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
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
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
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
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
Baseline Characteristics
Combined genotype groups
Subjects in combined genotype groups exhibited no differences in baseline
characteristics
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
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)
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
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)
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
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)
KHv0815 Ho K, Miller RG, Orchard TJ. J Am Soc Nephrol 2014;25:692A,P-246
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)
• Study population
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)
• Study population
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
KHv0815 Ho K…Miskulin D. J Am Soc Nephrol 2011; 22: 483:, P-1595
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
Personalized vs. Standardized
Computer-Assisted Epo Dosing RCT
KHv0815 Ho K…Miskulin D. J Am Soc Nephrol 2011; 22: 483:, P-1595
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
Personalized vs. Standardized
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…