0.0004

p=0.0004

Does Earlier AVF Placement Translate to Earlier AVF Use

& Fewer TDC-Associated Treatments in Year 1?

Modifying the Vascular Access Conversion Curve

Hypothesis

Predicting incident ESRD followed by inpatient Early Vascular

Access planning will reduce mortality & readmission in Year 1

TDC to Permanent Vascular Access Conversion

% P

ati

en

ts w

ith

AV

F (

AV

G)

Time (months)

Dis

ch

arg

e

EVA Usual Mo

rtality

& R

ead

mis

sio

n R

ate

KHv0815

Low Renal Recovery in Inpatients Starting Hemodialysis Transitioning to Outpatient Hemodialysis

54 Inpatients Known CKD Status in Prior 1 Yr(no ESRD/CKD-T) Initiated on Hemodialysis

Subset ofIncident HDInpatientsn=84

Incident HDInpatientsDischargedon HDn=459(4/08-6/10)

The rationale for Early Vascular Access planning initiative in inpatients

Andersen S…Ho K. J Am Soc Nephrol 2011;22:788A,P-2914

Pre-Admission CKD Stage

(GFR ml/min/1.73m2)

Total Patients

Prior Renal Care

Renal Recovery at 3 Months

Post-Initiation

n n (%) n (%)

No CKD 6 n/a 2 (33)

Stage 1 (90) 1 0 (0) 0 (0)

Stage 2 (60 to <90) 0 0 (0) 0 (0)

Stage 3A (30 to <45) 7 2 (29) 0 (0)

Stage 3B (45 to <60 14 6 (43) 0 (0)

Stage 4 (15 to <30) 19 15 (79) 0 (0)

Stage 5 (<15) 7 6 (86) 0 (0)

Total CKD Stages 1-5 48 29 (60) 0 (0)

Pre-Admission CKD Stage: Determined within 1 Yr prior to admission - nephrology documentation > other MD documentation > average of 3 baseline Cr values. Prior Renal Care: Nephrology followup (including renal transplantation followup for CKD-T) Renal Recovery at 3 Months: Discontinuation of hemodialysis as a result of improved renal function within 90 days of initiating hemodialysis as inpatient

39% CKD 3A+3B

35% CKD 4

13% CKD 5

11% No CKD

KHv0815

Inpatient AKI-on-CKD Leads to ESRD

• Multicenter, observational study of 9,425 Taiwanese post-surgical inpatients admitted to surgical ICU and surviving to hospital discharge

Baseline eGFR <45ml/min/1.73m2 = CKD Stages 3B+4+5

• High risk of ESRD in AKI-on-CKD vs AKI-without-CKD (AHR = 19.8) F

reed

om

fro

m D

ialy

sis

+ AKI / + CKD

– AKI / – CKD+ AKI R / – CKD+ AKI I / – CKD+ AKI F / – CKD

– AKI / + CKD

CKD status AKI status Long-term dialysis, HR (95% CI)

No prior CKD No AKI 1 (reference)

+ AKI 4.64 (2.51-8.56)*

+ Prior CKD No AKI 40.86 (20.01-83.50)*

+ AKI 91.6 (49.3-170.1)*

KHv0815 Wu V-C et al. Kidney Int 2011;80:1222-1230

Early Vascular Access (EVA) Planning Initiative

Dialysis Provider

Dialysis Unit start dateForm 2728 data

HD: incenter, SNF, homeethnicity, racemedical insurance typeemployment statusprior erythropoietin use

+

Vasc Access Data

Initial vasc access type1st AVF (AVG) use dateTDC treatment daysAVF (AVG) treatment daysRenal recovery, death

Dialysis Provider

Hospitalization Datahospital days / mo

admission diagnosis

The Renal Network& Dialysis Provider

Mortality Datadate of death

deaths/patient days at risk

Post-Discharge1-Y Outpatient Data

Data Sharing

Agreements

Dialysis Data

Renal replacement initiation date

1st acute hemodialysis date9999903 billing code1st CVVHD date

Renal RecoveryDeath, CMO

HD Vasc Access Data

TDC (1st) insertion date(existing AVF/AVG?)

