Embed Size (px)
Citation preview
TCmb
AD
ULT
CA
RD
IAC
Phosphodiesterase-5 Inhibition PreventsPostcardiopulmonary Bypass Acute KidneyInjury in SwineNishith N. Patel, MRCS, Hua Lin, MS, Tibor Toth, MD, MRCPath, Ceri Jones, BS,Paramita Ray, FRCA, Gavin I. Welsh, PhD, Simon C. Satchell, PhD, MRCP,Philippa Sleeman, BS, Gianni D. Angelini, MD, FRCS, and Gavin J. Murphy, MD, FRCSBristol Heart Institute, University of Bristol, Bristol Royal Infirmary, and Department of Histopathology, North Bristol NHS Trust,
Southmead Hospital, Bristol; Department of Anaesthesia and Critical Care, Weston General Hospital, Weston-Super-Mare; andAcademic Renal Unit, University of Bristol, Southmead Hospital, Bristol, United Kingdombd2tlosoppda
bjac
Background. Acute kidney injury after cardiac surgeryis common, has no effective treatments, and is associatedwith adverse outcomes. The aim of this study was todetermine whether administration of the phosphodies-terase-5 inhibitor sildenafil citrate (SDF) would preventthe development of post–cardiopulmonary bypass (CPB)acute kidney injury in swine.
Methods. Adult pigs (n � 8 per group) were random-ized to undergo sham procedure, CPB, or CPB plusadministration of SDF, with recovery and reassessmentat 24 hours.
Results. Cardiopulmonary bypass resulted in a signif-icant reduction in creatinine clearance relative to shampigs (mean difference CPB versus sham, �47.9 mL/min;95% confidence interval [CI]: �93.7 to �2.2; p � 0.039).
his was prevented by the administration of SDF duringPB (mean difference CPB�SDF versus CPB, �55.6 mL/in; 95% CI: �6.5 to �104.7; p � 0.024). Cardiopulmonary
ypass also resulted in a significant rise in the urinary
Institute, Bristol Royal Infirmary, Bristol BS2 8HW, UK; e-mail:[email protected].
© 2011 by The Society of Thoracic SurgeonsPublished by Elsevier Inc
iomarker interleukin-18 compared with sham proce-ures (mean difference 209.3 pg/mL; 95% CI: 120.6 to98.1; p < 0.001) that was prevented by SDF administra-ion. Post-CPB kidney injury was associated with vascu-ar endothelial injury and dysfunction, reduced nitricxide bioavailability, medullary hypoxia, cortical adeno-ine triphosphate depletion, inflammation, and evidencef proximal tubule epithelial cell stress manifest ashenotypic change. Administration of SDF to CPB pigsreserved nitric oxide bioavailability and prevented en-othelial dysfunction, regional hypoxia, inflammation,nd tubular changes.Conclusions. In this model, phosphodiesterase-5 inhi-
ition using SDF prevented post-CPB acute kidney in-ury by the preservation of nitric oxide bioavailability,nd warrants evaluation as a renoprotective agent inlinical trials.
(Ann Thorac Surg 2011;92:2168–76)
© 2011 by The Society of Thoracic SurgeonsAcute kidney injury after cardiac surgery is common,affecting as many as 40% of patients [1] and is
associated with a fourfold increased risk of in-hospitaldeath [2, 3], a doubling of healthcare costs [4], andprolonged hospital stay [4]. Despite the devastating ef-fects of acute kidney injury in cardiac surgical patients,our understanding of the underlying processes is poor,and there are no effective renoprotective strategies [5].
We have recently described a novel porcine recoverymodel of post–cardiopulmonary bypass (CPB) acute kid-ney injury. That has significant quantitative and qualita-tive homology to acute kidney injury in cardiac surgicalpatients [6]. In this model, nitric oxide (NO) bioavailabil-ity is an important determinant of functional, biochemi-cal, and histologic renal outcomes after CPB; and strate-gies, such as endothelin-A receptor blockade, thatpreserve NO bioavailability are renoprotective [7]. En-
Accepted for publication July 11, 2011.
Address correspondence to Mr Murphy, Cardiac Surgery, Bristol Heart
dogenous NO activity can be augmented by phosphodi-esterase-5 (PDE-5) inhibitors, which prevent the break-down of its secondary messenger, cyclic 3’-5’-guanosinemonophosphate [8]. We hypothesized that augmentationof endogenous NO pathways with the PDE-5 inhibitorsildenafil citrate (SDF) would prevent post-CPB acutekidney injury in the porcine model. Acute kidney injuryis characterized clinically by an acute decline in glomer-ular filtration rate [9]. We used changes in creatinineclearance, an index of the glomerular filtration rate thatwe have used in a previous clinical study [10], as ourprimary endpoint, with endothelial function, NO bio-availability, renal oxygenation, and histologic evidence ofepithelial and endothelial injury and inflammation assecondary endpoints.
Dr Murphy discloses that he has financial relation-ships with NovoNordisk, Ethicon, Maquet, andMedtronic.
