Gut 1995; 37: 791-797

Gut intramucosal pH and intraluminal Po2 in a

porcine model of peritonitis or haemorrhage

J B Antonsson, U H Haglund

AbstractThe tonometric method of detectingdecreased gut itramucosalpH (pH) is basedon the fact that carbon dioxide can diffusethrough the wall of the silastic balloon ofthe tonometer. By using deoxified salineand measuring Po2 as well as Pco2 thisstudy aimed to follow changes in mucosalP02 and relate them to changes in pHi inperitonitis versus haemorrhage. Twentyfive pigs were used. Five were controls, in10 peritonitis was induced by the instilla-tion offaeces in the abdominal cavity, and10 were bled, half ofthem stepwise duringthree hours, and half of them rapidlydown to a mean (SEM) arterial pressureof 30 (10) mm Hg. The drop in pHi corre-lated well with decreasing intraluminalPo2 (r=0.63 (0-13)) in haemorrhage. Inperitonitis this drop occurred within avery limited change in intraluminal Po2(r= 0.06 (0.17)). Thus oxygen seemed to bepresent in the mucosa at the same time asthere were signs of anaerobic metabolismas evidenced by a low intramucosal pH.Impaired oxygen extraction or utilisation,or both, is proposed as an explanation tothis seemingly paradoxical situation.(Gut 1995; 37: 791-797)

Keywords: oxygen consumption, oxygen delivery,sepsis, small bowel, tonometry.

Department ofSurgery, UniversityHospital, Uppsala,SwedenJ B AntonssonU H Haglund

Correspondence to:Dr J Antonsson, Departmentof Surgery, UniversityHospital, S-751 85 Uppsala,Sweden.

Accepted for publication21 April 1995

As the gut is an organ that suffers from earlyischaemia during shock and as the ischaemicgut can act as a 'motor of shock' and thusaggravate the shock state, it has been shown to beof benefit to monitor gut ischaemia in experi-mental animals and patients in shock.1 2 Thiscan be accomplished by the use of a tonometer,with which intramucosal pH (pHi) can be cal-culated. Low pHi is a sign of ischaemia, asanaerobic metabolism results in an accumula-tion of acid metabolites with increasing tissueacidity as a consequence. Ischaemia is mainlyconsidered to be caused by an inadequateoxygen delivery to the gut, but earlier experi-ments show that in septic states other factorsplay an important part.3 Rasmussen andHaglund have shown that in the early stages ofseptic shock there is a drop in pHi when oxygendelivery is still unaffected or even increased.4

It might seem more obvious to monitorgastrointestinal tissue oxygen tension directly,but technical difficulties have limited the use

of methods for direct measurements inclinical practice.5 6 In 1964 Bergofsky couldestimate Pco2 and P02 in urinary and gallbladders by measuring Pco2 and Po2 in saline

that had been instilled in them. He could, bythis method detect differences in P02 causedby different concentrations of 02 in airinhaled.7 Dawson et al later used this tech-nique in a loop of the small bowel.8 Bymeasuring Po2 intraluminally in a hollowviscus they could thus detect changes in tissueoxygen tensions. Using a gas permeablesilastic tube inserted subcutaneously andflushing it slowly with anoxic saline to deter-mine P02 of the perfusate Niinikoski andHunt could estimate subcutaneous tissuePo2.9 The tonometric method of determininggut intramucosal pH is a combination of thesetechniques. Here CO2 of gut mucosa andlumen diffuses through the wall of a perme-able balloon to its content of saline. 10The aim of this study was to monitor

intramucosal pH in experimental animals(pigs) during peritonitis or haemorrhage. Byusing deoxified saline in the tonometer andmeasuring Po2 of the fluid as well, intralumi-nal Po2 could also be monitored. This willallow the relation between pHi and intralu-minal Po2 to be studied and to find out ifthe relations differ in peritonitis versushaemorrhage.

