Pathophysiology of protracted acute renal failure in man

Pathophysiology of protracted acute renal

failure in man.

S M Moran, B D Myers

J Clin Invest.

1985;76(4):1440-1448. https://doi.org/10.1172/JCI112122.

Postischemic acute renal failure (ARF) induced by cardiac surgery is commonly prolonged

and may be irreversible. To examine whether persistence of postischemic, tubular cell injury

accounts for delayed recovery from ARF, we studied 10 patients developing protracted (36

+/- 4 d) ARF after cardiac surgery. The differential clearance and excretion dynamics of

probe solutes of graded size were determined. Inulin clearance was depressed (5.0 +/- 1.7

ml/min), while the fractional urinary clearance of dextrans (radii 17-30 A) were elevated

above unity. Employing a model of conservation of mass, we calculated that 44% of filtered

inulin was lost via transtubular backleak. The clearance and fractional backleak of

technetium-labeled DTPA ([99mTc]DTPA, radius = 4 A) were identical to those of inulin

(radius 15 A). The time at which inulin or DTPA excretion reached a maximum after an

intravenous bolus injection was markedly delayed when compared with control subjects

with ARF of brief duration, 102 vs. 11 min. Applying a three-compartment model of

inulin/DTPA kinetics (which takes backleak into account) revealed the residence time of

intravenously administered inulin/DTPA in the compartment occupied by tubular fluid and

urine to be markedly prolonged, 20 vs. 6 min in controls, suggesting reduced velocity of

tubular fluid flow. We conclude that protracted human ARF is characterized by transtubular

backleak of glomerular ultrafiltrate, such that inulin clearance underestimates […]

Research Article

Find the latest version:

Pathophysiology of Protracted Acute Renal Failure in Man

S. Mark Moranand Bryan D. Myers

Division ofNephrology, Stanford University School ofMedicine, Stanford, California94305

Abstract

Postischemic acute renal failure (ARF) induced by cardiac sur-gery is commonly prolonged and may be irreversible. To examine whether persistenceof postischemic, tubular cell injury accounts for delayed recoveryfrom ARF, we studied 10 patientsdeveloping protracted (36±4 d) ARF after cardiac surgery. The differential clearance andexcretiondynamics of probe solutes of graded size were determined. Inulin clearance was depressed (5.0±1.7

ml/

min), while the fractional urinary clearance of dextrans (radii 17-30A)were elevated above unity. Employing a model of con-servation ofmass, we calculated that 44%of filteredinulinwas lost via transtubular backleak. The clearance and fractional backleak oftechnetium-labeled DTPA

_(Ir"TcIDTPA,

radius

=4

A)

wereidenticalto thoseof inulin (radius 15

A).

The time atwhich inulinor DTPAexcretionreached amaximum after an intravenous bolusinjectionwas markedly delayed when compared with control subjects with ARFof briefduration, 102 vs. 11 min. Applyinga three-compartment modelof inulin/DTPA kinetics (which takes backleak into account) revealed the residence time ofintravenously administeredinulin/DTPAin the compartment occupied by tubular fluidand urine to be markedly prolonged, 20vs.6min in controls, suggestingreduced velocity oftubular fluid flow.We concludethatprotractedhuman ARFis charac-terizedby transtubular backleakof glomerularultrafiltrate,such that inulin clearance underestimates true glomerular filtration rate by -50%,and bysluggishtubularfluidflow,whichstrongly suggests the existence ofsevere and generalized intraluminal tubularobstruction.Because allpatientsalsoexhibitedextreme hyperreninemia (16±2 ng/mlperh)that wasinverselyrelated toinulin clearance (rvalue = -0.83) and urine flow (rvalue

= -0.70), wepropose thatpersistent, angiotensin II-mediated renalvasoconstriction may havedelayedhealing oftheinjured tubularepithelium.

Introduction

Areversible form ofacute renalexcretoryfailurehaslongbeen known to occur in maninthe wake ofan

episode

of transient circulatory failure. Formerly, this

postischemic

renal

injury

was

characterized byaninitial

period

of

oliguria

which

persisted

in thetypicalcasefor7-10 d.

Thereafter,

adiuretic

phase

ensued that heralded the onset ofa gradual recovery from azotemia overthesubsequent 7-10 d(1,2). Thisacuterenal

injury

has beenaccuratelysimulated inexperimental animals by transiently interrupting renal blood flow for 40-60 min.

Micropuncture

Addressreprint requests to Dr. Myers, Division of Nephrology, S2 15. Receivedfor publication7December 1984 and inrevisedform3 June 1985.

studies have revealed that the oliguric phase of postischemic, acuterenalfailure

(ARF)'

iscaused in large part by intraluminal tubularobstruction accompaniedby backleakofa large fraction of glomerular filtrateacross anecrotic tubular wall (3-7). As in thehuman injury,recoveryfrompostischemicARF in the rat typically begins duringthe second weekaftertheischemic insult. Itis characterizedby anincrement inrenalperfusionrate, relief of intraluminal obstruction of tubules, and restoration of im-permeability ofthe tubular wall to the smallfiltrationmarker, inulin(7,8).

Overthe past decadeit has become routine practice to ad-minister to susceptible patients protective substances such as mannitol, furosemide, and dopamine before or during an isch-emicrenalinsult. Withtheadoption of this practice, an isolated single episode ofrenalischemiano longer appearssufficientto induceovert renal excretory failure in most instances (9). Rather, moreprolonged and often repeated episodes of renal ischemia appear to berequired beforetheclinicalsyndromeofARF be-comes manifest. A typical example ofthe latteristhat which follows prolongedanddifficult cardiacsurgery,notwithstanding the use ofprotectiveagents (10-12). It occursin the wake of prolonged cardio-pulmonary bypass

(.180

min [13]), during which the kidneys are perfused at a low rate (-250 ml/min) and at ahydraulicpressure low enough(-50mmHg) tobring glomerular filtrationto ahalt (14).Usuallythis initiating ischemic insult isprolonged and compounded byimpaired,postoperative cardiac function (10, 13).Wehaveshownpreviouslythatonly a minority of patients with this form ofARF(21/44) exhibit recoverywithin 2 wk postinjury. Aswith theaforementioned experimentalanalogueofpostischemicARFinthe rat, recovery followed circulatory improvement,asmanifested by restoration tonormalof cardiacperformance (10).

