Thorax 1990;45:688-693

External detection of pulmonary accumulationof indium- 1 13m labelled transferrin in theguinea pig

Ursula Hultkvist-Bengtsson, Lena Martensson

AbstractAccumulation of radioisotope labelledtransferrin in the lungs of guineapigs was determined with an externaldetection system. The method is basedon the intravascular and extravasculardistribution of indium-113m labelledtransferrin compared with the intra-vascular distribution of technetium-99mlabelled red blood cells. Guinea pigswere given iloprost, a prostacyclinanalogue and potent pulmonary vaso-dilator, and noradrenaline, a pulmonaryvasoconstrictor, in an attempt to in-crease and decrease respectively theblood volume in the lungs. Neither agentaltered transferrin accumulation in thelung by comparison with a saline in-fusion. Iloprost infused before and afteroleic acid infusion reduced macro-molecular leakage when compared witholeic acid alone. These data suggest thatthe double isotope method can distin-guish between hydrostatic and injuryinduced pulmonary oedema.

Department ofZoophysiology,University ofLundU Hultkvist-BengtssonDepartment ofMedical Research,Gambro ABL MartenssonAddress for reprint requests:Dr Ursula Hultkvist-Bengtsson, Department ofMedical Research, GambroAB, Box 10101,S-22010 Lund, Sweden.

Accepted 25 April 1990

The basis of external radioflux detection ofisotope labelled plasma protein in thepulmonary circulation was described insheep.' Modifications of this method in largeexperimental animals have been reported"'but are not applicable to many laboratories.Clinical studies using this method have alsobeen reported.368 We have developed a

modification of the double isotope method foruse in the guinea pig.2 Two methodologicalquestions regarding this model are whetherchanges in lung blood pool volume affect thedetection of protein flux, and whether themethod can distinguish between protein poorand protein rich pulmonary oedema.

Procedures previously used to answer thesequestions,' 910 such as whole blood transfusionand left atrial balloon inflation, are notapplicable in the guinea pig because of its size.We therefore used pharmacological methods.Noradrenaline, a pulmonary vasoconstrictorthat causes pulmonary hypertension,"1increases pulmonary capillary pressure,'2and reduces pulmonary vascular volume bycausing predominant postcapillary vasocon-

striction,'135 was used to induce hydrostaticoedema. Iloprost, a prostacyclin analogue anda potent pulmonary vasodilator, was used toincrease lung blood volume." 16 The effect ofiloprost on pulmonary oedema induced byoleic acid was studied to show whether the

extemal detection of radioactivity could showchanges in plasma protein accumulation in thelung.

Materials and methodsThirty five male inbred albino guinea pigs(Dunkin-Hartley, conventionally bred) wereused (Sahlins Farm, Malmo, Sweden). Theyweighed 895 (SD 145) g and were kept under a12 hour daylight cycle, in 30-50% humidity, at20-220C. The animals were housed two to acage (19 x 34 x 56 cm), and they had freeaccess to food (EWOS, Sodertalje, Sweden)and drinking water.

EXPERIMENTAL PROCEDURESGuinea pigs were anaesthetised with ketamine(Ketalar 30 mg/kg body weight subcut-aneously) and xylazine (Rompun 3 mg/kgintramuscularly) and maintained on ketamine 4mg/ml Ringer solution infused at 3 0 ml/h. Asingle injection of xylazine (3 mg/kg intra-muscularly) was administered 2-5 hours afterthe start of anaesthesia. The animals breathedspontaneously during the experiments. Bodytemperature was kept at 36-38°C by means of arectal thermistor and a heating pad. Poly-ethylene catheters (diameter 0-96 mm outside,0-58 mnm inside) were introduced into the leftand right femoral arteries. A catheter wasplaced in the right internal jugular vein for theinfusion of drugs. Carotid artery bloodpressure was recorded with a Micro-Switch156 PC transducer (Honeywell, USA) zeroedto the heart level. When needed, mepivacain(Carbocain 5 mg/ml) was added drop by dropto prevent vascular spasm during catheterisa-tion. The catheters were flushed with heparin-ised saline (50 IE/ml saline).