Methods

Medication and instrumentationAfter 12 hours of starvation but with full accessto water, 25 Landrace pigs of both sexes with amean weight of 19.1 kg (range 15-22), werepremedicated with ketamine (Ketalar, Parke-Davis, NJ), 7 mg/kg. They were then anaes-thetised with an intravenous injection ofpentobarbital, 20 mg/kg. The animals wereintubated through a tracheostomy and con-nected to a respirator (Servo VentilatorSystem, Siemens-Elema, Sweden) and venti-lated with a mixture of oxygen and room air togive an 02 concentration of40% in air inhaled.Respiration was adjusted to give an arterialPco2 of40 (5) mm Hg. Anaesthesia was main-tained with a continuous infusion of pentobar-bital, 18 mg/h; this drug was given in 2.5%glucose (Rehydrex, Pharmacia, Uppsala,Sweden) at a rate of 15 ml/kg/h. Furthermore,an infusion of Ringer's acetate, 30 ml/kg/h, wasgiven from the induction of anaesthesia. Thus,the total amount of fluid infusion was 45ml/kg/h.

Bilateral neck cutdowns were performed: apolyethylene catheter was positioned throughthe carotid artery in the thoraric aorta and con-nected to a pressure transducer (PDCA 75,Druck Ltd, Groby, Leicestershire, UK) forrecording of arterial pressure. One cephalic

791

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

Antonsson, Haglund

vein was used for infusion and the othercannulated for injection of saline for determi-nation of cardiac output; the thermodilutionmethod was used and the results are thecalculated means from three consecutivemeasurements (Oximetrix 3 computer, AbbottLaboratories, North Chicago, IL). A Swan-Ganz catheter was positioned in the pulmonaryartery through a jugular vein and connected tothe pressure transducer, its position beingsecured by the typical pressure pattern.Through a skin incision the left femoral

artery was cannulated for blood sampling andbleeding.A midline laparotomy was performed. The

superior mesenteric artery was identified closeto its origin from the aorta and was cleanedfrom the surrounding tissue. An ultrasonicflow probe (Transsonic Systems, Ithaca, NY)was placed around the vessel and connectedto a blood flow meter (T 201, TranssonicSystems, Ithaca, NY). A vein in the mesen-terium of jejunum was cannulated and acatheter was positioned with its tip in thesuperior mesenteric vein. A small anti-mesenteric incision was made in the mid-ileumand a tonometer (Tonomitor, TonometricsInc, Bethesda, MD) was placed in the lumenof the gut and secured with a purse stringsuture. In 10 animals autologous faeces weresecured from a colostomy, which thereafterwas closed with a purse string suture. Thelaparotomy was then closed, and the animalswere allowed to rest for a period of onehour.

Figure 1: Mean (SEM) values ofmean arterial pressure(MAP - open circles andfull line) and cardiac index (CI -filled circles and broken line). For description ofgroups, seetext. * Values significantly (p< 05) differentfrombaseline: tsignificant (p<0 05) difference from controls.

Experimental designThe animals were allocated into five groups(n= 5 in all groups). One group served ascontrols (group C). In 10 animals the laparo-tomy was reopened after the resting period andperitonitis was induced by the instillation ofautologous faeces immersed in 50 ml of salineinto the abdominal cavity, whereafter thelaparotomy wound was closed. In five of these1 gxkg bw of faeces was used; in five 2 gxkgbodyweight. All these animals were infusedwith 45 ml/kg/h of fluid, as described above,from the start of anaesthesia till the end of theexperiment. In group B I blood was with-drawn: 25% of the estimated blood volume attime=0, 15% at t=l hour, 5%/o at t=2 hours,and 50/o at t=3 hours. Blood volume was cal-culated on body weight." From t=0 theseanimals were given only 15 ml/kg/h of 2.5%glucose with pentobarbital, no Ringer's acetatewas given. Group B II was bled over a period ofhalf an hour down to an arterial pressure of 30mm Hg. If needed, additional bleedings wereperformed to keep the arterial pressure at30± 10 mm Hg. The same fluid regimen as ingroup B I was used. These two differentmodels of haemorrhage were chosen to ensureone group with milder response (B I) and onewith severe response (B II) within a limitedtime scope.

All animals were followed up for five hourswith hourly haemodynamic measurements andsampling from arterial, central mixed venous,and mesenteric venous blood. The sampleswere analysed for blood gases (ABL 300,

120

10003)I 80EE 60a-< 402 20

0

c

**

I I0 1 2 3

lime (h)4 5

220

165 mC

E55 5j

0

P 1

*

*t *t

I I I0 1 2 3

Time (h)4 5

B I

T *tltv*t

0 1 2 3

Time (h)4 5

220

165 _he

110 E

55 5

,)

I

EE

[-

0

220

165 .0'-c

110 E5-5

55 U)

E

E

0-

120

100

80

60

40

20

n-

-il

*t -~*

I I0 1 2 3 4 5

Time (h)

120-

100180

60

40.