Because thephenomenonofpostischemicARFprotracted beyond2 wkhas becomeprevalent and carriesagrave

prognosis

(1

3, 15, 16),and because thereisnoanimalmodelofARFafter prolongedandpartial renalischemia,wehave attemptedto elu-cidate

directly

thefunctionalcorrelatesof thiscommonvariant ofARF in man. To do thiswe studied 10

patients

inwhom ARFafter cardiacsurgery wasprotracted beyond 14 d. We used the excretory

dynamics

ofsmallfiltration markerstoexamine the roleof intraluminal tubular obstruction,and thedifferential

1.Abbreviations usedinthis paper: ARF,acuterenalfailure; CRF,chronic renalfailure;

GQD,

inulin clearance;FEN.,fractional excretion ofsodium;

GFR,glomerularfiltration rate;

kDi,

fractional dextranbackleak;ki,

fractional backleak ofinulin;0,sievingcoefficient of theglomerular cap-illary wall for dextran; ,p,apparentsievingcoefficient for the i'th dextran basedsolelyonplasma and urine concentration measurements;OD, frac-tional clearances ofnarrowdextranfractions; PRA, plasmareninactivity; rD, radius ofthedextran;rDi,any dextranspeciesof the ith molecular radius; rn,molecularradiusofinulin;T.,time after bolusinjection

whenurinary inulin (or DTPA) excretionrateismaximal; T*, mean

residence time ofinulin/DTPAin plasma;T.,urinarydead spacemean

residencetime;UDS, urinarydead spacevolume; U/P,urine-to-plasma; V, urine flow rate; V,rateof urineflow in milliliters per minute.

J.Clin.Invest.

©The AmericanSocietyforClinical

Investigation,

Inc.

0021-9738/85/10/1440/09 $1.00

clearance of filtrationprobesofgradedsize to evaluate

transtu-bular backleak of

_glomerular

filtrate. Ourfindingsreveal arenal injurythat isquantitatively more severethanandqualitatively

different from human

postischemic

ARF of shorterduration.

Methods

Patient

population

and clinical

background.

Usingaprotocoland consent

procedure approved bythe Stanford

University

Committeeonthe Use of Human

_Subjects

in

Research, Stanford,

CA,westudied 10patients with

satisfactory premorbid

renal function whodevelopedprotracted ARF

subsequent

tosurgery of the heart andgreat vessels. There were

eight

menandtwo womenwho

ranged

in age from 34to77,with a

mean

_(±SE)

of64±4 yr. The

development

of

ARF, initially

nonoliguric

inall

_subjects,

wasmanifestedbythe presence of

progressive

azotemia

accompanied

by

isosthenuria,

and

by

the increased fractional excretion of sodium

(>1%)

andwater

(urine-to-plasma

inulin concentration ratio

<

_{20) (17).}

These

_subjects

wereselected for

study

because ofunremitting ARF

_despite

several weeksof intensive medicaland

surgical

care.

Ac-cordingly,

studies of renal

_physiology

were

_performed

23±4 d after the

onsetof renalexcretoryfailure

(Table

I).

For

_comparison,

twoother groups of

subjects

wereevaluated. These

were:

(1)

BriefARF group. This group included 32age-matchedpatients

(63±2 yr),

21 males and 1 I

females,

in whomARFresolved within 2 wk. In these

subjects,

ARFalsodevelopedafter

major

cardiacsurgery. Studies ofrenal

_physiology

were

_{performed during}

a

period

ofincreasing

azotemia,

on average 6±1 daftertheonset of renal

_{insufficiency.}

(2) Stablechronicrenalfailure

(CRF)

group.Thisgroup included 30patients whohad stable CRF ofdiverse

etiology

(43±3yr,15males, 15females).

Hemodynamic findings

weresimilarin

protracted

and brief ARF

(control

group

1)

populations,

asshown in Table II. Meanarterialpressure

averaged

74±3 vs.80±3 mmHg,

respectively.

Mean increase inbody

weights

(comparedwith admissionweights)wereidentical,5±3vs.5±1%.

Large

butsimilarpercentages ofpatientsin each groupwerereceiving vasoactivedrugsatthe time of the studies of renalphysiology.Inaddition,

thecolloid osmoticpressureofplasmawas not

significantly

differentin thetwogroups, 22.1±1.0 vs. 23.3±0.5 mmHg. Measurementsof he-matocrits alsodidnotsignificantly differ,33±3vs.35±1%. Thecardiac index tendedtobe lowin controlsubjects with briefARF(2.47±0.12

liters/min

per

M2).

Becauseappropriatevascular catheterswere nolonger in

place

in

subjects

withprotractedARF, cardiac indexcould notbe determined.Thetendencytowardalowarterialpressuredespitea

volume-expanded

stateand theadministration ofdopamine (n = 10),and in

TableI.ClinicalFeaturesof10 Patientswith ProtractedARF

Day

Day dialysis Dayof Day

Patient Age/Sex ofstudy initiated recovery of death

1 65/M 15 41

2

_34/M

25 1 49

3

64/M

16 14 29

4 60/M 21 26

5 77/M 35 33 46

6

49/F

46 44 47

7

77/F

25 8 30

8

_68/M

47 17 53

9 _75/M 15 12 23

10 73/M 14 14

Mean 64 23 18

±SE ±4 ±4 ±5

Table II.

Hemodynamic

Findings

in

_Brief

and Protracted ARF Prtracted BriefARF ARF

No.

patients

32 10 Meanarterialpressure

(mmHg)

80±3 74±3

Colloid osmotic pressure

(mmHg) 23.3±0.5 22.1±1.0

Increment in

_body

weight (%)

5±1 5±3

Hematocrit(%) 35±1 33±3

Inotropicagents (%) 87 100

After

load-reducing

agents(%) 50 20

some casesepinephrine(n = 5), suggested that they too manifesteda

low cardiac output state.

Study protocol. Patientswereexaminedduringanepisode of ARF thatwasprotracted,asjudged byfailure of recovery of renalexcretory functiondespitetwo or more weeks ofintensive medical care. All studies

werebegun24 hor moreafter the most recenthemodialysistreatment whenpatientswerehemodynamicallystable and nothypovolemic.Fluid balancewasmaintainedduringthestudy.Clearance studies were per-formedby giving asingle primingdose of inulin(60 mg/kg ofbody weight) and technetium-labeled diethylenetriamine pentaacetic acid

([9'Tc]DTPA,

ImCi).Apolydispersesolution of neutral dextran 40 (Rheomacrodex,Pharmacia FineChemicals, Uppsala,Sweden, 140mg/ kg)wasthereaftergivenbyinfusionovera20-min period.Anin-dwelling Foleycatheterwasusedtocollect urinespecimens.To evaluate the ki-netics of filtration marker clearance and distribution within thebody, blood and urinespecimenswerecollectedduringthe immediate post-prime periodat 1, 3, 5, 10, 15, 30, and 60min.Thereafter, clearance studieswerecommenced with blood and urine collected each half hour for2 h and thenhourlyfor an additional 2-4 h. From these plasma and urinesamples,theurinaryclearances(C)of inulin andDTPA,the frac-tional excretion of sodium

(FEN.)

and theurine-to-plasma(U/P)

os-molality ratiowere determined. Clearanceswerecalculated from the ratio (UV)/(PT), where U = urinaryconcentration of the clearance

marker,V=urine flow rate, T=time ofcollection,andP = logmean

of clearance marker concentration in plasma samples bracketingthe urine collection.