0003-4975/$36.00doi:10.1016/j.athoracsur.2011.07.002
nppd
2169Ann Thorac Surg PATEL ET AL2011;92:2168–76 SILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI
AD
ULT
CA
RD
IAC
Material and Methods
Twenty-seven female Large-White-Landrace crossbredpigs approximately 4 months old and weighing 50 to 70kg were used. Animals received care in accordance with,and under license of, the Animals (Scientific Procedures)Act 1986 and conformed to the “Guide for the Care andUse of Laboratory Animals” published by the NationalInstitutes of Health (NIH Publication No. 85-23, revised1996). The study had received local Institutional ReviewBoard approval. A schematic of our experimental designis shown in Figure 1. All experimental methods havebeen described in detail previously [7].
InterventionWe allocated 24 animals (n � 8 per group) to thefollowing groups: group 1, sham operation—neck dissec-tion with 2.5 hours of general anesthesia; group 2, CPB—2.5 hours of CPB; or group 3, CPB plus SDF—2.5 hours ofCPB plus the selective PDE-5 inhibitor SDF (Revatio;Pfizer, Sandwich, UK), 10 mg in 50 mL 0.9% saline infusedover 30 minutes at the commencement of CPB. This dosewas identified in a preliminary study (n � 3 pigs) as havingminimal hypotensive side effects. The timing of SDF admin-istration was chosen to reflect the likely clinical scenariowere the agent to be administered in a future clinical trial.
Anesthesia and CPBAnesthesia and CBP were performed by a modification ofour protocol described previously [6, 7]. Briefly, CPB was
Abbreviations and Acronyms
CI � confidence intervalCPB � cardiopulmonary bypassDBA � dolichos biflorus agglutinineNOS � endothelial nitric oxide synthaseIL � interleukiniNOS � inducible nitric oxide synthaseNO � nitric oxidePDE-5 � phosphodiesterase-5SDF � sildenafil citrate
established between minimally invasive Smart Cannu-lae (Smartcanula LLC, Lausanne, Switzerland) placedin the aorta and right atrium through the right internalcarotid artery and external jugular vein, respectively.Postbypass animals were recovered, reanesthetized,and reevaluated after 24 hours. Nephrectomy wasperformed at 24 hours before euthanasia.
Biochemical Markers of Renal InjuryCreatinine clearance, free water clearance, and fractionalsodium excretion were calculated from serum samplesand urine samples taken over three periods, namely, 90minutes before CPB; 90 minutes after weaning from CPB,and 90 minutes at 24 hours after CPB, as previouslydescribed [6], using accepted formulas [11]. Urinaryprotein and the albumin-to-creatinine ratio were de-termined by immunoturbidimetry on the Cobas Miraanalyzer (Roche Diagnositcs Limited, West Sussex,UK). Interleukin-18, a specific marker of acute kidneyinjury detected in the urine more than 12 hours afterrenal injury [12], was measured in urine samples usingenzyme-linked immunosorbent assay (Bender Med-Systems, Vienna, Austria).
Endothelial Function and Renal HypoxiaAt 24 hours after CPB, renal artery blood flow (T106Transonic flow meter; Transonics Systems, Ithaca, NY)was measured. Endothelial function was determined bythe change in renal blood flow in response to acetylcho-line (0.1 to 10 �g · kg�1 · min�1). Urinary nitrate:nitriteconcentration was used as a measure of renal NObioavailability. Nitrate:nitrite concentration was mea-sured using a colorimetric NO assay kit (Calbiochem;Merck, Nottingham, England) that is based on theGriess reaction [13] and expressed as NO concentra-tion, as described previously [14]. Outer medullaryoxygenation at 24 hours after CPB was measured bytissue O2 sensors (Oxylite pO2 E Series; Oxford Optro-
ix, Oxford, UK). Renal cortical adenosine triphos-hate levels were measured using reverse-phase high-erformance liquid chromatography, as previouslyescribed [15].
Fig 1. Schematic of study design. (CPB � car-diopulmonary bypass; ICC � immunocyto-chemistry; mins � minutes; postop � postop-erative; preop � preoperative; SDF �sildenafil citrate.)
2170 PATEL ET AL Ann Thorac SurgSILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI 2011;92:2168–76
AD
ULT
CA
RD
IAC
Renal TissueFormalin fixed, paraffin embedded, 5-�m transverse re-nal sections stained with hematoxylin and eosin werescored for renal tubular injury and inflammation by anexperienced renal pathologist (T.T.) blinded to the exper-imental conditions, as described previously [6]. Immuno-cytochemistry and immunofluorescence was performedas previously described [16-18] for endothelial nitric oxidesynthase (eNOS [Santa Cruz Biotechnology, Santa Cruz,CA]), dolichos biflorus agglutinin (DBA) lectin, induciblenitric oxide synthase (iNOS [Thermo Fisher Scientific UK,Loughborough, UK]), endothelin-1 (Acris Antibodies, Her-ford, Germany), and inflammatory cells (MAC 387; Abcam,Cambridge, UK). Qualitative differences in staining wereconfirmed by Western blotting [16].