2010

B II

'\\

*t*t

*t *t *t *t *t

0 1 2 3 4 5

Time (h)

0)

IEa-

EECL

120

100

80

60

40

20r

120

100

80

60

40

20

220

165 .'

110 Ef_E

55 5

0

220

165 .'._h

110 E_E

55 5

0J

792

*t *.~

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

IntramucosalpH and intraluminal Po2

Radiometer, Copenhagen, Denmark) andoxygen saturation and haemoglobin (OSM 3HemoximeterTM, Radiometer, Copenhagen,Denmark).

Gut intramucosalpH (pH) and gut intraluminalP02The tonometric method for calculation of gutintramucosal pH (pHi) has been describedbefore.10 12 The balloon of the tonometerplaced in the mid-ileum was filled with 2.5 mlof saline. After one hour of equilibrationthrough the wall of the balloon the saline wasaspirated and Pco2 and Po2 of the sampledetermined in the blood gas machine. Pco2from the tonometer and simultaneouslyobtained arterial [HCO3-] were used in theHenderson-Hasselbalch equation for calculationof pHi.The saline was flushed with nitrogen for at

least one hour before the start of the experi-ment. Po2 of the saline was tested on fourseparate occasions during these procedureswith three samples at each time: mean (SEM)Po2 was 173-9 (7.4) mm Hg before, and 23.4(1. 6) mm Hg after one hour of nitrogenflushing.

Figure 2: Mean (SEM) values ofgut oxygen delivery (gutDO2 - open squares andfull line) and gut oxygenconsumption (gut V02 -filled squares and broken line).For description ofgroups, see text. * Values significantly(p<O0OS) differentfrom baseline; tsignificant (p<005)difference from controls.

Calculation ofoxygen delivery (DO2) andconsumption (V02)Oxygen content (Co2) was determinedaccording to the equation: Co2=(Po2xO0003) + (HgbX So2X 1 *36), where Hgbis haemoglobin and So2 is oxygen saturation.D02=Cardiac indexXCaO2, where CaO2 isoxygen content of arterial blood.V02= Cardiac indexX (CaO2- C,02), whereCV02 is oxygen content of central mixedvenous blood.Gut D02=(SMA flow kg- 1)X CaO2.Gut V02=(SMA flow kg- 1) X (CaO2-CsmvO2),where CsmvO2 is oxygen content of blood fromthe superior mesenteric vein.

Statistical methodsData were analysed using a two way (group andtime) analysis of variance (ANOVA) forrepeated measures with Dunkan's new multiplerange test. Differences with respect to baselinewithin a group were analysed by using a one wayANOVA with Dunkan's new multiple rangetest. Differences with respect to controls atmatched time points were assessed by Mann-WhitneyU test. p Values <005 were consideredsignificant. All results are given as mean (SEM).

-0c)

4.03.5

Cm -

> g0X x 2.5' X 2.0o c

o E 1.5

E 1.00.5

P 1

*

0 1 2Time (h)

3 4

.0C)

Ei-.E

4.03.53.02.52.01.51.0

0.50.0

4.0

3.5CM3O> .0

*t -0 m 2.5'Xx20o ._o E 1.5(EE (¾ E1.0

0.50.0

5

c

,_----t-~ __-- - _---X-

0 1 2 3 4 5Time (h)

Pll*

X _ <*~~~~~~~

N--~~~~-- Il U- ---M*

0 1 2Time (h)

3 4 5

B I 4.0

3.5

C v> x -0' X 2.0O *-o E 1.5= E 1.00D-

0-5

0 1 2Time (h)

3 4 5

B II

t It *t

It0 1 2

Time (h)

4.0

3.5

0 ~:3.0>O B-0v a) 2.5C -v' X 2.0o .'c E 1.5

(

0.50.0 3 4 5

n-nU.W

nrnn 1^ ^

793

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

Antonsson, Haglund

TABLE I Changes in mean (SEM) packed cell volume(%) at baseline, three hours, and endpoint

Baseline Three hours Five hours

Controls 28-0 (1-4) 28-4 (1-2) 29-0 (1 1)P I 29-0 (0.5) 38-3 (2.2)*t 38-0 (2.5)*tP II 28-8 (1-3) 36.5 (2.5)*t 37-2 (2.2)*tB I 29-6 (1 0) 27-8 (1-6) 22-8 (2O0)*B II 28-4 (0.7) 21-2 (20)*t 20-5 (2.8)*t

*Values significantly (p<0 05) different from baseline; tvaluessignificantly (p<005) different from controls.