[p'Tc]DTPA

wascounted in urine andplasma witha Packard gamma scintillation spectrometer(model 3002, Packard In-strumentCo., Inc.,DownersGrove, IL).The assay oftheremainingtest

solutesemployedin the presentstudyhas been described in detail

pre-viously (9, 10).Inaddition, plasmareninactivity(PRA) was determined

asthe rate ofimmunoassayableangiotensinI generated in plasma in-cubatedatpH7.4 andat370C(18).

Aftergelchromatographicseparationof plasma and urine on pre-calibratedSephacryl S-300 columns(Pharmacia FineChemicals), the fractional clearances ofnarrowdextranfractions(0D) were computed

usingtheequation

OD=

_{(U/P)D/(U/P)IN,}

₍₁₎

where

(U/P)D

and

_(U/P)N

referto

_{urine-to-midpoint}

plasma

concentra-tion ratios of dextran and

_inulin,

_{respectively.}

In thetwocontrolgroups,clearancestudieswere_performedby

ad-ministering

the

_priming

dosefollowed

_by

aconstantinfusionof inulin

resulting

ina

steady

state

plasma

concentration. Afteranequilibration period

lasting

60 min, four30-mintimed urine collectionsweremade and bloodwasobtainedatthe

_beginning

and endof each_period.

Clear-ances werecalculatedas_reported

_previously

_(9-12).

(inulin anddextran) from damaged tubules have been reported in detail previously(I1).In brief, we have assumed that glomerular permeability toinulin and dextran was not different between patients with protracted ARF and those with CRF in whom inulin clearance was depressed to similarlylowlevels (< 15ml/minper 1.73m2).It followsthat any elevation of6D inARFwhen compared with 0Din CRF might be due to the backleakofinulin through damaged tubules at a greater rate than backleak ofthe largerfiltered dextran molecules. The use of a population with CRF as a control group is novel. It has the advantage that even the smallestcomponent molecules in dextran 40 are retained in serum be-cause of thelow glomerular filtration rate (GFR), thus permitting for thefirst timeacomparison between the experimental and control group ofthefractionalclearanceof dextran molecules of radius below 20A.

When thepermeability properties ofthe tubule are normal, all inulin and dextrancleared by the glomerulus and appearing in the glomerular filtrate isassumed to reach thefinal urine. It is thisassumption that permits estimation ofthe dextran ultrafiltration characteristics of the glomerularcapillarywall(19). For the kidney in ARF, however, we pro-posethattubule backleak of inulin(and alsodextrans) occurs; if we furtherassumethat suchleakageis"passive"(thatis,occurs solely by ultrafiltration and/or diffusionthrough the damaged tubulewall), then the rateof leakage ofany such solute should benearlyproportionalto theconcentration ofthatsolute in theglomerularfiltrate. Under such circumstances,thefractionalbackleakof inulin(definedasthe rateof backleakdivided bytheGFR)will be constant(kjn).Preciselythe same criteriaapply to any dextranspecies oftheith molecular radius

_(rDi),

leadingto afractionaldextran backleakof

kDi.

However, thepermeability ofthedamagedtubule wall to dextrans would be expected to be adeclining function ofthesize ofthe dextranmolecule,

kDi

=

_f(rDi),

(2)

where

_kDi

-

_k,.

as

_rD.

-

_{r..; kD}

-*0as r> -* 00.Simplemass-balance computations of inulinanddextran clearedbytheglomerulusleadto

thefollowingrelationship(I1):

Oam

=O[I -

kDJ/(l

-kj)= [I -

_f(rD)]/(l

-k1), (3) where

_O.,,

is the apparentsieving coefficientforthei'th dextran based solelyonplasmaand urine concentration measurements, and ( isthe sieving coefficient oftheglomerularcapillarywallfor that dextran. Qual-itatively, we can deducefrom Eq. 3 that, for largedextran molecules where

kD;

= 0,

Oapp

> e(since [I -

k,,]

>0), while forsmall dextran

moleculeswhere

_kDi

= ki,,O1 -p 0.

Assumingthat the dextransieving characteristicsof theglomerular capillarywall inprotractedARFarethesame asthose insevereCRF,

Eq. 3 can be reduced to:

Oam/0

=(I -k)/(l -

kjn)

=[1/(1-

kjn)]

- [kD/l-kin)], (4) where 0 nowequalsthefractionaldextran clearance in CRF.

Further,

kinis a constantforagivensubject,whereas

kD;

isafunction of ri. As

rD1-

r3,k,,- _kj,andas_rDi-oo,

kDi

- 0.Hence,the ratio

(O.pp/O)

should increase withrDifromavalueof

unity (for

small

rDi)

to alimiting

valueof1/(1-

_k1.)

as rincreases.Since,as r-

_00,

0-*0,a

_plot

_of(O../

0)vs.rshouldpermitareasonablyaccurateextrapolationtothelimiting

valueof

(O,./O)

-1/(1-

_ki,).

Thispermitscalculationof

_kn,,.

Since

k,

is the fraction of filtered inulin that leaks

back,

correctedGFR=[l ]Cil] (5)

whereC,, denotes inulin clearance.

Measurementofthe inulinresidencetimein theurinarydead space. Aftertheintravenousadministrationof inulinorDTPA,the filtration marker distributes inplasma,inanextravascular,extracellular space and in a third volumeextendingfrom Bowman's spacetothe

urinary

bladder (theso-calledurinarydead space[20]).There isadelayin the appearance oftheinjectatein the final bladder urine

(10,

20).Inulin excretion rises

slowlywithtime,reachesapeak,and thendeclinesparipassuwith the

falling

plasmaconcentrationofinulin.Thetimerequired for inulin to

traversethe dead space

(T.)

andthetimerequiredfor peakurinaryinulin excretiontobe observed_dependupon severalfactors.Thefirstis the existence ofa_poolof tubular fluid and urine knownastheurinarydead space, which consists of partorallof the volume of the renal tubules,

pelves,

andureters(10,20). Largerdead space volumes (or lower tubular fluid flowrates)will result inagreaterdelayin inulin appearance anda

longertimeto_peakexcretion. The secondpotentialfactoraffectingboth

T.

and thetimetopeak

urinary

inulin excretion is leakage ofglomerular filtrateacrossthedamagedrenal tubularepithelium.Ifpresent, backleak decreasesnet urinary clearanceofinulin and,therefore,the apparent GFRaswell. A thirdsetoffactorsaffecting

T.

is therelationship of

plasma

volumetoapparent GFR and theassociatedrateof clearanceof inulinfromplasmaintothe extravascular space. Asplasmavolume

in-creasesortherateof extravascularaccumulation decreases, T,, prolongs. Wehavepreviouslyanalyzedthekinetics of inulinexcretion by

de-vising

a3-compartmentmodel thatpermitsdeterminationof themean

residence timeof inulin inplasmaandtheurinarydead space(10).In the presentstudywehavemodifiedtheearlier3-compartmentsolution toincludetheeffectsof transtubular backleakofthefilteredsolute. The modification

yields

the

following

formulation:

(Tm..)-

=

((T.)-'

-

(T*)-')/(Ln[(T*/T,)]).