Power of the Study and Statistical AnalysisCreatinine clearance was our primary endpoint. Wecalculated that a study with 24 animals (8 per group)would have a 90% power to detect a large effect size of 0.7,equivalent to a difference of 16.5 mL/min in creatinineclearance between groups assuming a within-group stan-
40
50
60
70
80
90
100
110
0 30
BPCerP
Mean A
rteria
l Pre
sure
(m
mH
g)
-20
-15
-10
-5
0
5
10
15
0 30
BPCerP
CV
P (cm
H2O
)
A
Fig 2. (A) Mean arterial bloodpressure and (B) central venouspressure (CVP) in the sham group(black diamonds), the cardiopulmo-nary bypass (CPB) group (graysquares), and the CPB plus silde-nafil (SDF) group (blue triangles)over the course of the experiments.Data represent mean (SD).
B
dard deviation of 23.5, with one baseline and two postin-tervention measures per animal, and an assumed corre-lation between before and after measures and betweenafter measures of 0.9 (estimated from pilot data (6)) and a5% Bonferroni corrected statistical significance. Differ-ences between groups were calculated using analysis ofvariance with adjustment for baseline differences forrepeated measures. For normally distributed data, valuesare expressed throughout as mean (� SE), and treatmentdifferences are reported as mean difference (95% confi-dence intervals [CI]). Nonnormally distributed data areexpressed as geometric means, with treatment differ-ences expressed as ratios of geometric means (95% CI).All p values less than 0.05 were considered to be statis-tically significant. All analyses were carried out usingSPSS 14.0 (SPSS, Chicago, IL).
ResultsAnesthesia, Monitoring, and CPBAll animals survived the study to recovery, reanesthesia,reevaluation, and sacrifice. Baseline weight, serum creat-
30 60 90 120 0 30 Day 1
1 yaDBPCtsoPBPC
Time Period
Sham CPB CPB+SDF
30 60 90 120 0 30 Day 1
1 yaDBPCtsoPBPC
Time Period
0
0
Sham CPB CPB+SDF
Table 1. Biochemical Markers of Acute Kidney Injury
Outcome
Mean (�SE) Difference (95% CI)ANOVA(p Value)Sham CPB CPB � SDF CPB Versus Sham CPB�SDF Versus Sham CPB�SDF Versus CPB
Creatinine clearance, mL/mina
Baseline 109.9 (8.0) 113.1 (13.0) 123.2 (12.5) 3.2 (�38.8 to 45.1) 13.3 (�28.7 to 55.2) 10.1 (�31.8 to 52.0) 0.696p � 1.00 p � 1.00 p � 1.00
1.5 hours 146.1 96.1 152.8 �50.0 6.7 56.724 hours 132.7 81.3 154.6 �51.4 21.9 73.3Pooled over time 144.9 (9.7) 91.9 (10.1) 153.6 (12.4) �53.0 (�90.5 to �15.5) 8.8 (�34.5 to 52.0) 61.7 (19.3 to 104.2) 0.001
p � 0.005 p � 1.00 p � 0.004Urinary P:C ratio, mg/mmolb
Baseline 23.15 23.83 22.56 1.03 (0.68 to 1.57) 0.97 (0.64 to 1.48) 0.95 (0.62 to 1.44) 0.945p � 1.00 p � 1.00 p � 1.00
1.5 hours 23.14 39.56 29.59 1.71 1.28 0.7524 hours 27.57 36.34 24.26 1.31 0.88 0.67Pooled over time 25.29 39.08 22.08 1.54 (1.16 to 2.05) 0.87 (0.64 to 1.19) 0.56 (0.41 to 0.78) � 0.001
p � 0.003 p � 0.757 p � 0.001Urine IL-18 at 24 hours, pg/mLa 27.54 (13.53) 238.32 (36.74) 0.02 (0.003) 210.78 (127.62 to 293.94) �27.52 (�110.68 to 55.64) �238.30 (�321.46 to �115.14) � 0.001
p � 0.001 p � 1.00 p � 0.001
a Least squares means, adjusted for baseline creatinine clearance estimated at 120.3. b Nonnormally distributed data expressed as geometric means, adjusted for baseline protein:creatinine (P:C) ratioestimated at 23.81, and difference between groups expressed as ratios of geometric means. The test for an interaction between treatment and time was p � 0.911 for creatinine clearance and p � 0.846 forprotein:creatinine ratio. Data from postintervention time points were therefore pooled to estimate the overall effect for these outcomes.
ANOVA � analysis of variance; CI � confidence interval (after Bonferroni adjustment); CPB � cardiopulmonary bypass; IL � interleukin; SDF � sildenafil; SE � standard error.
2171A
nn
Th
oracSu
rgP
AT
EL
ET
AL
2011;92:2168–76
SILD
EN
AFIL
PR
EV
EN
TS
PO
STC
AR
DIA
CSU
RG
ER
YA
KI
ADULT CARDIAC
(t
e(cp9p(9st0
stdstWoCe2
ble
2.En
doth
elia
lFu
nctio
nan
dR
enal
Hyp
oxia
tcom
e
Mea
n(�
SE)
Diff
eren
ce(9
5%C
I)
AN
OV
A(p
Val
ue)
Sham
CP
BC
PB
�SD
FC
PB
Ver
sus
Sham
CP
B�
SDF
Ver
sus
Sham
CP
B�
SDF
Ver
sus
CP
B
nal
bloo
dfl
ow,m
L·k
g�1
·min
�1a
asel
ine
4.15
(0.6
1)2.