In establishing the relation between pHi andintraluminal P02 in peritonitis and haemor-rhage correlation coefficients were calculatedfor each animal separately, based on all sixmeasure points in every animal. The slopespresented in Fig 5 are the means of the correla-tion coefficients for animals with peritonitisand haemorrhage, respectively.13

EthicsThis study was approved by the InstitutionalReview Board for the Care ofAnimal Subjects,and the care and handling of the animals was inaccord with National Institutes of Healthguidelines for the use of experimental animals.

ResultsThe response to the introduction of faeces intothe abdominal cavity differed considerablyfrom animal to animal and was not correlatedto the amount used. Our aim was to have onegroup with mild and one with a more severeperitonitis, and this was not accomplished by

Figure 3: Mean (SEM) values ofgut intramucosalpH(pHi - open triangles andfull line) and changes inintraluminal Po2 (filled triangles and broken line). Fordescription ofgroups, see text. * Values significantly(p<OOS) differentfrom baseline; tsignificant (p<0 05)difference from controls.

increasing the amount of faeces. This can becaused by individual differences in theresponse to faecal challenge and by differencesin the amount of bacteria in each faecalsample. Therefore, the results from theanimals with peritonitis are presented in twogroups independently of the amount of faecesused. Those with the five lowest pHi - values atend point form a group with more 'severe'peritonitis (P I) whereas those remaining fivewith higher pHi at end point constitute thegroup with 'mild' peritonitis (P II).The packed cell volume remained unchanged

in the control group, whereas a haemoconcen-tration was noted in the septic groups (P I and PII), and a haemodilution in the haemorrhagicgroups (B I and B II) (Table I).Mean arterial pressure remained stable in

controls after an initial rise, probably becauseof the large amount of fluids given; in theexperimental groups there was a continuousdecrease in mean arterial pressure (meanarterial pressure in group B II of course had arapid drop as an arterial pressure of 30 (10)mm Hg was a guideline for the amount ofblood to be shed) (Fig 1). A progressivedecrease in cardiac index was noted in groupC, probably as a result of the cardiodepressingeffects of anaesthesia. The decrease in CI was,however, more profound in all experimentalgroups (Fig 1).

Figure 2 shows the changes in gut oxygendelivery (gut DO2) and consumption (gutV02). Peritonitis as well as haemorrhageresulted in a significant drop in gut DO2,whereas no changes in gut V02 were noted.

C~

I0.c.

7.4

7.3

7.2

7.1

7.0

6.9

c

0 1 2 3 4 5Time (h)

110

100

90

80

70

60

50

Ca1)

U)In

-00

0

P1

*t *t

0 1 2 3Time (h)

4 5

B I

1--,~ * * -

t

*t *t *tl

0 1 2 3 4 5Time (h)

110

100

90

80

70

60

50

110

100

90

80

70

60

50

C

a)InCo

-0

00~

a)Into.0

4-

00~

7.4

-, 7.3

C 7.2

= 7.1

*_ 7.0

6.96.8

7.4

7.2

I 7.1

._ 7.0

6.96.8

*t *t

0 1 2 3 4 5Time (h)

B 11

e* *t

*t -*tI

02u0 1 2 3 4 5Time (h)

Cu

Ia-

cl

7.4

7.3

7.2

7.1

7.0

6.9

6-8

110

100

90

80

70

60

50

C'a)n

Co4.00

0~[L

7.4-

en~ 7.3

CQ 7.2 F-

7.QI 7.0a6-6.9

6.8-

110a)

100 c._E90 U)

800

70 R

60 &5CL

Fi.>

__p_ _irn

794

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

IntramucosalpH and intraluminal P02

Both these parameters were unaffected incontrols.pHi did not change during the experiment in

controls. In peritonitis there was a significantdrop in both groups: from 7.35 to 7-08 in P Iand from 7-31 to 7.18 in P II. It should, how-ever, be noted that pHi at end point was usedto group these animals. There was a drop from7.32 to 7.02 in B I and from 7.32 to 7.09 in BII (Fig 3). A drop in pHi can result from eitherincreased gut lumen Pco2, or decreased[HCO3-], or both. Table II shows [HCO3-] atbaseline, three hours, and at end point. Fromthis it can be seen that [HCO3-] showed asignificant decrease from baseline only ingroup B II.