(6)

T.,,,,,,

time after bolus injection when urinary inulin (or DTPA) excretion rateis maximal. T,,urinary dead space mean residence time where

_T.

=

_UDS

_*V;

_UDs

isthe volume inmilliliters ofthedead space andVisthe

rateof urine flowin milliliters per minute.

_Tp,

meanresidence time of inulin/DTPA in plasma. It is determined by the plasma volume and by theclearances ofinulin or DTPA: (a) from plasma into the urine by glomerularfiltration; (b) from plasma into the extravascular space; and finally(c),from tubularfluid back into the vascular and extravascular spacesbytherateof backleak. During ARF, with low urinary excretion rates, T* iscalculated at longtimes after the bolus injection of inulin/ DTPA, whenplasma and extravascular spaces have equilibrated and the "'clearance" of inulin or DTPA into the extravascular space is nil. The rateofbackleakof a given solute depends upon its molecular size and theextentof increased permeability oftheinjuredtubularepithelium. Thus, T*= plasma volume/[corrected GFR- (I -_ka,)].As backleak increases,neturinaryclearancediminishesand the meanplasmaresidence time rises.Therefore, thetime topeak urinary inulin excretion may becomesubstantially prolonged.

Statisticalanalysis. Allresultsare expressed as mean±SEM.Data fortheexperimental andtwocontrolpopulationswere analyzed and compared with the Stat Plus General Statistical Package (Human Systems Dynamics, Northridge,CA).Anunpaired two-tailed t test or the Mann-WhitneyUtest wasusedto testthesignificanceofthedifferencesobserved. Linear regressionanalysiswasused toexamine the strength ofassociations betweenvariables.

Results

Clinical andlaboratoryfindings.Theresults of10studiesduring thecourse ofprotractedARFareshown in Table III and

com-pared with those of 32subjects havingARFof brief duration. As noted previously, episodesof ARFwere lengthyin the

ex-perimentalgroup,averaging36±4d until deathsupervenedor

recoveryoffunctionwasobserved.7of 10patientswere_receiving dialytic support. Theprofound depression in renal function is

indicatedbythe inulinclearance,whichaveragedonly 5.0±1.7,

avaluesignificantlyless than the_{corresponding}value of14±1 ml/minper 1.73 m2observed in the brief ARFcomparison pop-ulation(controlgroup 1).5 of 10patientswithprotracted ARF

were oliguric(V. 0.3 ml/min); forall studies, urine flow

TableIII.LaboratoryFeaturesof10 Patients with ProtractedARF

C. CDTPA U/P U/P

Patient (ml/min/1.73 i2) (ml/min/1.73i2) V FEN inulin osmol

mlmin %

1 10.0 10.0 1.4 7.1 7.0 1.2

2 12.0 12.0 0.5 11.5 4.0 0.8

3 0.4 0.4 0.07 3.3 6.0 1.0

4 8.0 8.0 1.1 2.7 7.0 1.0

5 1.0 1.0 0.3 36.6 3.0 1.2

6 0 0 0.01 -

-7 1.0 1.0 0.1 1.0 8.0 1.0

8 1.0 1.0 0.2 6.3 7.0 0.9

9 2.0 2.0 0.5 3.3 6.0 1.1

10 15.0 14.0 1.4 0.9 11.0 1.2

Mean 5.0 4.9 0.6 8.1 7.0 1.0

±SE ±1.7 ±1.7 ±0.2 ±3.5 ±0.7 ±0.04

Nonoliguric ARF of brief duration(n=32)

Mean 14.0 ND* 1.9 8.0 6.0 1.05

±SE ±1.0 ND* +0.4 +2.0 ± 1.0 ±0.05

Pvalue <0.05 ND* <0.01 NS NS NS

* ND, not done.

elevated(8.1±3.5%); isosthenuriawaspresent(U/P osmolality 1.0±0.04); the urine-to-plasma inulin concentration ratio was low (7±0.7). Alloftheforegoing findingsareindicativeofsevere ARF(17). Wealsocomparedthe clearance of

[99mTc]DTPA

to that of inulin and found the clearancestobeessentiallyidentical, 5.0±1.7 vs.4.9±1.7 ml/min per 1.73 m2, respectively.

MeasurementsofPRA wereavailableinnineof thepatients with prolonged ARF. All patients exhibited significant hyper-reninemiawhen compared with normal euvolemic volunteers studied in ourlaboratory(16.1±1.7 vs.0.6±0.2ng/ml per h, P

<0.001). PRA levels wereinversely related tobothurineflow rate and inulin clearance (coefficients of correlation = -0.70 and-0.83, respectively).There was also a weakerinverse

rela-100 I

10 - NORMAL

E

E Figure 1.

Urinary

inulin

ex-BRIEF ARF cretion(Ui.,,fnV)vs.time

curvesfor13 control

pa-1 _ _ tients withARFofbrief

du-ration (middle) and 10 ex-perimentalpatients with

/fPROTRACTED

_{PROTRACTED ARF}ARF _protractedARF

_(bottom).

Theinulin excretion/time

curvefor normal postopera-tive subjects studied by us

0 30 60 90 120 150 previously(10)is also

TIME(min) shown for comparison (top).

tionshiptotherateofsodium excretion(coefficientofcorrelation

--0.5).

Timetopeak inulinexcretionand

T,.

Asshown in Fig.

1,

meaninulin excretionratereached amaximumat - _120min

in patients with protracted ARF. In comparison, subjects with ARF of brief duration (controlgroup 1) achieved maximal ex-cretionratesmuchearlier,at I

min,

albeitathighlydifferent maximal rates commensurate with renal function. The latter value is similartothat observed byuspreviouslyinpostoperative subjects with normal renal function(10).The estimated

urinary

dead spacevolume

(UDs)

in patients with protracted ARFwas

smaller than that of brief ARF controls (10±1.6 vs. 15.6±2.0 ml, P<0.05). Further insightintothe physiology of theurinary deadspace maybe obtained with the aid of the linearregression analysis depicted in Fig. 2, which demonstrates therelationship between urine flow rate V and

UDs.

Asshown,

UDS

isdirectly relatedtoV(coefficient of correlation=0.7). Thus,UDsexhibits akinetic compliance. Hence, with lower urine flowrates, asin subjects with prolonged ARF,

UDS

is also diminished. From the relationship in Fig. 2, itcanbe predicted that aminimaldead

spacevolume of 6 ml exists(3mlperkidney and ureter),even

with complete anuria.2

The Tu was markedly increased in subjects withprotracted ARF (80±45 min), andgreatlyexceededthatofcontrolpatients withbrief ARF (9±1

min,

P < 0.01). However, Tuis directly proportional to the volume of the urinary dead space but in-versely relatedtourineflowrate,i.e.,

2.Ourfindings are in keeping with an earlier study in the dog by Bojesen (20). Hedemonstrated both kinetic compliance and a similarly small volumeof the dead space at low rates of urine flow in canine kidneys weighing 100-150 g. Bojesen (20) validated the veracity of estimates of dead space volume from the kinetics of inulin excretion by making si-multaneous direct determinations of pelvic volume over a wide range of urine flow rates.