Renal RN education dateVein mapping dateVascular Surg Consult date

AVF (AVG) surgery dateConversion time TDC to VATDC (last) removal date

TriggersData

Collection

EVA Mechanism

EMR EVA

ActivationRenal RNEducator

Vascular SurgRN Coordinator

VeinMapping

AVF (AVG)Placement

Cerner PatientTracking List

VascularAccessDatabase

Discharge &Admission

to OutpatientDialysis Unit

KHv0815

Effect of Inpatient Early Vascular Access Planning on Outpatient Hemodialysis

Vascular Access Outcomes, Hospitalization, Mortality

Evaluate

• Pre-implementation status of inpatient vascular access planning• Post-discharge outpatient effectiveness of QI mechanism

• Primary Measures

o EVA planning event occurrence (vein mapping, Vascular Surgery consultation, vascular surgery, followup visit)

o Conversion time from initial inpatient TDC placement to (a) initial AVF/AVG placement (inpatient or outpatient) and (b) initial AVF/AVG use (two-needle) in outpatient dialysis center setting

o Hospitalization & mortality rates of incident HD inpatient patients during the first 6 months and 12 months post-index hospital discharge

- date of initial TDC placement- date of vein mapping procedure- date of Vascular Surgery inpatient / outpatient consultation- date of Vascular Surgery outpatient follow-up appointment- date of initial AVF/AVG placement (inpatient or outpatient)- date of conversion from TDC use to AVF/AVG use for outpatient HD treatments- duration of pre-ESRD nephrology care prior to hemodialysis initiation- occurrence of pre-dialysis erythropoiesis stimulating agent (ESA) therapy- total hospital days post-index hospitalization within 1 year- mortality events post-index hospitalization within 1 year

KHv0815

AKI to ESRD Predictive Analytics Data Capture

Dialysis Provider

Dialysis Unit start dateForm 2728 data

HD: incenter, SNF, homeethnicity, racemedical insurance typeemployment statusprior erythropoietin use

+

Vasc Access Data

Initial vasc access type1st AVF (AVG) use dateTDC treatment daysAVF (AVG) treatment daysRenal recovery, death

Dialysis Provider

Hospitalization Datahospital days / mo

admission diagnosis

ESRD Renal Network& Dialysis Provider

Mortality Datadate of death

deaths/patient days at risk

Post-Discharge1-Y Outpatient Data

Discharge &Admission

to OutpatientDialysis Unit

TriggersData

Collection

Data Sharing

Agreements

Goal: Real-Time Data Capture

KHv0815

AKI to ESRD Predictive Analytics

KHv0815

Outpatient

ESRD

Outpatient AKI Evolution

DEVELOP

Acute

IHD2nd HIT1st HIT

NoRecovery

3RD HIT

Risk ofAKI

Risk ofAKI

Risk ofESRD

Inpatient

AdmissionHospitalization

Pre-ExistingRisk Factor(s)

AmplifyingFactor(s)

AcutePrecipitatingFactor(s)Age

AlbuminuriaProteinuriaChronic Kidney DiseaseESLDHyperuricemiaGenomics

Acute MICHFSepsisHemodynamicsCardiac SurgeryIV ContrastMedications

AKI Severity

AKI Duration

AKI Recurrence

Predicting ESRD & Early Vascular Access Planning

Conclusions

• Already known: (1) Tunneled dialysis catheters (TDCs) account for

largest attributable fraction of mortality risk during Year 1 in incident

hemodialysis patients when risk is greatest, (2) conversion of TDC to

permanent vascular access (preferably AVF) reduces mortality risk

• We/others observed that the low rate of renal recovery following acute

initiation of hemodialysis for AKI on CKD in inpatients (discharged on

hemodialysis) represents a major pathway for incident ESRD

• Data suggests that occurrence of incident ESRD in this inpatient

population is likely to be highly predictable

• In identified inpatients at risk for incident ESRD, early vascular access

planning initiated in the inpatient setting may reduce healthcare costs,

hospital readmission, mortality by significantly reducing vascular access

conversion time

KHv0815

http://images.3d4medical.com/blood-in-circulation-51

Anemia Management in ESRD

Personalized vs. Standardized Computer-

Assisted Erythropoietin Dosing

Anemia in ESRD: Expensive Economic Issue

• Erthropoiesis stimulating agents (ESAs, e.g. epoetin alfa, darbepoetin alfa) are expensive

– largest modifiable economic factor determining HD treatment cost

– $1.9 billion of total $28 billion annual Medicare cost for ESRD (2010)

• Reducing use of ESAs -- a priority for CMS (Centers for Medicare & Medicaid Services) -- has driven health economic policy and quality improvement in Nephrology

– 2011: “Bundling" of prior separately reimbursable injectable drugs (targeting ESAs) into the total hemodialysis treatment cost shifted the cost of dialysis from payer to dialysis provider (Medicare Improvement for Patients and Providers Act of 2008/MIPPA)

– 2012: First CMS pay-for-performance Quality Incentive Program (QIP): focused on anemia management in 2 of the 3 metrics (Jan. 2012)

KHv0815 USRDS Annual Data Report 2012

CHOIR Study in Chronic Kidney Disease

Open-label, RCT in CKD Stages 3B/4 patients, Mdn duration=16 mo.