17(0
.77)
10.6
4(2
.17)
�1.
98(�
6.52
to2.
57)
6.49
(1.9
4to
11.0
4)8.
47(3
.22
to13
.72)
0.00
2p
�0.
704
p�
0.00
7p
�0.
003
cety
lch
olin
e7.
22(0
.54)
1.67
(0.5
0)12
.08
(1.5
6)�
5.55
(�9.
01to
�2.
10)
4.86
(1.4
0to
8.31
)3.
27(0
.34
to6.
21)
�0.
001
p�
0.00
1p
�0.
99p
�0.
005
bioa
vaila
bilit
yat
24h
ours
,�m
ol/L
a19
3.0
(24.
2)53
.4(1
6.3)
275.
1(1
6.4)
�13
9.6
(�21
0.7
to�
68.5
)82
.1(1
1.0
to15
3.1)
221.
6(1
50.6
to29
2.7)
�0.
001
p�
0.00
1p
�0.
020
p�
0.00
1rt
ical
AT
P/A
DP
rati
oa0.
79(0
.04)
0.21
(0.0
3)0.
92(0
.22)
�0.
58(�
0.97
to�
0.19
)0.
13(�
0.28
to0.
54)
0.71
(0.2
8to
1.14
)0.
001
p�
0.00
4p
�1.
00p
�0.
002
du
llary
O2
ten
sion
at24
hou
rs,m
mH
ga87
.1(7
.2)
23.7
(3.1
)79
.4(0
.4)
–63.
3(�
84.8
to�
41.8
)�
7.6
(�32
.7to
17.4
)55
.7(2
9.8
to81
.6)
�0.
001
p�
0.00
1p
�1.
00p
�0.
001
ast
squ
ares
mea
ns,
adju
sted
for
base
line
ren
albl
ood
flow
esti
mat
edat
5.28
.
P�
aden
osin
ed
iph
osp
hat
e;A
NO
VA
�an
alys
isof
vari
ance
;A
TP
�ad
enos
ine
trip
hos
ph
ate;
CI
�co
nfi
den
cein
terv
al(a
fter
Bon
ferr
oni
adju
stm
ent)
;C
PB
�ca
rdio
pu
lmon
ary
ass;
NO
�n
itri
cox
ide;
SDF
�si
lden
afil;
SE�
stan
dar
der
ror.
2172 PATEL ET AL Ann Thorac SurgSILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI 2011;92:2168–76
AD
ULT
CA
RD
IAC
inine, urine output, and total volume of crystalloid andmetaraminol administered over the course of the studywere not different between the groups (data availablefrom authors on request). Mean arterial blood pressurewas significantly lower in CPB pigs versus sham (meandifference �14.5 mm Hg; 95% CI, �19.9 mm Hg to �9.2mm Hg; p � 0.001) and in CPB plus SDF pigs versus shammean difference �10.8 mm Hg; 95% CI, �17.7 mm Hgo �3.9 mm Hg; p � 0.002) during the intervention period
(Fig 2). There was no significant difference in perfusionpressure between the CPB and the CPB plus SDF groups(p � 0.498). In addition, mean arterial blood pressure wassimilar at baseline, immediately after intervention, and at24 hours for all three groups (Fig 2). There was nosignificant difference in pump flow rates between theCPB and the CPB plus SDF groups over the course of theexperiments (ratio of geometric means CPB�SDF versusCPB 1.01; 95% CI, 0.839 to 1.21; p � 0.929). There was nodifference in core body temperature (p � 0.276) or bloodhemoglobin concentrations (p � 0.232) between groupsduring the intervention period.
Biochemical Markers of Renal InjuryThe study was powered to detect significant differencesin creatinine clearance between the groups over two timeperiods after intervention. There was no interaction be-tween the time the postintervention measurements weretaken and treatment group (p � 0.911); therefore, pooled
stimates of the differences between groups are givenTable 1). Cardiopulmonary bypass resulted in a signifi-ant reduction in creatinine clearance relative to shamigs (mean difference CPB versus sham, �53.0 mL/min;5% CI, �90.5 to �15.5; p � 0.005; Table 1). This wasrevented by the administration of SDF during CPB,
mean difference CPB�SDF versus CPB, �61.7 mL/min;5% CI, �19.3 to �104.2; p � 0.004; Table 1). There was noignificant difference between groups in fractional excre-ion of sodium (p � 0.845) and free water clearance (p �.725; data not shown).The test for an interaction between treatment and time of
ample was p � 0.846 for urine protein:creatinine ratio;herefore, estimates of treatment effects were pooled. Car-iopulmonary bypass caused significant proteinuria, a non-pecific marker of kidney injury, relative to sham pigs, andhis was prevented by the administration of SDF (Table 1).
e also measured interleukin (IL)-18, a specific biomarkerf acute kidney injury. This was increased eightfold in thePB group compared with the sham group (mean differ-nce CPB versus sham 210.8 pg/mL; 95% CI, 127.6 pg/mL to93.9 pg/mL; p � 0.001) at 24 hours. Such a rise is diagnostic
of acute kidney injury [12], and this was prevented by theadministration of SDF (Table 1).