Intraluminal P02 did not change in controlsand peritonitis (C, P I, and P II). A progressivedrop was seen in B I. The rapid bleeding in BII resulted in a rapid drop of intraluminal Po2at t=one hour, followedby a slight decrease (Fig

Figure 4 illustrates thpHi and intraluminal Porhage, and peritonitis, r(represents an individua

C~

IQ

ac

Q.

C,,

C

I

aa

7.6

7.4

7.2

7.0

68

6.6

6.4-50

7.6

7.4

7.2

60 70

P02 (% '

P I a

7.0 H

6.8

6.6 H

6.4-50 60 70

Po2 (% '

TABLE II Changes in mean arterial bicarbonateconcentration (mmol/) at baseline, three hours, and endpoint

Baseline Three hours Five hours

Controls 26-8 28.9* 29.2*P I 25-4 24-2t 24.OtP II 25-9 24-5t 24OtB I 25-0 23.8t 23-7tBII 25-6 190*t 22-5*t

*Values significantly (p<005) different from baseline; tvaluessignificantly (p<005) different from controls.

calculations using each measure point in everyanimal a good correlation was noted in haemor-rhage (r=063 (0-13)) whereas there was nocorrelation at all in peritonitis (r= 006 (0O 17)).The difference between them was significant(p<0002) (Fig 5).

by a plateau and then Discussion3). From a theoretical point of view some objec-e correlation between tions can be raised against the tonometric2 in controls, haemor- method of measuring intraluminal Po2. Firstly,espectively. Each line when using the tonometric method subcuta-Ll animal. Based on neously, the silastic tube is in full contact with

the surrounding tissue,9 but the tonometer inc the small bowel may not necessarily be in full

contact with the mucosa, parts of it may be infree lumen. Secondly, CO2 diffuses easily

= through tissue planes and it has been shown thatintramucosal Pco2 is in equilibrium with that ofthe lumen.'4 Whether this is also the case withoxygen, has not been investigated. Thirdly, thetime to reach an equilibrium between the gutlumen and the contents of the tonometerballoon is not known and no correlation factor

I has been established. Finally, the implication of80 90 100 110 the counter-current exchange and its impacts)f baseline) on the diffusion of oxygen in the mucosal villi

has to be studied.'5 It could be argued that the3nd P II method in itself needs more investigation and

validation and we therefore present our data asper cent of baseline, as we are confident thatchanges in intraluminal P02 reflect changes inmucosal Po2. Probably it is not necessary todetoxify the saline in the tonometer as theoxygen most probably will equilibrate irrespec-tive of direction of flow. It should also beemphasised that this method is very sensitive to

I errors caused by negligent handling. Even a80 90 100 110 small amount of air, accidentally sucked into the)f baseline) syringe, may give falsely high values.