• n=1432, eGFR=15-50ml/min

• Randomized to: High-Hb Grp, target/maintain Hb13.5 g/dl (actual mean Hb12.6 g/dl) (715 pts) vs. Low-Hb Grp, target/maintain Hb11.3 g/dl (717 pts)

• Primary end point: composite of death, MI, CHF hospitalization, or CVA

• Study was halted early:125 events in High-Hb Grp vs.97 events in Low-Hb Grp, HR=1.34 (1.03-1.74; p=0.03)

Singh AJ et al. N Engl J Med 355:2085-2098, 2006

Month

Pro

ba

bil

ity o

f

Co

mp

os

ite

Eve

nt

High-Hb

Low-Hb

KHv0815

Optimal Hemoglobin Target in ESRD?

The optimal Hb target remains controversial

• In contrast, large observational studies of ESRD patients on HD suggest better survival with Hb = 10.0 to 12.0(-13.0) g/dl range but increased mortality for Hb <10.0 g/dl

– 21,899 pts (Madore F et al. J Am Soc Nephrol 8:1921-1929,1997)

– 5,302 pts (Locatelli F et al. Nephrol Dial Transplant 13:1642-1644, 1998)

– 95,273 pts (Ma JZ et al. J Am Soc Nephrol 10:610-619,1999)

– 58,058 pts (Regidor DL et al. J Am Soc Nephrol 17:1181-1191, 2006)

• Recent evidence suggests ESA hypo-responsiveness (comorbidities, acute illness, demographics) and not simply Hb level or ESA dose is associated with increased mortality, poor patient outcomes

KHv0815

Current Non-Agreement in Hemoglobin Targets

K-DOQI : Kidney Disease Outcomes Quality Initiative

K-DIGO : Kidney Disease: Improving Global Outcomes

CKD StageHb to

Start ESA

Hemoglobin Target

Range

Max Hb

Threshold

K-DOQIStages 3 to 5

11.0 - 12.0 > 13.0Stage 5+Dialysis

K-DIGOStages 3 to 5

< 10.010.0 - 11.5

> 13.0Stage 5+Dialysis 9.0 - 11.5

FDAStages 3 to 5 (None) - 10.0 (> 10.0)

Stage 5+Dialysis (None) - 11.0 (> 11.0)

CMS(Medicare & Medicaid)

Stages 3 to 5 < 10.0 10.0 - 12.0(> 12.0)

Stage 5+Dialysis (None) - 12.0

KHv0815

Personalized Pharmacodynamic Dosing Algorithm?

• General hypothesis

A pharmacodynamic algorithm to dose Epo (epoetin alfa) in a patient-specific

manner will increase the % patients on hemodialysis achieving a defined

hemoglobin (Hb) target while reducing overall Epo utilization

• Aim

Create a model able to predict a Hb value in individual patients for a given dose

of Epo based on each subject’s historical EPO-Hb dose-response

• Intelligent Dosing SystemTM (IDS)

FDA-approved algorithm (class II device) which models drug dosing based on

individual patient drug-responsiveness through an iterative process (stochastic

loop function)


DCI database of 11,803 incenter hemodialysis (HD) patients with complete

demographics, lab data, EPO doses (1.4 million thrice-weekly iv) in 1/05-11/06

• Retrospective analysis

Assess the difference between a predicted Hb level 2 (IDS model without

iteration) and observed Hb level no. 2 by using 3 parameters (starting EPO

dose no. 1, starting Hb value no. 1, EPO dose no. 2) in patients

KHv0815 Mcmichael J…Ho K. J Am Soc Nephrol 2010; 22: 483A, P-1107

Modeling Individual Hemoglobin Responses to Epo

• Relationship between Epo dose and Hb values was non-linear

• Predicted Hb no. 2 levels were predicted within 0.3 g/dL, 0.5 g/dL, and 1.0 g/dL of the observed Hb no. 2 level in 34%, 50%, and 75% of cases