Endothelial Function and Renal OxygenationBaseline renal blood flow at 24 hours was similar be-tween sham and CPB pigs, but was significantly in-creased in CPB plus SDF pigs (mean differenceCPB�SDF versus CPB 8.47 mL · kg�1 · min�1; 95% CI,3.22 mL · kg�1 · min�1 to 13.72 mL · kg�1 · min�1; p �
0.001; Table 2). Administration of supraaortic acetylcho- Ta Ou
Re B A
NO
Co
Me
aL
e
AD
byp
2173Ann Thorac Surg PATEL ET AL2011;92:2168–76 SILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI
AD
ULT
CA
RD
IAC
line to CPB pigs caused a paradoxic reduction in renalartery blood flow indicative of endothelial permeabilityand dysfunction. That was prevented with the adminis-tration of SDF with increased renal blood flow in re-sponse to acetylcholine, comparable to that of shamcontrols (Table 2). Vascular endothelial function, perme-ability, and activation are influenced by NO bioavailabil-ity. We demonstrated a significant reduction in NObioavailability at 24 hours in CPB pigs relative to shampigs (mean difference �139.6 �mol/L; 95% CI, �210.7�mol/L to �68.5 �mol/L; p � 0.001; Table 2). The admin-istration of SDF resulted in levels of NO bioavailabilitysuperior to that observed in both CPB (mean difference221.6 �mol/L; 95% CI, 150.6 �mol/L to 292.7 �mol/L; p �0.001) and sham groups (CPB�SDF versus sham, 82.1�mol/L; 95% CI: 11.0 �mol/L to 153.1 �mol/L; p � 0.002).The CPB-mediated endothelial dysfunction also resultedin depletion of cortical adenosine triphosphate concen-
trations and significant hypoxia at the level of the outermedulla at 24 hours (Table 2). Preservation of NO bio-availability and endothelial function in CPB plus SDFtreated kidneys resulted in cortical adenosine triphos-phate levels and medullary oxygen tensions similar tothose observed in sham controls (Table 2).
Tissue AnalysisWe have previously demonstrated that post-CPB acutekidney injury is not associated with acute tubular necro-sis [7]. We did, however, demonstrate a flattened pheno-type of proximal tubular epithelial cells causing a pseu-dodilation of the proximal tubules in CPB kidneys, acharacteristic of proximal tubular epithelial cell stress(Fig 3A). We used immunocytochemistry to evaluate thebasis of the endothelial dysfunction and NO depletiondemonstrated in our functional and biochemical studies.Immunoreactivity of eNOS (Fig 3B and Fig 4A) and DBA
Fig 3. Tubular injury, endothelial injury andactivation, and renal inflammation at 24hours. Representative photomicrographs of (A)hematoxylin and eosin (H&E) stained corticaltubules. Immunocytochemistry/immunofluo-rescence stains for (B) endothelial nitric oxidesynthase (eNOS), (C) dolichos biflorus agglu-tinin (DBA) lectin, (D) inducible nitric oxidesynthase (iNOS), and (E) MAC-387, in sham,cardiopulmonary bypass (CPB), and CPB plussildenafil (SDF) groups. Scale bars represent140 �m (A), 46 �m (B), 70 �m (C), 75 �m(D), and 58 �m (E).
44tptD
en
ocnb
2174 PATEL ET AL Ann Thorac SurgSILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI 2011;92:2168–76
AD
ULT
CA
RD
IAC
lectin, a marker of the endothelial glycocalyx (Fig 3C andFig 4B) were significantly attenuated in glomeruli andvascular endothelium, respectively, in CPB pigs, indica-tive of endothelial injury. Conversely, CPB was associ-ated with increased expression of iNOS localized to theglomeruli and renal tubular epithelium (Fig 3D and FigA), increased cortical expression of endothelin-1 (FigA), and inflammatory cell infiltration most marked inhe glomeruli (Fig 3E and Fig 4B). Administration of SDFrevented CPB-induced phenotypic changes in proximal
ubular epithelial cells, preserved eNOS expression and
Fig 4. Quantification of protein expression and inflammation at 24hours. (A) Representative Western blots of protein expression, and(B) graph showing quantification by densitometry (n � 4) normal-ized to �-actin expression. (eNOS � endothelial nitric oxide syn-thase [white bars]; ET-1 � endothelin-1 [black bars]; iNOS � in-ducible nitric oxide synthase [gray bars].) (C) Shown is MAC-387cell count (white bars) per mm2 renal cortex and percentage stainingof dolichos biflorus agglutinin (DBA) lectin (black bars [n � 4]) inthe sham group, the cardiopulmonary bypass (CPB) group, and theCPB plus sildenafil (SDF) group. *p less than 0.05 CPB versussham; †p less than 0.05 CPB versus CPB plus SDF.