RADIOACTIVITY MEASUREMENTSRed blood cells were labelled in vivo withtechnetium-99m (half life 6-0 hours) and trans-ferrin (molecular weight 76 000) was labelled invivo with indium-I 13m (half life 100 min). Anexternal dual probe scintillation detector sys-tem was used to sample the entire photonenergy spectrum for both isotopes contin-uously. One scintillation probe was placedvertically over the heart at midsternal level andthe other in the horizontal plane just caudal tothe right shoulder. Both probes werecollimated with lead (fig 1) and registeredactivity from two conically shaped volumes inthe chest. The collimators reduced radiationfrom surrounding tissues by 68-94%. Theefficiency of sampling decreased with increas-

688 on A

pril 15, 2022 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.9.688 on 1 Septem

ber 1990. Dow

nloaded from

External detection ofpulmonary accumulation of indium-i 13m labelled transferrin in the guinea pig

Lead

Stainless steel

LSII

Nal crystal

1T

, X1

- 14.5- --

'.5.54I- 26 -4

Figure 1 Structure of the collimator used in theexperiments (measurements in millimetres).

ing distance from the detector (>600% reduc-tion 2 0 cm into the thorax). The base of the"sampling cone" in the thorax was 15 mm indiameter 2 cm below the opening of thecollimator for the energy level of 99"Tc. Thecalculated cone volume was smaller with theenergy level of l"smIn. Thus the heart probemainly sampled activity from the heart and theother probe mainly the right lung with a verysmall contribution from the heart. The areasunder the four energy peaks were integratedand used for further calculations. Counts werecollected for three minute periods for 3-5hours.To calculate the contribution of is3mIn radia-

tion to the 99mTc peak, ll3mIn was injected 15minutes before 99"Tc. Thirty minutes wereallowed for distribution and equilibrium of theradionuclides after the 99mTc injection. Thelatter part of this period was used for baselinemeasurements.

PROTOCOLTreatment groups are given in table 1. Oleicacid (0 03 ml/kg/10 min intravenously; Sigmachemical, St Louis, USA) or saline infusionswere started at time zero. In the combinedexperiments iloprost (supplied by E Schillin-ger, Schering AG Research, Berlin) infusion(5 ng/min/kg) was started 60 minutes beforeoleic acid administration and lasted for a total

Table 1 Treatment of the five groups

No ofanimals Treatment Dose

6 Iloprost 40 ng/min/kg, 120 min6 Noradrenaline 1 pg/min/kg, 120 min10 Saline 0-03 ml/kg, 10 min7 Oleic acid 0-03 ml/kg, 10 min6 Iloprost + 5 ng/min/kg, 180 min +

oleic acid 0-03 ml/kg, 10 min

of 180 minutes with a break for the oleic acidinfusion. In the single agent experimentsnoradrenaline (1 1ug/min/kg) or iloprost (40 ng/min/kg) infusions were started at time zero andcontinued for 120 minutes. The drugs werediluted in saline and given to the animals in atotal volume of 4 ml.At the end of the experiment, the animals

were killed with 1 ml of pentobarbital sodium(60 mg/ml). The lungs were dissected andweighed immediately. Pieces of lung for wet-dry weight analysis were dried to a constantweight at 80'C. Lung water was expressed as apercentage of wet weight (wet weight -dryweight)/wet weight x 100.17

BLOOD ANALYSESBlood cell counts were performed on 200 plwhole blood taken from the femoral artery intoEDTA microtainer tubes at 0, 15, 30, 60, 90,and 120 minutes. White blood cell and plateletcounts were determined in a Twincounter 187(Analys Instrument, Stockholm) and expres-sed as percentages of the initial values. Dif-ferential white blood cell counts were perfor-med on blood films stained with Giemsa-May-Gruinewald at Medi-Lab, Stockholm. Fordetermination of arterial oxygen tension (Pao2)400 pul whole blood was taken anaerobicallyfrom the femoral artery in heparinised syringesat 0, 15, 30, 60, 90, and 120 minutes andanalysed with an OSM Hemoximeter(Radiometer, Copenhagen) and an ABL 3Bloodgas Analyser set at 38°C. The resultswere expressed as percentages of the initialvalues.As a measure of free oxygen radical activity,

plasma concentrations of uric acid weremeasured spectrophotometrically (AbbottBiochromatic Analyser 100; A-GENTTM UricAcid, Abbott Diagnostica Division). Wholeblood (500 p1) was collected in 50 pl 388%citrate buffer at 0, 15, 30, 45, 60, 90, and 120minutes in the group of animals having oleicacid and iloprost. The samples wereimmediately centrifuged and the plasma wasfrozen for subsequent analyses. The red bloodcells of the same samples were washed threetimes with saline and kept at 4'C for up to 72hours before determination of catalaseactivity.'8 Catalase activity was measured in redcell lysates by observing the decrease in absor-bance at 240 nm of 30 mM H202. The decreaseof absorbance per time unit is a direct measureof the catalase activity, which is expressed asthe rate constant/g haemoglobin (K/g Hb).