• Thus, IDS model plausibly beneficial in improving Epo dosing

Starting Epo Dose (1st), New Epo Dose (2nd),

% Change in Hb vs. IDS-Based Model

KHv0815 Mcmichael J…Ho K. J Am Soc Nephrol 2010; 22: 483A, P-1107

Personalized vs. StandardizedComputer-Assisted Epo Dosing RCT

• Aim

Compare 2 computer-assisted algorithms in dosing iv Epo 3x-weekly to

achieve Hb=11.0 g/dl with a target Hb=10.0-12.0 g/dl in hemodialysis patients

using a double-blinded randomized controlled trial:

(1) Personalized (individual-specific) approach based on the IDS algorithm(2) Standardized (fixed-rule ‘one-size-fits-all’) approach based on the DCI-CDS

algorithm (DCI Corporate)


Incenter HD patients (n=88) at primary dialysis unit for Univ of Pittsburgh

• Study design: computer-assisted Epo dosing

– 48 incenter HD patients were block-randomized to IDS or DCI-CDS

computer-assisted algorithm interventions for 6 mo. (post-iv Venofer

protocol x 2 mo.)

– Both algorithms were administered blinded thru a central DCI server

– Biweekly Hb values triggered automatic calculation of Epo doses by

algorithms (monitored by Anemia Manager); Epo held for Hb>12.5 g/dl

– Epo orders were securely transmitted by MIS to unit and implemented

KHv0815 Ho K…Miskulin D. J Am Soc Nephrol 2011; 22: 483:, P-1595

Personalized vs. StandardizedComputer-Assisted Epo Dosing RCT

• Outcomes

Target Hb=10.0-12.0 g/dl

– All Hb values categorized as: 10.0-12.0 g/dl (at target),

<10.0 g/dl, >12.0g/dl

Average EPO dose (IUnits)/Treatment

Erythropoetin Resistance Index (ERI)/Treatment

– ERI = Epo dose (IU) / weight (kg) / Hb (g/dl)

• Statistics

Categorical variables – chi-square; continuous variables –

Kruskal Wallis

Pre-intervention (run-in=6 mo.) and intervention (6 mo.)

measurements (repeated measures in same subjects):

Discrete outcomes – mixed logistic regression

Continuous outcomes – mixed linear models


IDS vs. DCI-CDS Baseline Characteristics

Ho K…Miskulin D. J Am Soc Nephrol 2011; 22: 483:, P-1595KHv0815

No significant baseline

differences overall

with exception in IDS group:

Lower no. missed treatments

Lower mean ferritin level

Personalized vs. Standardized

Computer-Assisted Epo Dosing RCTPrior to intervention, the IDS patient group exhibited:

• Lower % of patients within the Hb target range

• Higher Hb variability, Epo requirement/treatment, and Epo Resistance Index

During intervention, % patients (both groups) achieving Hb target increased 52.1% to 61.9%

Ho K…Miskulin D. J Am Soc Nephrol 2011; 22: 483:, P-1595KHv0815

Target Hb = 10.0 to 12.0 g/dl


Computer-Assisted Epo Dosing RCT


IDS vs. DCI-CDS Computer-Assisted EPO Dosing

% Patients achieving

hemoglobin target

• On average over 6 mo.:

59.4% of IDS vs.

54.6% of DCI-CDS

• IDS patients more likely

to achieve target Hb

than DCI-CDS patients (OR=2.1, p=0.043)

• IDS patients were less

likely to overshoot,

Hb>12.0 g/dl (OR=0.35,

p=0.005) -- CMS QIP

(quality incentive

program) metric


Computer-Assisted Epo Dosing RCT

Conclusion

• The IDS DCI-CDS trial provides evidence that a computer-assisted

erthropoiesis stimulating agent (Epo) dosing program may operate

autonomously to increase the % patients achieving a defined Hb target

range (whatever that target may be)

• Such a personalized, patient-specific computer-assisted algorithm may not

only improve anemia outcomes, but also reduce Epo utilization and reduce

healthcare costs at the dialysis unit provider level

KHv0815

Thank You

…for your kind attention &

for the honor of both meeting

and speaking with you here

at Geisinger…

geisinger research talk incorporating addendum 090215

Documents