BA lectin staining, significantly reduced iNOS and
ndothelin-1 expression, and significantly reduced theumber of inflammatory cells (Fig 3A–3E, Fig 4A and 4B).
Comment
Main FindingsThis study has demonstrated that PDE-5 inhibition pre-vents post-CPB acute kidney injury, as defined by areduction in creatinine clearance and a rise in the specificurinary biomarker IL-18, by the preservation of NObioavailability and attenuation of endothelial injury, in-flammation, and hypoxia.
Strengths and LimitationsThe porcine recovery model of post-CPB acute kidneyinjury represents an ideal platform for the translationaldevelopment of novel renoprotective strategies. We haveevaluated the renal response to a clinical event, CPB, in amodel with significant homology to humans with respectto renal anatomy, function, hemodynamics, and the re-sponse to injury [6]. In this study, animals were recoveredwith assessment of outcomes and mechanisms at 24hours. That is important as changes in conventionalfunctional and biochemical markers (serum creatinine) ofacute kidney injury are observed clinically at 24 to 48hours after cardiac surgery [19]. The strength of themodel is underlined by the similarities between CPBinduced changes in creatinine clearance, urinary IL-18,and nitrates in this model and those reported in clinicalstudies [10, 12, 20]. The major limitation of this model isthat it does not reflect the multifactorial etiology ofpost-CPB acute kidney injury. Patient factors such asincreasing age, obesity, diabetes mellitus, preoperativerenal dysfunction, anemia, red cell transfusion, and ad-verse postoperative events are all important independentpredictors of acute kidney injury [2]. This study reflects
nly the beneficial effect of SDF with respect to the CPBomponent of acute kidney injury, and these effects mayot be apparent where injury has been caused primarilyy other events.
Translational RelevanceNitric oxide is central to numerous renal homeostaticmechanisms including the regulation of renal plasmaflow, glomerular filtration, and salt and water excretion[21, 22]; eNOS-derived NO also has an important role inmaintaining vascular endothelial structural integrity,promoting vasorelaxation and protecting against oxida-tive stress [23]. In this study, CPB resulted in endothelialinjury, loss of eNOS expression, and reduced NO bio-availability. There was no effect of CPB on salt or waterretention; however, that may be attributable to the highvolumes of crystalloid administered in an attempt toprevent type I errors arising from prerenal hypovolemia.Post-CPB loss of NO activity was associated with the lossof endothelial integrity, diapedesis of inflammatory cells,local vasoconstriction, hypoxia, and reduced glomerularfiltration. In contrast to the loss of eNOS, we observed an
increase in iNOS expression: iNOS was localized pre-
2175Ann Thorac Surg PATEL ET AL2011;92:2168–76 SILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI
AD
ULT
CA
RD
IAC
dominantly to the glomeruli and renal tubular epithe-lium, where increased local NO production is associatedwith the generation of reactive oxygen species and cel-lular injury [24, 25] as manifest in this study by renaltubular phenotypic change. We also observed increasesin expression of endothelin-1, a potent vasoconstrictorand mediator of inflammation and oxidative stress, whichin turn are likely to have contributed to the observedreductions in NO bioavailability at 24 hours.
Administration of the PDE-5 inhibitor SDF restoredNO bioavailability, and prevented the wide-rangingfunctional and histologic changes to the vascular endo-thelium, glomeruli, and interstitium that characterizedpost-CPB acute kidney injury in the model. That mayhave occurred either as a direct result of improvedperfusion during CPB, or by the preservation of eNOSactivity with its attendant antiinflammatory and homeo-static effects on the vascular endothelium. These findingsunderscore the importance of endogenous NO synthesisin the renal response to injury, and suggest that therapiesthat preserve or augment NO bioavailability may trans-late into effective primary preventions for acute kidneyinjury. Sildenafil citrate was approved by the EuropeanMedicines Agency in 2005 for the management of pa-tients with pulmonary hypertension in both oral andintravenous formulations [26] and has an excellent safetyprofile [27]. Our current findings support the evaluationof intravenous SDF for the prevention of acute kidneyinjury in a randomized controlled trial in patients under-going cardiac surgery.
In conclusion, we have demonstrated that SDF pre-vents post-CPB acute kidney injury by preserving NObioavailability and eNOS expression, and warrants eval-uation as a renoprotective agent in a randomized con-trolled trial. Our findings also highlight the homologybetween post-CPB acute kidney injury in pigs and hu-mans and support the development of large animal CPBmodels as translational steps in the development ofrenoprotective strategies in cardiac surgery.
The authors would like to express their thanks to Mr David Balland Dr Wolf Woltersdorf for their assistance with the biochem-ical assays. We would also like to thank Dr Anita Thomas for hertechnical assistance. This study was supported by grants fromthe Royal College of Surgeons of England, the British HeartFoundation (PG/08/044/25068), the Higher Education FundingCouncil for England (Walport) scheme, and the NIHR BristolBiomedical Research Unit in Cardiovascular Medicine.
References
1. Haase M, Bellomo R, Matalanis G, Calzavacca P, Dragun D,Haase-Fielitz A. A comparison of the RIFLE and AcuteKidney Injury Network classifications for cardiac surgery-associated acute kidney injury: a prospective cohort study.J Thorac Cardiovasc Surg 2009;138:1370–6.