DATA HANDLINGAll counts from 99mTc and lls3mIn were correctedfor background radiation, physical decay, andchannel overlap. Loss of activity through bloodsampling was considered to be almost constant.The integrals of the four energy peaks werecalculated by the multichannel analyser. Eachdata point was then calculated with the use ofthese integrals as follows:

ismlIn lung cpm/99mTc lung cpm

ii3mIn labelled heart cpm/99"Tc heart cpmThe activity ratios were plotted against time

689

.................................................................................................................................................................................................................................................................................................................

on April 15, 2022 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.45.9.688 on 1 S

eptember 1990. D

ownloaded from

Hultkvist-Bengtsson, Mdrtensson

Table 2 Baseline data (mean (SD) values) on the anaesthetised guinea pigs before the infusions

Group Saline Oleic acid Noradrenaline Iloprost Iloprost + oleic acid

White blood cells (109/l) 7-1 (1-0) 5 7 (3 0) 7-6 (4-1) 8-6 (1-3) 11 0 (0 9)Neutrophils (109/1) 3-2 (0 9) 2-6 (2 3) - - 6-6 (1-5)Platelets (109/l) 268 (86) 315 (67) 273 (25) 310 (63) 321 (49)Arterial oxygen tension (mm Hg) 61 (13) 74 (16) 67 (13) 67 (12) 65 (4)Mean arterial blood pressure(mm Hg) 54 (5) 57 (7) 52 (3) 58 (7) 62 (9)

Heart rate (beats/min) 177 (27) 181 (26) 179 (16) 206 (29) 180 (12)Plasma uric acid (jumol/l) 32 (18) 37 (10) 56 (23) 59 (32) 64 (25)Red blood cell catalase(K/g Hb) 180 (44) - - - 292 (259)

Conversion to SI units: 1 mm Hg t 0-133 kPa.

and by linear regression (method of leastsquares) of the steepest part of the curve theregression line slope (RLS) and the intercept ofthe regression line were derived. To avoidartefacts from uneven mixing of activity orfailure to reach equilibrium between intra-vascular and extravascular spaces the first 30minutes were excluded from this analysis.

STATISTICSMean values and standard deviations are given.Statistical evaluation was performed with the Ftest for equality of variance, Student's t test,and one way analysis of variance (ANOVA).Probability values below 0 05 were consideredstatistically significant.

ResultsBaseline haematological, biochemical andhaemodynamic measurements are shown intable 2.

EFFECT OF NORADRENALINE AND ILOPROSTThere were no significant differences in trans-ferrin accumulation (expressed as RLS x

* SALINEA ILOPROST* NORADRENALINE

60

TIME (min)

Figure 2 White blood cells (WBC), as percentages of initial values, in saline(n = 10), noradrenaline (n = 6; 1 pg/min/kg), or iloprost (n = 6; 40 ng/min/kg)treated guinea pigs: mean values versus time. Noradrenaline versus saline, iloprostversus saline: *p < 0O05, **p < 0 01, ***p < 0 001 (Student's t test).