7.6B and B 11

7.4

U).4.~ 72c

I 7.0a/// I~~~~~~~` 6.8

~~~~~~~~~~~~6.6K

I I I I 6.4L60 70 80 90 100 110 50

Po2 (% of baseline)Figure 4: Relation between gut intramucosalpH ((pHd)and changes in intraluminal P02 in each animal in controls(C), peritonitis (P I and P II), and haemorrhage (B I andB II).

Haemorrhage

60 70 80 90

P°2 (% of baseline)Figure 5: Relation between gut intramucosalpH ((pH)and gut intraluminal P02 in peritonitis and haemorrhage.r=0-06 (0.17) in peritonitis and r=0-63 (0.13) inhaemorrhage (p<0002).

7.6

7.4

7.2

7.00-

Ia

cl6.8

666

6.4 L50

Peritonitis

100 110

795

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

Antonsson, Haglund

The mean (SEM) value for all tonometricmeasurements of Po2 at baseline (n=25) was71.9 (1.83) mm Hg. Bergofsky found a Po2 of48 mm Hg in saline instilled in the urinarybladder after three hours of equilibration; thisvalue increased after the breathing of pureoxygen.7 Dawson et al estimated mucosal P02in the small bowel of dogs to range between 35and 45 mm Hg. In one normal volunteer it wasfound to be 45 mm Hg.8 Hamilton et al founda wide range of Po2 values in dogs using atonometric method with fluids instilled inloops of the bowel. Their maximum value was57 mm Hg.'6 Korsback measured Po2 indeoxified saline slowly flushed through a gaspermeable catheter implanted subserosally inthe jejunum of pigs and found a mean at base-line of 35 mm Hg.17 In Korsback's experimentthe 02 of the tissue equilibrated over the wallof the implanted tube to saline flushed throughthe tube, whereas, in our experiment, intra-luminal 02 could equilibrate with saline thatwas kept in the balloon of the tonometer for afull hour. Bohlen, using oxygen micro-electrodes in the mucosa of rats, found a Po2 ofonly 14.8 in the villus apex.18 The tissue Po2 is,however, known to increase towards the baseof the villus, and it can be assumed that intra-luminal Po2 is influenced by that of the crypts,or, more specifically, by Po2 in the fluid of thecrypts.

In all of the above experiments, as well as inexperiments by Sheridan et al on rats'9 and onhumans20 using a Clark oxygen electrode onthe serosal surface of the bowel, changes in guttissue oxygen tension occurred simultaneouslywith changes in arterial Po2. The animals inthe present series had 40% oxygen in theirbreathing air and consequently a high arterialP02, 183 (4.9) mm Hg (n=25) at baseline.This, together with our long equilibration time(one hour), might explain our seemingly highP02 values.When comparing the mean values for

changes in intraluminal P02 in the separategroups no statistical differences were found incontrols and peritonitis. In haemorrhagevalues were different from baseline from t=4hours in B I and from t=2 hours in B II; a dif-ference from controls was noted only at endpoint in B II. This can be explained by a con-siderable overlap because of the individual dif-ferences in the response to the challenge ofintraperitoneal faeces or haemorrhage. It mightbe assumed that, if an individual animal in, forexample, haemorrhage only developed a minorchange in pH,, it would also show only a minorchange in intraluminal Po2 as well as in othervariables. More profound changes in pH,would, on the other hand, be accompanied bymore profound changes in intraluminal P02.To test this and to establish the relationbetween mucosal ischaemia (that is, a fall inpH,) and intraluminal P02 each individualmeasurement of P02 was plotted against themeasurement of pH, at the same timepoint.The slope was then calculated for the animalswith peritonitis (groups P I and P II) and forthose with haemorrhage (groups B I and B II)(Fig 5).

From this graph, where the correlationbetween pHi and intraluminal P02 in haemor-rhage was 0.63 (0.13), it can be seen that adrop in pH, from 7.42 to 6.46 was concomitantwith a change in P02 from 100 at baselinedown to 590/o, or vice versa. If it is acceptedthat intraluminal P02 mirrors the Po2 of thegut mucosa, it follows that a drop in mucosalPo2 is accompanied by a drop in pHi, indicat-ing increasing anaerobic metabolism. Thisbecomes even more clear if we look at thedecreasing DO2 in these two haemorrhagicgroups. The chain of events in haemorrhagethus seems to be as follows: DO2 droppedbecause of a decreased mesenteric blood flowand decreased oxygen carrying capacity, as aresult of a fall in the packed cell volume. Theresult of this was a drop in tissue oxygentension to the extent that aerobic metabolismcould not be maintained and was thereforereplaced by anaerobic processes. As a conse-quence tissue acidity increased, which here wasshown by a fall in pHi.The pattern for the septic animals was dif-