2. Karkouti K, Wijeysundera DN, Yau TM, et al. Acute kidneyinjury after cardiac surgery: focus on modifiable risk factors.Circulation 2009;119:495–502.
3. Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality inacute kidney injury: a systematic review. Kidney Int 2008;73:
538–46.4. Dasta JF, Kane-Gill SL, Durtschi AJ, Pathak DS, KellumJA. Costs and outcomes of acute kidney injury (AKI)following cardiac surgery. Nephrol Dial Transplant 2008;23:1970 – 4.
5. Patel NN, Rogers CA, Angelini GD, Murphy GJ. Pharmaco-logical therapies for the prevention of acute kidney injuryfollowing cardiac surgery: a systematic review. Heart FailRev 2011 Mar 13 [E-pub ahead of print].
6. Murphy GJ, Lin H, Coward RJ, et al. An initial evaluation ofpost-cardiopulmonary bypass acute kidney injury in swine.Eur J Cardiothorac Surg 2009;36:849–55.
7. Patel NN, Toth T, Jones C, et al. Prevention of post-cardiopulmonary bypass acute kidney injury by endothe-lin A receptor blockade. Crit Care Med 2011;39:793–802.
8. Muzaffar S, Shukla N, Srivastava A, Angelini GD, Jeremy JY.Sildenafil citrate and sildenafil nitrate (NCX 911) are potentinhibitors of superoxide formation and gp91phox expressionin porcine pulmonary artery endothelial cells. Br J Pharma-col 2005;146:109–17.
9. Kellum JA. Acute kidney injury. Crit Care Med 2008;36(Suppl):141–5.
10. Ascione R, Lloyd CT, Underwood MJ, Gomes WJ, AngeliniGD. On-pump versus off-pump coronary revasculariza-tion: evaluation of renal function. Ann Thorac Surg 1999;68:493– 8.
11. Shoker AS. Application of the clearance concept to hypona-tremic and hypernatremic disorders: a phenomenologicalanalysis. Clin Chem 1994;40:1220–7.
12. Parikh CR, Mishra J, Thiessen-Philbrook H, et al. UrinaryIL-18 is an early predictive biomarker of acute kidney injuryafter cardiac surgery. Kidney Int 2006;70:199–203.
13. Nims RW, Cook JC, Krishna MC, et al. Colorimetric assaysfor nitric oxide and nitrogen oxide species formed from nitricoxide stock solutions and donor compounds. Meth Enzymol1996;268:93–105.
14. Shukla N, Jones R, Persad R, Angelini GD, Jeremy JY. Effectof sildenafil citrate and a nitric oxide donating sildenafilderivative, NCX 911, on cavernosal relaxation and superox-ide formation in hypercholesterolaemic rabbits. Eur J Phar-macol 2005;517:224–31.
15. Lim KH, Halestrap AP, Angelini GD, Suleiman MS. Propofolis cardioprotective in a clinically relevant model of normo-thermic blood cardioplegic arrest and cardiopulmonary by-pass. Exp Biol Med (Maywood) 2005;230:413–20.
16. Muzaffar S, Shukla N, Angelini G, Jeremy JY. Nitroaspirinsand morpholinosydnonimine but not aspirin inhibit theformation of superoxide and the expression of gp91phoxinduced by endotoxin and cytokines in pig pulmonary arteryvascular smooth muscle cells and endothelial cells. Circula-tion 2004;110:1140–7.
17. Coward RJ, Welsh GI, Yang J, et al. The human glomerularpodocyte is a novel target for insulin action. Diabetes 2005;54:3095–102.
18. Murphy GJ, Johnson TW, Chamberlain MH, et al. Short- andlong-term effects of cytochalasin D, paclitaxel and rapamycinon wall thickening in experimental porcine vein grafts.Cardiovasc Res 2007;73:607–17.
19. Haase-Fielitz A, Bellomo R, Devarajan P, et al. Novel andconventional serum biomarkers predicting acute kidneyinjury in adult cardiac surgery—a prospective cohort study.Crit Care Med 2009;37:553–60.
20. Lema G, Urzua J, Jalil R, et al. Decreased nitric oxideproducts in the urine of patients undergoing cardiac surgery.J Cardiothorac Vasc Anesth 2009;23:188–94.
21. Mount PF, Power DA. Nitric oxide in the kidney: functionsand regulation of synthesis. Acta Physiol (Oxford) 2006;187:433–46.
22. Kwon O, Hong SM, Ramesh G. Diminished NO genera-tion by injured endothelium and loss of macula densanNOS may contribute to sustained acute kidney injury
after ischemia-reperfusion. Am J Physiol Renal Physiol2009;296:F25–33.