10-3 min-') in the lung after infusion ofnoradrenaline (mean RLS 0.52) or iloprost(1 69) and after saline infusion (1 -01) (table 3).Mean arterial blood pressure increased to a

maximum value that was 190% (SD 8%) oftheinitial value four minutes after the start of thenoradrenaline infusion and fell to 49O% (9%)within two minutes of the end of the infusion.Iloprost caused a substantial decrease in meanarterial blood pressure, which reached 64%(10%) of the initial value six minutes after thestart of the infusion, remained unchangedthroughout the rest of the experiment, and roseto 81% (15%) of the initial value after theinfusion had stopped. Saline infusion did notchange blood pressure. Heart rate did not altersignificantly with any infusion, though it wassomewhat higher in the noradrenaline group at15 minutes (131O% (17%) of baseline).Noradrenaline caused the white blood cell

count to increase by 82% (13%) at 15 minutesand it remained raised until the end of theinfusion (fig 2, table 4), returning to near basalvalues (120% (25%)) 30 minutes after theinfusion stopped. The white cell count did notchange with the iloprost infusion. The plateletcount did not change significantly in anygroup. There was no change in Pao2 in anygroup. Lung water measurements after theinfusion showed no difference betweennoradrenaline (81 3% (0 7%)), iloprost(81 2% (1-5% )), and saline (79-9o (144%)).

EFFECT OF OLEIC ACID WITH AND WITHOUTILOPROSTWhen iloprost was given with oleic acid trans-ferrin accumulation (expressed as RLSx 10-3min-') in the lung was less (RLS 1-59)than after oleic acid alone (3-86; p = 0-026):see table 3. The first 60 minutes of infusionwere used to control for the effects of iloprostalone. The changes in transferrin accumulationdid not differ from those seen with saline.During the oleic acid infusion period wheniloprost was stopped temporarily transferrinaccumulation increased significantly more(RLS 13-0) than in the group receiving oleicacid alone (3 86; p = 0-011).The changes in blood cell counts and Pao2

did not differ significantly between the animalsreceiving iloprost and oleic acid and the grouphaving oleic acid only except that the maximumfall in white blood cell count seen at 15 minuteswas greater in the iloprost plus oleic acid group(58% (30%) of the initial value) than in the

LU

-J

-1

z

0

0

690 on A

pril 15, 2022 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.9.688 on 1 Septem

ber 1990. Dow

nloaded from

External detection ofpulmonary accumulation of indium-I 13m labelled transferrin in the guinea pig

Table 3 Regression line slope (RLS) valuesfor transferrin accumulation (x l0- min-')from individual animals in thefive treatment groups

Saline Oleic acid Noradrenaline Iloprost Iloprost +oleic acid

1-20 5 59 -0-42 0-81 0-96-1-34 3 00 -0 37 1-49 1-932 40 3-57 1-62 2-73 -0-241-14 4-38 0-25 1-45 0-221 99 3-85 1-21 1-25 5-481-33 2-20 0-86 2-39 1-20

-0-48 4-43 - - -

0-49 -1-761-63 - - - -

Mean (SD)1 01 3-86 0-52 1-69 1-59

(1-16) (1 09) (0 84) (0-73) (2 05)

group receiving oleic acid only (78% (23%)).Neutrophil counts were reduced significantlyin the group receiving iloprost (fig 3, table 4).Both groups receiving the oleic acid infusionhad a greater fall in Pao2 at 15 minutes than theanimals receiving saline (p < 0-001, table 4).Mean arterial blood pressure did not changewith saline but it fell by about 300o in theiloprost group and by 400,, after oleic acid alonewithin 20 minutes of the start of the infusion; itthen recovered slowly during the rest of theexperiment. Heart rate was unchanged in allgroups. There were no differences in lungwater between the two oleic acid groups (oleicacid only 81 90% (2 3%), iloprost plus oleic acid81 9% (0-9%)).The plasma uric acid concentration did not

differ significantly between the animals receiv-ing oleic acid plus iloprost and those havingsaline but it was substantially higher in thegroup having oleic acid alone (table 4); thisincrease was significantly greater than thechange after saline at 15 minutes (p < 0-01).

Red blood cell catalase activity fell in theiloprost group, the lowest value (58 2%(29*3%) of the initial value) being seen at 120minutes.