ferent, however, and less straight forward. Thedrop in pHi preceded that in gut DO2 - that is,at a timepoint where no statistically significantdecrease had appeared in DO2, pHi haddecreased significantly. It can therefore not beassumed that low DO2 was the sole, or eventhe most important, cause for gut mucosalacidity. This is in accordance with the findingsby Fink et al.3 When pHi dropped from 7.42 to7.06 only a minor, insignificant, change in P02was noted, from 100 to 96. 1 (the haemorrhagicanimals had a lower pH, at end point but at apH, of 7.06 in this group the correspondingP02 was 84. 1). Although a significant drop inDO2 takes place in both groups with peritonitisonly minor changes are detected in intra-luminal P02. Furthermore, these minorchanges are accompanied by a significant dropin pHi. Thus we have a seemingly paradoxicalsituation where decreased DO2 only results inminor, or even no, changes in intraluminal P02but still there is a drop in pH,, implyinganaerobic metabolism. As the animals withperitonitis did not have as low pH, at end pointas those with haemorrhage, we cannot deter-mine the relation between pHi and intral-uminal P02 at pH, below 7.06, which was ourlowest calculated pH, value in this series ofperitonitis. The possibility cannot be excludedthat, with pH, values in a lower range, thecorrelation curve would reach a deflectionpoint where a further drop in pH, would beaccompanied by a steeper drop in P02.Furthermore, the difference between peri-tonitis and haemorrhage might be questionedas the range was wider in haemorrhage, withpH, values below 7-0, whereas the single lowestvalue in peritonitis was 7.06. We, therefore,recalculated the correlation between pH, andchanges in P02 in haemorrhage excludingthose three animals that had a pH, value atany timepoint below 7.0. We still found acorrelation (r=0*55 (0*17)).

Earlier experiments have shown that thedrop in pH,, seen in septic states, is not solelydependent upon decreased oxygen delivery.

796

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from

Intramucosal pH and intraluminal Po2 797

Rasmussen and Haglund recently found that,in the early stages of septic shock, pHi wasdecreasing even when DO2 was still normal orincreased - that is, during the hyperdynamicphase of shock.4 Fink et al found that endo-toxaemia in pigs resulted in a decrease of gutDO2 and of pHi. If, however, the blood flow inthe superior mesenteric artery was mechani-cally decreased to the same values as in theseptic animals no change in pHi was noted.3 It,thus seems justified to conclude that low DO2alone cannot account for the gut intramucosalacidity seen in sepsis.

In our model of peritonitis, we thus haveobtained evidence that normal concentrationsof oxygen are present in the mucosa and yetthere is anaerobic metabolism as evidenced bya drop in pHi. An explanation of this seeminglyparadoxical situation could be that the uptakeof oxygen in the individual cell is blocked insepsis, or that the utilisation of oxygen in theintracellular metabolism is somehow limited.The most probable function to be disturbed insepsis seems to be the oxidative phosphoryla-tion taking place in the mitochondria, as a lackof oxygen here leads to an anaerobic pathwaywith production of an excess of [H'] explain-ing the acidity detected in the mucosa as wellas in other tissues in septic states.

Although Dawson et al assessed that mito-chondrial function is enhanced by sepsis,21there are abundant data supporting the ideathat mitochondrial function is impaired insepsis, either by bacterial endotoxins directlyor, more probably, by toxic mediators releasedby them. Mela et al, in 1971, could showdefects in liver mitochondrial energy linkedfunctions in haemorrhagic and endotoxinshock. They proposed the destruction of mito-chondrial membranes by lysosomal enzymes asan explanation.22 Astiz and coworkers, in anexperiment using caecal ligation in rats,showed an impaired oxidative metabolism andenergy production in muscle tissue, expressedas increased lactate/pryuvate ratio anddecreased high energy phosphates. This wasthe case even if blood