Figure 2. Therelationship of

_UDs

to_V,

_{demonstrating}

kineticcompliance.

Tua(UDS)AP)

V).

Since

UDS

alsodeclinesasVdecreases

(Fig. 2),

any

prolongation

of T0caused

by

falling

urine flow isto somedegree abbreviated

by

thecoincident reduction in UDS.

Fractional inulin

leakage.

Wehave

previously

demonstrated the transtubular backleakof filtered inulin and dextran during ARFof brief duration

(1

1,

12).

In

Fig.

3,

the meanfractional dextran clearance

profile for patients

with

protracted

ARF

(OAF)

is

plotted

asafunction of the

gel

chromatographic

radiusof the

dextran,

rD, in the interval 17-40 A. For

comparison,

also

de-picted

is thefractional dextranclearance

profile

of the 30

subjects

of the second controlgroupwith stable CRF

(OCRF).

As

shown,

OARP

iS strikingly and significantly (P < 0.01) elevated above0cRF for all dextran fractions in therDinterval 17-40A. Several

points

deserveadditional

emphasis:

(1)

OCRF approaches

but doesnot

exceed

unity

for

rD

17-20

A (0.96

'

0.

_'0.92);

(2) OCRF

for

larger

dextran fractions

(rD>

20

A)

declines

monotonically

toward zero;

(3)

in

striking

contrast,

only

when rD> 28

A

is

OARF

less than

unity; (4)

when a

single patient (#3),

in whom therewas

no detectable backleak is

excluded, OARF

significantly

exceeds

unity

forrD= 18 A and 19 A

(P

<

_0.05);

and

₍₅₎

forrD= 17 and 20

A <rD

< 26

A,

OARY

also exceeds

unity, although

the

difference

fails

narrowly

toachieve statistical

significance.

16 20 24 28 32 36 40

DEXTRANRADIUS(A)

Figure3. Themeanfractional dextranclearanceprofilein 10patients withprotractedARF(top)iscomparedwith that of 30 controlsubjects with CRF(bottom).All resultsare_expressedasmean±SE.

From Eq. 4, it may be calculated for the experimental pop-ulation that -44%of filtered inulin emigrates from the tubular lumen(fractionalinulin leakage,

kin)

and thusfails to reach the final urine. Because of interpatient variability of fractional dex-tranclearance profiles for the protracted ARF studies, fractional inulin leakage was determined for each subject, with one excep-tion. The value of

kCn

wasthen used to correct those experimental variables affected by inulin backleak: inulin clearance,

_Tm.,,

and

TU.

Data are showninTable IV. There was asinglesubject (#3) whoapparentlyhadno backleakdespitesevereARF.Otherwise, values for

kin

varied from 0.21 to 0.66, reflecting considerable variability in the extent of increased tubular permeability, with the result that transtubular inulin leakage may differ more than threefold between individuals.Insubject6, no inulin or dextrans were detected in the final urine despite the presence of both in plasma, which could theoretically reflect an extreme backleak of100%, i.e.,

kin

= 1. (In theabsenceof measurable dextran and inulinclearances, however, actual fractional backleak cannot be calculated.) For nine patients, inulin clearances corrected for backleak varied from 0.4 to 27.9 ml/min, averaging only 9.6 ml/min.Clearly,inulinbackleak cannot wholly account for the gravereductionin GFR associated with protracted ARF. In these nine subjects,the time to peak inulin excretion, when corrected for backleak, was shortened from 120 to 102min. The

Tu

de-creasedfrom 33.8to20 min. Even when correctedfor transtu-bular inulin backleak, both the time to peakinulin excretion and the

Tu

are considerably prolonged when compared with corresponding values in the controlpopulation with brief ARF (Table IV).

Fractional dextran leakage. Not only filtered inulin leaks back from the damaged tubule, however.Weresuchto be true, then afractional dextran clearanceprofile forthe ARF popu-lationverydifferentfrom that actually observed would be pre-dicted.

Fig.

4

displays

thecalculatedfractionaldextran clearance

(O@aculatd)

that wouldobtainifthe tubulediscriminated

perfectly

against molecules larger than inulin,sothatapproximately one-half

of

filtered inulin,but nofiltered dextran,were to escape the

injured

tubule. Thehypothetical, calculated fractional clearance profile (top curve)representsthereforeatwofold multiplication of control fractional dextran clearances (bottom curve). The fractionaldextran clearanceprofile observed for protractedARF

(middle

curve) clearly differs from both the upper and lower curves:it iselevated above the lower curve but

depressed

below the upper curve because both dextran and inulin leak

back,

al-thoughthe percentage of inulin lost

(rinulin

= 15

A)

exceeds that ofthelargerdextran

molecules.

Thedifferencebetween observed (o) and

predicted

(p)curvesreflectsfractional dextran

leakage

(kD

= [p - o]/p). As shown, the difference between the two

curves varies

continuously

as rD

changes, declining

as rD in-creases.

The quantities

kin

and

kDi

(where

Di referstoan

arbitrary

dextran fraction of radius

i)

represent the individual rates of transportof inulinand the ith dextran acrossthe

damaged

tu-bular

epithelial

barrier. The ratio

(kDdkin)

may betaken, there-fore,as ameasureoftherelative

permeability

of the

epithelium

totheithdextranwith respecttoinulin. An examination of(kD(/

kin)

for dextrans in the radius interval 17-40 A is therefore in-structive.The results of these calculationsare

plotted

in

Fig.

5. The

accompanying

fractional dextran clearance

profile

of the glomerulus in CRF

(OCRF,

Fig. 3)

is included for

comparison.

Relativetothe

_glomerulus,

itappearsthat the

damaged

tubule

exhibits somewhat diminished

permeability

to dextrans of all

20 -18 w 16 o 14 0

uj 12

4W 10

I6

4

2

8

4x

2

:

/

r=0.7

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 URINE FLOW RATE(ml/min)

1.4 1.2

z

1.0

Z4 uiz

<1_< 0.6

JCJ

-iQ.0,

u

* P<O.01

** p<0.001

-0ARF e

_o _~

___ _____9_______ ___

I-0.4

-0.2

Table IV. Fractional InulinLeakageand CorrectedExperimental Variables in ProtractedARF

Cia T. T.