DiscussionThe absence of changes in transferrinaccumulation in the lung after an infusion ofnoradrenaline or iloprost suggests that thisindex of permeability corrects for blood poolchanges in the lungs. Noradrenaline was

infused slowly to prevent endothelial celldamage from pressure effects. The unchangedheart rate indicates that little, if any, of theaction of noradrenaline and iloprost is due tosystemic effects. The high mean arterial bloodpressure at the start of the noradrenalineinfusion is likely to have led to compensatorymechanisms or tachyphylaxis as it returnedtowards baseline during the experiment despitethe continued infusion. The low arterial pres-

sure seen with iloprost stayed at the nadir levelduring the experiment, returning to baselineonly after the end of the infusion. The infusionof iloprost alone to the guinea pigs in one ofour groups served as an additional control forthe stability of the transferrin accumulationmeasurements in the initial part of the study inthe group receiving iloprost plus oleic acid,though the doses given to the two groups were

different.The increased pressure in the pulmonary

circulation that would be expected from thenoradrenaline infusion" 12 could have caused a

protein poor hydrostatic oedema undetectableby the scintillation probes. We were unable tomeasure pulmonary wedge pressure to confirmthe increased vascular pressure in the lung afternoradrenaline infusion directly, but the trend

Table 4 Mean (SD) values for white blood cells, neutrophils, arterial oxygen tension, and plasma uric acid concentration as

percentages of initial values in the five treatment groups

Time (min) 0 15 30 45 60 90 120

White blood cells (% initial value)Saline 100 103 110 - 122 116 120

(10) (14) (18) (21) (20)Iloprost 100 122 130 130 126 127

(I11) (16) (18) (18) (18)Noradrenaline 100 182 188 186 183 195

(13) (24) (39) (45) (56)Neutrophils (% initial value)Saline 100 103 98 122 129 116

(30) (17) (25) (30) (54)Oleic acid 100 79 113 - 118 132 156

(28) (32) (41) (28) (50)Iloprost + oleic acid 100 42 64 73 58 88

(44) (29) (45) (23) (34)Arterial oxygen tension (% initial value)Saline 100 101 110 105 112 129

(13) (15) (18) (19) (26)Oleic acid 100 65*** 72** 71** 78* 90*

(19) (24) (26) (30) (33)Iloprost + oleic acid 100 62*** 91 - 89 90 83**

(19) (32) (23) (35) (25)Uric acid (% initial value)Saline 100 104 95 118 97 92 107

(16) (25) (30) (23) (23) (34)Oleic acid 100 151** 159 190 154** 161** 150**

(45) (36) (68) (99) (89) (108)Iloprost + oleic acid 100 116** 138* 123 109 117 104*

(41) (60) (52) (36) (40) (36)

p < 0-05, **p < 0-01, ***p < 0 001 in the comparison of each treatment with the saline control.

691 on A

pril 15, 2022 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.9.688 on 1 Septem

ber 1990. Dow

nloaded from

Hultkvist-Bengtsson, Mdrtensson

Lu

z

n

z

TIME (min)

Figure 3 Neutrophils as percentages of initial values in saline (n = 10), oleic acid;(0-03 ml/kg/10 min); (n = 7), or iloprost (5 ng/minfkg) + oleic acid (0-03 ml/kglO0min) (n = 6) treatedguinea pigs: mean values versus time. Saline versus oleic acid,saline versus iloprost + oleic acid: *p < 0-05, **p < 0O01 (Student's t test).

towards increased lung water content supportsthe presence ofhydrostatic pulmonary oedema.This suggests that the permeability index we

used is not influenced by changes in pulmonaryblood volume or increased pulmonarymicrovascular pressure. These results are inaccordance with reports that such externalradioisotope measurements are specific forpulmonary oedema with increased proteinpermeability.4 Gorin et al' showed a closecorrelation between lymph protein concentra-tion measured directly and external radiofluxdetection in sheep after bacteraemia and Suger-man et al,'9 using computerised gamma scin-tigraphy, showed that the lung:heart ratio of99MTc labelled human serum albumin was

significantly greater after oleic acid administra-tion to dogs than the lung:heart ratio of `nTclabelled red blood cells. These studies confirmthat external detection of radiolabelled proteindetects lung injury and not oleic acid inducedchange in pulmonary vascular volume. Thefindings in these animal studies have beensupported by clinical studies in which patientswith adult respiratory distress syndrome were

compared with patients with renal or car-

diogenic pulmonary oedema.38 The results ofthese studies support the specificity of thismethod as shown in our small animal model.The reduction in oleic acid induced

macromolecular permeability seen when oleicacid was preceded by iloprost infusion con-

firms reports of studies using other meth-ods.'6202' The transferrin accumulation ratewas linear throughout the control experiments,and the stability of the transferrin accumula-tion after the infusion of iloprost alone showsthat the decreased transferrin accumulationafter oleic acid when given with iloprost is notartefactual. The favourable effects of iloprostafter oleic acid injury may have been produced