2176 PATEL ET AL Ann Thorac SurgSILDENAFIL PREVENTS POSTCARDIAC SURGERY AKI 2011;92:2168–76
AD
ULT
CA
RD
IAC
23. Goligorsky MS, Brodsky SV, Noiri E. Nitric oxide in acuterenal failure: NOS versus NOS. Kidney Int 2002;61:855–61.
24. Araujo M, Welch WJ. Oxidative stress and nitric oxide inkidney function. Curr Opin Nephrol Hypertens 2006;15:72–7.
25. Wu L, Mayeux PR. Effects of the inducible nitric-oxide
synthase inhibitor L-N(6)-(1-iminoethyl)-lysine on microcir-culation and reactive nitrogen species generation in thecolleagues [4] have demonstrated, in a swine model, that
CnhaetNi
lccrt
fsanotpfi
P
DS3Se
© 2011 by The Society of Thoracic SurgeonsPublished by Elsevier Inc
kidney following lipopolysaccharide administration in mice.J Pharmacol Exp Ther 2007;320:1061–7.
26. Agency EM. Revatio: sildenafil. Available at: http://www.ema.europa.eu/ema. Accessed October 9, 2010.
27. Katz SD, Balidemaj K, Homma S, Wu H, Wang J, MaybaumS. Acute type 5 phosphodiesterase inhibition with sildenafil
enhances flow-mediated vasodilation in patients withchronic heart failure. J Am Coll Cardiol 2000;36:845–51.INVITED COMMENTARY
Postoperative acute kidney injury (AKI) is still a majorproblem in cardiac surgery. It contributes to highermorbidity, longer length of hospital stay, and highercosts and mortality [1]. Although cardiopulmonary by-pass (CPB) has been described as the most important riskfactor for the development of AKI, off-pump procedureshave not convincingly been shown to prevent this com-plication. A recent meta-analysis of randomized con-trolled trials comparing renal outcomes in on-pump withoff-pump coronary artery surgery could not confirm abeneficial effect of either of these procedures [2].
Important risk factors for postoperative AKI includeage, diabetes mellitus, preoperative impairment of renalfunction, low ejection fraction, obesity, perioperative he-modynamic derangement, red blood cell transfusions,and anemia. Several of these risk factors are not avoid-able, which encourages researchers to study differentways of preventing the development of postoperativeAKI or the worsening of preexisting renal dysfunction.
It is generally accepted that AKI is the result of isch-emia/reperfusion injury during cardiopulmonary bypassresulting from the combination of the enhanced systemicinflammatory response, lower perfusion pressure, andanemia. The focus of prevention of AKI has been primar-ily on pharmacologic interventions to improve renalblood flow, such as the administration of dopamine andfenoldopam. No improvement was demonstrated withthese strategies. The different approach of prevention oftubular obstruction and decreasing tubular oxygen con-sumption by administration of mannitol and diureticagents has been studied without showing a beneficialeffect on AKI. In the same hope of improving natriuresis,atrial natriuretic peptide and recombinant B-type natri-uretic peptide (nesiritide) have been studied, with noconvincing evidence of long-term protection of kidneyfunction. In a recent study, nesiritide did not prove tohave a beneficial effect on renal function in patients withacute decompensated heart failure [3].
Because of these disappointing outcomes, the focus ofresearch has shifted toward minimizing the effects ofsystemic inflammation. Initially N-acetylcysteine wasstudied in the setting of cardiovascular operations, with-out a positive effect on kidney function.
In an elegant study published in this issue, Patel and
PB-induced AKI can be prevented by intravenous silde-afil, a phosphodiesterase-5 inhibitor. The rationale be-ind the study includes the finding that inflammationnd oxidative stress result in endothelial injury, loss ofNOS, and an increase in inductible nitrous oxide syn-hase (iNOS) and endothelin-1 expression. By restoringO availability, sildenafil prevented loss of endothelial
ntegrity and preserved renal blood flow.Although the results of this study are fascinating, the
imitations are obvious. Cardiac operations are compli-ated procedures, with multiple factors apart from CPBontributing to the development of AKI. Multicenterandomized controlled trials in humans have to showhat this approach has significant merits.
It would be interesting to study the effects on renalunction of sildenafil in patients undergoing cardiacurgery without CPB. Independent of CPB, major surgerys such can induce systemic inflammation with potentialegative effects on renal function as has been shown inff-pump coronary operations. One could even speculatehat studies in noncardiac major operations should beerformed to evaluate the beneficial effects of sildena-l in the prevention of AKI.
ieter van der Starre, MD, PhD
epartment of Anesthesiatanford University School of Medicine00 Pasteur Drtanford, CA 94305-mail: [email protected]
References
1. Kumar A, Suneja M. Cardiopulmonary bypass–associatedacute kidney injury. Anesthesiology 2011;114:964–70.
2. Nigwekar S, Kandula P, Hix J, Thakar C. Off-pump coronaryartery bypass surgery and acute kidney injury: a meta-analysis of randomized and observational studies. Am JKidney Dis 2009;54:413–23.
3. O’Connor C, Starling R, Hernandez A, et al. Effect of nesirit-ide in patients with acute decompensated heart failure.N Engl J Med 2011;365:32–43.
4. Patel NN, Lin H, Toth T, et al. Phosphodiesterase-5 inhibitionprevents postcardiopulmonary bypass acute kidney injury in
swine. Ann Thorac Surg 2011;92:2168–76.0003-4975/$36.00doi:10.1016/j.athoracsur.2011.07.049