Patient kin Raw Corrected Raw Corrected Raw Corrected

ml/min mllmin min min min min

1 0.21 10.0 12.7 48 42 14 11

2 0.57 12.0 27.9 90 60 38 16

3 0 0.4 0.4 300 300 71 71

4 0.55 8.0 17.8 30 27 9 5

5 0.66 1.0 3.0 66 54 13 6

7 0.58 1.0 2.4 300 240 90 38

8 0.60 1.0 1.7 180 132 45 18

9 0.59 2.0 4.9 54 42 i4 6

10 0.22 15.0 19.2 30 27 10 8

Mean* 0.44 5.6 9.6 120 102 33.8 20

±SE* +0.06 ±1.7 ±3.0 ±36 ±30 ±9 ±7

Nonoliguric briefARFcontrols (n = 13)

Mean 0.46 14.0 28.0 11 9 7 6

±SE ±0.05 ±2.0 ±3.8 ±1 1 ± 1 ±1

Pvalue NS <0.01 <0.01 <0.001 <0.001 <0.01 <0.01

* n=_{9, because no inulin or dextran was detected in the urine ofpatient6 upto7h}

after

_{bolusdespite adequate plasmalevels.Calculation of}

fractional backleak for this patient wasthereforenotpossible.

molecular sizes examined, and that the permeabilitydeclines at arate (slope of the curve) which is similar to that of the glo-merular sieving curve for rD> 24 A. The relative tubular

per-meability is zero at rD>34 A, even in the presence of a severe tubularinjury. This finding provides an explanation for the ob-servationinthe rodent model of postischemicARFthatproteins ofmolecular radius .30 Acannot traversethedamagedtubular wall(3, 6).

Recoveryoffunction after protracted ARE.4ofthe 10patients with sustained ARF recovered satisfactory renal excretory ca-pacity.Allsurvived (patients 1,2, 4, and 10). The remainingsix

patients failed

torecoverrenalfunctionand eachdied.Data are summarized in Table V.Inulin clearance, though profoundly subnormal, wasnonethelessgreater in the subgroup that sub-sequently recovered(11± 1 vs. 1±0.3ml/ninper 1.73m2).The superiorrateof inulin clearancewasfurtherreflected by a lesser timetopeak inulin excretion(50±14 vs. 170±60 min).Those

2.0 WCALcuLATED 1.8 -'K

DEXTRAN BACKLEAK 1.6

AtZ 1.4

a 2 1.

_Figure

_{4.Fractional dextran}

Z 1.0 9AR clearanceprofileof control

LU 0.8- populationwithCRE

(bot-4W 0.

.J_{UQ 06 _ \}

atom),

the

hypothetical

pro-0.6

-filecalculated as if one-half

0.4- offiltered inulin butno

0.2 dextranleaked from the

tu-o__.,,__.,__.,__., bule(top),and observed

16 20 24 28 32 36 40 profile in subjects with

pro-DEXTRAN RADIUS (A) tractedARF(middle).

who ultimately recovered satisfactory function hadsignificantly higherurine flow rates, while those who exhibited no recovery werefrankly oliguric (1.1±0.2 vs. 0.2±0.01mI/min). Commen-suratewith the urineflow,urinary dead space volumes also dif-fered

(I6±2

vs.7±1 ml). The combination ofhigherdead space volume but greatertubularflow nevertheless resulted in a shorter

Tu

(18±6vs.47±13min).Patients who recovered also exhibited significantly less hyperreninemia(2.3±0.3 vs. 27.0±3.0

ng/ml

perh). Thus, theextentofinjurytothenephron in ARF may represent a graded response to the degree of underlying renal ischemia. Asthe latter increases in severity and duration, the manifest injury appears to advance through a brief form of ARF (controlgroup 1)to aprotracted, but reversible injury, or ulti-mately to an irreversible renalinjury.

1.2 1.0- _0CRF

2°

0.8

S

T

X

U. I

'U

o

0.6->C.4- 0 U,

TUBULeARF

0.2-18 22 26 30 34 38 42 DEXTRAN RADIUS ₎

Figure S.Thedextransieving behaviorofthe

_glomerular

_capillarywall in CRF(upper curve)iscomparedwith thatcalculatedforthe

dam-agedtubule inARE(lowercurve).

TableV. Clinical and Biochemical Characteristics ofSubjects Who Recovered vs. Failed to Recover Renal Function after Protracted ARF

Sur- Cow

Recovery Age vived V (ml/min/1.73m2) FEN,. kin

_T.

T PRA

n mlmin % ml % min min ng/ml/h

Yes 4 58±7 4 1.1±0.2 11±1 5.6±2 16±2 39±8 50±14 18+6 2.3±0.3

No 6* 68±4 0 0.2±0.1 1±0.3 10.1±5 7±1 49±10 170±60 47±13 27.0±3.0

Pvalue NS <0.05 <0.01 <0.01 NS <0.05 NS <0.01 <0.05 <0.01

*

_{FEN,, T.o, T.,}

and PRA:n=_5.

Discussion

Tubular obstruction. The severityofthe renal injury sustained by the subjects of thecurrent study isattestedtoby the loss of 96%ofnormal renalcapacity toclear inulin from plasma. Fur-thermore, this impairment of function persistedon averagefor

5wk,terminating either in recovery in a minority of patients

ormoreoften in the patients' demise. Compared with the results

ofprevious studies from this laboratory, both the timetopeak inulinexcretion and

_T,

weremarkedly prolonged (10). Profound

oliguriawasalsomore frequently observed. These findingsare

similartothoseof themaintenance phase ofpostischemicARF inthe rodent. In keeping with themoredirectmeasurementsin therat(3-7), the prolongation of

Tu

inourpatientsmayreflect

significant tubular obstruction. Inaprior study of nonoliguric postischemic ARF inman,inwhich measured renal impairment

was less severe and of shorter duration,

T,

was found to be normal (10). Given the demonstrated variabilityamong indi-vidualnephrons of tubularobstructionin the rodent model (4, 5),apossible explanation for this findingwastheexistenceofa

smallsubpopulation of widelypatent, unobstructed tubules(10). Theexistence ofa sparsecollection of tubules having unimpeded

flowmaybeabettedbyhigher tubular fluidflowrates(signaled by nonoliguria); tubule cellularflotsam which would otherwise obstruct the tubular lumina couldbeliterallywashedaway(10). Theseverereduction in urine flowrateobservedinourpatients

withprotracted ARFmay havereflectedsluggishtubularfluid

flow. This, in turn, could have contributedto morediffuse in-tratubularobstruction, and hencetothe prolongationof both timetopeak inulin excretion and

Tu.

Thus,the threefold pro-longation of

Tu

observedinthecurrentstudysuggests the absence ofcompletelypatenttubulesand thepossibilityofamuchmore

homogeneouspatternoftubularinjury.Thisfindingisnecessary,

but insufficient by itself, to validate the concept that tubular obstruction maybecome generalizedinthisprolongedvariant

ofhuman ARF.

Tubularleakage. The presenceoftranstubular backleakin ourexperimental subjectswithprotractedARF isattested toby

thefindingoffractional dextran clearances thatexceedunity,a

phenomenon not observed in controls with CRF. Fractional

dextran clearances in excess ofunity must signify that either

smallerinulin molecules leak back fromthe tubule withgreater facility than larger dextrans, orthat larger dextrans permeate theglomerularfiltermoreeasilythan thesmaller inulin molecule.