indirectly by the decreased pressure in thepulmonary circulation rather than from a directeffect on the pulmonary microvascular mem-brane, though prostacyclin may also be impor-tant in preserving the cellular integrity of thelungs.2022 An explanation of why the Pao2 andlung water did not return towards controlvalues despite a decrease in transferrinaccumulation could be that the permeabilityindex is a more sensitive indicator of changesin pulmonary extravasation than traditionalmethods of assessment, such as wet-dryweight, and that it detects a change inextravasation earlier.4The failure of iloprost to prevent the leuco-

penia and thrombocytopenia induced by oleicacid infusion is in keeping with the results ofstudies in which prostacyclin has prevented orattenuated oedema formation.2" The protectiveeffects of iloprost on platelet function areprobably concentration and species depen-dent.2324 Prostacyclin also counteracts oxygenfree radical activities.25 We found diminishedred blood cell catalase activity and uric acidproduction after iloprost was added to oleicacid treatment. Winterbourn et al6 showedthat red blood cells are an important compon-ent of the extracellular antioxidant defencesystem. The inhibition of uric acid productionseen with iloprost might indicate a beneficialaction of iloprost on oxygen free radicalmediated injury and suggests a link betweenfunctional inhibition of neutrophils anddecreased uric acid production. The neutrophilcount in our study fell to a greater extent afteriloprost than with oleic acid alone. Reducedcirculating neutrophils are considered toindicate possible sequestration in the lung inmany shock states.27 Despite this, the rate ofmacromolecular extravasation was decreased inour study, which might indicate inactivation ofthe sequestered neutrophils by iloprost.

External detection of the accumulation ofradiolabelled protein in the lung is a reliablemeans of separating protein poor oedemaresulting from increased pulmonary pressureand protein rich oedema resulting from lunginjury. The method of calculating the regres-sion line slope corrected for changes in thevolume of the lung blood pool. Iloprost atten-uated oleic acid induced pulmonary oedemapossibly through the inactivation of seques-tered neutrophils. This model in the guinea pigmay be of value for studying treatments thatinhibit development of protein rich pulmonaryoedema.

We express our gratitude to Ewa Dahlberg, Christina Johans-son, and Anna-Maria Larsson for their excellent technicalassistance. This study was supported by Gambro AB, theHierta-Retzius Foundation, and the Crafoordska Foundation.

1 Gorin AB, Weidner WJ, Demling RH, Staub NC. Non-invasive measurement ofpulmonary transvascular proteinflux in sheep. J Appl Physiol: Respirat Environ ExercisePhysiol 1978;45:225-33.

2 Hultkvist U, Bjellin L, Martensson L. External "'Ine and"Tcm radiation detection of lung edema induced by oleicacid in the guinea-pig. Res Exp Med 1988;188:277-88.

3 Byrne K, Sugerman HJ. Experimental and clinical assess-ment of lung injury by measurement ofextravascular lungwater and transcapillary protein flux in ARDS: A reviewof current techniques. J Surg Res 1988;44:185-203.

4 Dauber IM, Pluss VT, van Grondelle A, Trow RS, Weil

692 on A

pril 15, 2022 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.9.688 on 1 Septem

ber 1990. Dow

nloaded from

External detection ofpulmonary accumulation of indium-113m labelled transferrin in the guinea pig

JV. Specificity and sensitivity of noninvasivemeasurement of pulmonary vascular protein leak. J ApplPhysiol 1985;59:564-74.

5 Sugerman HJ, Strash AM, Hirsch JI, et al. Sensitivity ofscintigraphy for detection ofpulmonary capillary albuminleak in canine oleic acid ARDS. J Trauma 1981;21:520-7.

6 Basran GS, Byrne AJ, Hardy JG. A noninvasive techniquefor monitoring lung vascular permeability in man. NuciMed Commun 1985;3:3-10.

7 Mishkin FS, Niden A, Kumar A, Thomas A, Reese IC,Vasinrapee P. Albumin lung/heart ratio change. A simpleclinical means of documenting increased pulmonaryendothelial permeability to protein. JAMA 1987;257:953-6.