Itisdifficulttoconceiveofanycircumstanceinwhichalarger molecule couldby passive forces move more rapidly through theglomerular capillarywall thanasmaller molecule.Weinfer,

therefore,thatthe formerexplanation iscorrect and that there

is inprotracted ARF a size-dependent backleak of filtered mol-eculesto which the tubule is normally impermeable.

From Eq. 3 we calculate that

_kins

the fraction offiltered inulinthat leaked back, was 0.44, a finding in good agreement with themagnitude of inulin backleak in the rat with an anal-ogous acuterenal injury (4, 6, 7). However, the calculation of ki,restsupon the assumption that the Bowman'sspace-to-plasma concentration ratio(or

glomerular

sieving coefficient) of each dextranmolecule in ARF does not differ from the corresponding value in the control population (group 2). Transcapillary dextran flux into Bowman's space is governed in part by the same de-terminantsthat controlthe ultrafiltration process. GFR in ex-perimental

ARF

declines primarily because of a reduction in the net transmembrane ultrafiltrationpressure(4, 8, 21), while

this

latterquantity is elevated inCRF,inwhich GFR depression iscaused by areduction ofthe glomerular ultrafiltration coef-ficient, Kf (22). Despitethesedifferencesin thedeterminantsof GFR inthe twodisorders, theirneteffect is to lower the filtration fraction. A lowerfiltration fraction will blunt the rise in con-centration ofaretainedmacromolecule as plasma transits along theglomerular capillaries (23),thereby lowering bothdiffusive

and

convectivecomponentsof transcapillarydextran transport (19). Sincetheextent bywhichthefiltration fractionislowered inprotractedARFisunknown, the assumptionofdextransieving coefficients for the glomerulus identical to those in the CRF group (and hence ourcomputation of kin)may be in error.

Ac-cordingly,

thecorrectedvaluesfortrueGFR andmeanresidence timeof inulin in the urinarydead spacein protractedARF(Table IV)should beregardedasapproximationsonly.

Uncertainty

regarding

theactualvalueof

kn

alsolimitsthe precision with whichtheratio

kDi/kin

can be used tocharacterize the

permselectivity

ofthe damaged

tubular

wallin protracted ARF. Inthiscase, moreover,it isnecessary to makean additional

assumption

that the transport of inulin through the

damaged

tubular

epithelial

barrier in

protracted

ARFis unrestricted.

Only

inthis latter circumstance can the ratio

kD/kifl

beregarded

un-equivocally as a

sieving

coefficient for the

injured

tubule. That inulinindeedpassesthrough

the injured

tubularbarrierin

un-restricted

fashion

issuggested bythe

identity

of its

urinary

clear-ance with that ofDTPA. DTPA has a molecular

weight

of 393 daltons andamolecular radius of -4 A. Inulin is heavier by tenfoldand

substantially

larger

(radius

= 15

A).

Since both solutes are

freely

filtered across the

glomerular capillary

wall (24),the

equality

of their

urinary

clearances inthe present

study

suggests thattheir

respective

ratesof backleakacrossthe

_injured

tubular wallare notdifferent. This

implies

that in ARF the

reasonably accurate description ofthe size-selective

properties

of the tubular epithelial barrier. From the relationshipbetween the ratio

_kD/1k,

and dextran radius we infer that the

injured

tubular wall behaves as a membrane with a rather sharp

cut-off,

displaying

almostcomplete impermeabilitytodextransof radius above 34 A. This membranesize-selectivity is inkeeping with earlierobservationsin

postischemic

rodent ARF in which neither horseradish peroxidase (r=30A)noralbumin (r = 36A) leaked backacrossthedamaged tubular wall (6,3).

Thestructure intheinjuredtubular wall mostlikely to behave as aneffective barriertothe passage of macromolecules may be thetubular basement membrane.Thiscollagenousstructurehas a

composition

notunlikethat of theglomerularbasement

mem-brane.

Thus,

thesimilaritybetween thepermselective properties ofthedamaged tubulein ARF and that of the glomerular

cap-illary

wall

(Fig.

5) could be construedaspointingto the tubular basement membrane asthe structure in protracted ARF that

ultimately

prevents the egress ofdextrans of 34 A radius and

larger

from tubule fluid into the interstitium.The microdissection studiesbyOliver and his coworkers(25) of individual

nephrons

from

patients

with protractedARF are inkeepingwith this

pos-sibility. They

observeda

patchy

detachment of necrotic tubular

epithelial

cellsfrom the

underlying

tubular basementmembrane, a

phenomenon

for which they coined the term tubulorrhexis. The denuded basement membrane remained intact, however,

forming

acontinuous layer along the length ofthenecrotic tu-bule.

Recovery offunction after ARF. In the rat, recoveryfrom

postischemic

ARFhasbeenpresaged by increasing renalblood flow

(8). Presumably

restoration of their blood supply isnecessary before necrotic tubular cellscanregenerate. Becausetheability of

proximal

tubular cells to secrete

p-amino-hippuric

acid is

profoundly impaired during

ARF

(2,

8, 9), the clearance of

p-amino-hippuric

acidcannotbe used to measure renal plasma flowinthis circumstance.Inthe absenceofanalternative, non-invasive method, we were unableto confirm the presence of renal

underperfusion

in thepresent study. That renal plasma flow inour

patients

with protractedARF flow wasseverely

de-pressed

nonetheless, is suggested by the observation ofextreme

hyperreninemia.

Similar elevations of active plasma reninlevels

by

20-or30-fold havebeenobserved by others during theoliguric

phase

ofARF.These levelshave tendedtodeclineas early

oli-goanuria

gives

waytothediuretic

phase (26,

27).It maybethat extremeelevations ofPRA(and hence the vasoconstrictor

an-giotensin

II) delay recoveryfrom ARF. The findings inthose

subjects

destinedtorecoverof larger urine flowrates and greater inulin clearances coupled with less hyperreninemia are

com-patible

withthisview.Inkeeping withthefindingthat maneuvers thatsuppress therenin-angiotensin axisdo not protect uniformly

against subsequent

ARF(28), hyperreninemiaduringARFmay beasecondary phenomenon,anditssuppression may have im-portance

only

in

facilitating

theonsetofrecovery.

We conclude that the brief and protracted forms of

ARF

describedinthepresent study represent a continuum of post-ischemic renal injury. Protracted ARF differs from the abbre-viatedinjuryin that the reductions in urine formation and inulin clearancearemore profound. Even when corrected for backleak, theGFR issubstantially lower in protracted than in brief

ARF,

10±3vs.28±4 ml/min per 1.73m2, respectively.Clearly, factors other thanbackleak mustbeinvoked to

_explain

the

_disparity.