8 Rocker GM, Morgan AG, Pearson D, Basran GS, Shale DJ.Pulmonary vascular permeability to transferrin in thepulmonary oedema of renal failure. Thorax 1987;42:620-3.

9 Welsh CH, Dauber IM, Weil JV. Endotoxin increasespulmonary vascular protein permeability in the dog.J Appi Physiol 1986;61:1395-402.

10 Prichard JS, Rajagopalan B, Lee G de J. Transvascularalbumin flux and the interstitial water volume inexperimental pulmonary oedema in dogs. Clin Sci1980;59: 105-13.

11 Reeves JT, Grover RF, McMurtry I, Weir EK. Pulmonaryvascular reactivity. Bull Eur Physiopathol Respir1985;21:583-90.

12 Malik AB, Selig WM, Burhop KE. Cellular and humoralmediators of pulmonary edema. Lung 1985;163:193-219.

13 Rippe B, Allison RC, Parker JC, Taylor AE. Effects ofhistamine, serotonin, and norepinephrine on circulation ofdog lungs. J Appl Physiol: Respirat Environ ExercisePhysiol 1984;57:223-32.

14 Minnear FL, Malik AB. Mechanisms of neurogenicpulmonary edema. Ann N Y Acad Sci 1982;384:169-90.

15 Minnear FL, Barie PS, Malik AB. Effects of epinephrineand norepinephrine infusion on lung fluid balance insheep. J Appl Physiol: Respirat Environ Exercise Physiol1981;50:1353-7.

16 Malik AB, Perlman MB, Cooper JA, Noonan T, Bizios R.Pulmonary microvascular effects of arachidonic acidmetabolites and their role in lung vascular injury. FedProc1985;44:36-42.

17 Hambrecht GS, Hilton JG. A modified technique forpermeability studies in the superfused hamster cheekpouch. Life Sci 1984;34:971-6.

18 Aebi H. Catalase. In: Bergmeyer HU, ed. Methods ofenzymatic analysis 2nd ed. New York: Academic Press,1974:673-84.

19 Sugerman HJ, Hirsch JI, Tatum JL, Strash AM, Sharp DE,Greenfield LJ. Comparative scintigraphy in oleic acidpulmonary microvascular injury. Crit Care Med 1982;10:31-3.

20 Ikeda T, Hirose T, Aoki E, Ishibashi M, Domae M, TanakaK. A pathophysiological role of endogenous prostacyclinin endotoxin induced increase in lung vascularpermeability in dogs. Prostaglandins, Leukotrienes andMedicine 1983;12:385-97.

21 Hirose T, Aoki E, Ishibashi M, Domae M, Ikeda T,Tanaka K. The effect of prostacyclin on increasedhydraulic conductivity of pulmonary exchange vesselsfollowing microembolization in dogs. Microvasc Res 1983;26:193-204.

22 Sugerman HJ, Blocher CR, Hirsch JI, Strash AM,Tatum JL. Failure of various agents to decrease oleic acidpulmonary albumin leak. J Surg Res 1983;34:456-62.

23 Addonizio Jr VP, Fisher CA, Jenkin BK, Strauss JF, MusialJF, Edmunds LH. Iloprost (ZK36374), a stable analogueof prostacyclin, preserves platelets during simulatedextracorporeal circulation. J Thorac Cardiovasc Surg1985;89:926-33.

24 Smith III EF, Tempel GE, Wise WC, Halushka PV, CookJA. Experimental endotoxemia in the rat: Efficacy ofprostacyclin or the prostacyclin analog iloprost. CircShock 1985;16:1-7.

25 Fantone JC, Kinnes DA. Prostaglandin E, and prosta-glandin I modulation ofsuperoxide production by humanneutrophils. Biochem Biophys Res Commun 1983;113:506-12.

26 Winterboum CC, Stem A. Human red cells scavengeextracellular hydrogen peroxide and inhibit formation ofhypochlorous acid and hydroxyl radical. J Clin Invest1987;80:1486-91.

27 Modig J, Hillgren R. Pathophysiologic significance of lunggranulocytes in human adult respiratory distresssyndrome induced by septic or traumatic shock. Acta ChirScand 1987;153:267-71.

693 on A

pril 15, 2022 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.45.9.688 on 1 Septem

ber 1990. Dow

nloaded from