More marked and diffuse

intraluminal

tubularobstruction in

protracted

ARF,

with consequent reduction ofthe

transmem-brane

hydraulic

pressure

gradient,

may beone such factor. A greaterdecline in

glomerular

perfusion

rateduetoextreme

sec-ondary

hyperangiotensinemia

may be another. We wishto

em-phasize,

however,

that there isatpresentno

_analogous

animal model of greatly

protracted

ARF after

prolonged

and

_partial

(and

frequently

repetitive)

ischemicrenal insults. Confirmation of the

findings

inferred fromour

_necessarily

indirect

_approach

in human

protracted

ARF must await the

development

ofa

corresponding

renal

injury

in

experimental

animals.Because of thelethal natureof

protracted

ARF in man, the

development

ofan

appropriate

animal model is

urgently

required.

Only

by

learning howtoattenuatethissevereformofARF in the labo-ratoryisaneffective

therapeutic

strategy for the human renal

injury

likely

to eventuate.

Acknowledaments

The authors thank Mary Petersonforably preparingthemanuscript. This studywassupported bygrant 5 ROI AM29985 from the Na-tionalInstitutesofHealth.Dr.Moran'sfellowshipwassupportedjointly by the Marilyn SimpsonTrust Foundation andbytraining grantAM 07357 from the National Institutes of Health.

References

I.Swann, R. C., andJ. P.Merrill.1953. The clinical course ofacute

renal failure. Medicine (Baltimore). 32:215-292.

2.Bull, G. M.,A.M.Joeckes, andK. G. Lowe. 1950. Renal function studies inacutetubular necrosis. Clin. Sci.9:379-404.

3.Tanner, G. A., K. L. Sloan, and S. Sophasan. 1973. Effects of renalarteryocclusiononkidneyfunction in the rat.KidneyInt. 4:377-389.

4.Arendhorst, W.J.,W. F.Finn,C. W.Gottschalk,and H. K. Lucas. 1976.Micropuncture studyof acute renal failurefollowing temporary renal ischemia in therat.KidneyInt.1O(Suppl.):SI00- 105.

5. Tanner, G.A., and S. Sophasan. 1976. Kidneypressures after temporary renal arteryocclusion inthe rat. Am. J.Physiol.

230:1173-1181.

6.Donohoe, J. F., M.A.Venkatachalam,D. B. Bernard, and N. G.

Levinsky. 1978.Tubularleakage and obstruction after renalischemia: structural-functionalcorrelations.KidneyInt. 13:208-222.

7.Eisenbach, E. M., andM.Steinhausen.1973. Micropuncture studies aftertemporaryischemia ofratkidneys.

Pjluegers

Arch. Eur. J.Physiol. 343:11-25.

8.Finn, W. F., and R. L. Chevalier.1979. Recoveryfrom postischemic

acuterenalfailure intherat.KidneyInt. 16:113-123.

9.Myers, B.D., C. Miller,J. Mehigan, C. Olcott, H. Golbetz,G. Derby,R. Spencer,andS. Friedman. 1984. Nature of the renalinjury

followingtotal renalischemiain man. J. Clin. Invest.73:329-341. 10.Myers,B. D., B. J.Carrie,R. R.Yee, M.Hilberman,andA.S. Michaels. 1980. Pathophysiology of hemodynamically mediated acute renalfailurein man.KidneyInt. 18:495-504.

11.Myers,B.D.,F.Chui, M. Hilberman, and A. S. Michaels. 1979. Transtubular leakage ofglomerular filtrate in human acute renalfailure. Am.J.Physiol.237:F319-F325.

12. Myers, B. D., M. Hilberman, B. J.Carrie, R. J.Spencer,and R. L.Jamison. 1982. Glomerular and tubular function in non-oliguric

acuterenalfailure.Am. J. Med. 72:642-649.

13.Hilberman,M., B. D. Myers, B. J. Carrie, G. Derby,R.L.Jamison, andE. B.Stinson. 1979. Acute renal failure following cardiacsurgery. J. Thorac. Cardiovasc.Surg.77:880-888.

14.Johnston,P. A., D. B. Bernard, J. F. Donohoe, N. S.Perrin,and N.G. Levinsky. 1979. Effect of volume expansion on hemodynamics of thehypoperfusedrat kidney. J. Clin. Invest. 64:550-558.

15.Anderson, R. J., and R. W. Schrier. 1980. Clinical spectrum of

oliguric and non-oliguric acute renal failure. In Acute Renal Failure. B. M. Brenner and J. H. Stein, editors. Churchill Livingstone, Inc., New York. 1-17.

16. Anderson, R. J., S. L. Linas, A. S.Bemns,W.L.Heinrich,T. R. Miller,P. A.Gabow, and R. W. Schrier. 1977. Non-oliguric acute renal failure.N.Engl.J.Med. 296:1134-1138.

17.Miller,T.R., R. J. Anderson, S. L.Linas,W. L.Heinrich, A. S. Berns, P. A. Gabow, and R. W.Schrier.1978.Urinarydiagnostic indices in acute renalfailure.Aprospectivestudy. Ann. Intern.Med.89:47-50. 18. Bryer-Ash, M., R. A. Ammon, and J. A. Luetscher. 1983. In-creasedinactive reninindiabetes mellitus without evidence of nephrop-athy.J.Clin.Endocrinol. Metab. 56:557-561.

19.Chang,R.L.S., C. R.Robertson,W. M.Deen, and B. M. Brenner. 1975.Permselectivity oftheglomerularcapillarywalltomacromolecules. I.Theoretical considerations. Biophys.J. 15:861-906.

20.Bojesen, E. 1954. The transportof urine in the upperurinary tract. ActaPhysiol. Scand. 32:39-62.

21. Oken,D. E. 1984. Hemodynamic basis forhumanacuterenal failure(vasomotornephropathy).Am.J.Med.76:702-7 10.

22.Hostetter,T.H.,H.G.Rennke,and B. M. Brenner. 1982. The

casefor intrarenalhypertensionin the initiation andprogressionof di-abetic and otherglomerulopathies.Am.J.Med. 72:375-380.

23.Deen, W. M.,C. R.Robertson, and B. M. Brenner. 1972. A

model ofglomerularultrafiltration in therat.Am.J.Physiol. 223:1178-1183.

24. Chervu, L. R., and M. D. Blaufox. 1982. Renal radiopharma-ceuticals-anupdate.Semin.Nucl.Med.XII(3):224-245.

25.Oliver, J.,M.MacDowell,and A.Tracy. 1951.Pathogenesisof

acuterenalfailure associated with traumatic and toxic injury. J. Clin. Invest.30:1305-1351.

26.Brown,J.J.,R. L. Gleadle, D. H. Lawson, A. F. Lever, A. L. Linton,R. F. Macadam,E.Prentice,J. I. S. Robertson, and M. Tree. 1970. Renin andacuterenalfailure: studies inman.Brit. Med. J. 1:253-258.

27.Tu,W. H. 1965. Plasma renin activity in acute tubular necrosis and other renal diseases associated with hypertension. Circulation. 31: 686-695.

28.Madias,N.E.,and J. T.Harrington. 1983. Postischemicacute