circulation was kept atcontrol values by albumin infusion and tissueP02 was unchanged.23 Hotchkiss et al couldnot detect signs of cellular hypoxia in septicrats using [I8F]fluoromisonidazole. Yet therewas increased blood lactate concentrations intheir experimental settings indicating anaero-bic metabolism. However, the gut was notexamined in their experiment.24 Schaefer andcoworkers have recently shown that cellularoxidative phosphorylation in endotoxic ratswas impaired. In the brain they noted thatthese changes were correlated to the decreasedblood flow induced by endotoxaemia, but inthe small bowel they found this defect to bedisproportionally great and independent ofmaintenance of normal blood flow.25-27

In conclusion, we found a drop in mucosalpH to correlate well with a drop in gut intra-luminal P02 in haemorrhage. In peritonitis, onthe other hand, a significant drop in pH,occurred within a very limited and insignificantchange of intraluminal P02. This can be

explained by a limited oxygen uptake in theindividual cell, possibly resulting from animpairment of mitochondrial function, as hasbeen seen in sepsis by earlier investigators.This study was sponsored by grants from the Swedish MedicalResearch Council, project 4502, the Swedish Medical Society,and Trygg Hansa.

1 Gutierrez G, Palizas F, Doglio G, Wainsztein N, Gallesio A,Pacin J, et al. Gastric intramucosal pH as a therapeuticindex of tissue oxygenation in critically ill patients. Lancet1992; 339: 195-9.

2 Maynard N, Bihari D, Beale R, Smithies M, Baldock G,Mason R, et al. Assessment of splanchnic oxygenation bygastric tonometry in patients with acute circulatoryfailure. JAMA 1993; 270: 1203-10.

3 Fink MP, Antonsson JB, Wang H, Rothschild HR.Increased intestinal permeability in endotoxic pigs.Mesenteric hypoperfusion as an etiologic factor. ArchShock 1991; 126: 211-8.

4 Rasmussen I, Haglund U. Early gut ischemia in experi-mental fecal peritonitis. Circ Shock 1992; 38: 22-8.

5 Gottrup F. Measurement and evaluation of tissue perfusionin surgery (to optimize wound healing and resistance toinfection). In: Leaper DJ, Branicki FJ, ed. Internationalsurgical practice. Oxford: Oxford University Press, 1992:15-39.

6 Hunt TK, Rabkin J, Jensen JA, Jonsson K, von Smitten K,Goodson III WH. Tissue oximetry: an interim report.World JSurg 1987; 11: 126-32.

7 Bergofsky EH. Determination of tissue 02 tensions byhollow visceral tonometers: effect of breathing enriched02 mixtures. Jf Clin Invest 1964; 43: 193-200.

8 Dawson AM, Trenchard D, Guz A. Small bowel tonometry:assessment of small gut mucosal oxygen tension in dogand man. Nature 1965; 206: 943-4.

9 Niinikoski J, Hunt TK. Measurement of wound oxygenwith implanted Silastic tube. Surgery 1972; 71: 22-6.

10 Grum CM, Fiddian-Green RG, Pittenger GL, Grant BJB,Rothman ED, Dantzker DR. Adequacy of tissue oxygena-tion in intact dog intestine. J Appl Physiol 1984; 56:1065-9.

11 von Engelhardt W. Swine cardiovascular physiology - areview. In: Bustad CK, McClellan RO, ed. Swine inbiomedical research. Seattle: Frayn Printing Co, 1966.

12 Antonsson JB, Boyle III CC, Kruithoff KL, Wang H,Sacristan E, Rothschild HR, et al. Validation of tono-metric measurement of gut intramural pH duringendotoxemia and mesenteric occlusion in pigs. Am JPhysiol 1990; 259: G519-23.

13 Kalln R, Borgstrom A, Falt K. Pancreatic enzymes inserum and urine as indicators of pancreatic allograftrejection in the pig. Transplantation 1989; 48: 376-81.

14 Cunningham JA, Cousar CD, Jaffin JH, Harmon JW.Extraluminal and intraluminal PCO2 levels in theischemic intestines of rats. Curr Surg 1987; 44: 229-32.

15 Lundgren 0, Haglund U. The pathophysiology of theintestinal countercurrent exchanger. Life Sci 1978; 23:1411-22.

16 Hamilton JD, Dawson AM, Webb JPW. Observations uponsmall gut 'mucosal' P02 and PCO2 in anesthetized dogs.Gastroenterology 1968; 55: 52-60.

17 Korsback C. Tissue P02 and pCO2 in the gastrointestinaltract and liver during intravenous vasopressin infusion. Anexperimental study in pigs. Acta Chir Scand 1984; 150:593-7.

18 Bohlen HG. Intestinal tissue P02 and microvascularresponses during glucose exposure. Am J Physiol 1980;238: H164-71.

19 Sheridan WG, Lowndes RH, Williams GT, Young HL.Determination of a critical level of tissue oxygenation inacute intestinal ischaemia. Gut 1992; 33: 762-6.

20 Sheridan WG, Lowndes RH, Young HL. Intraoperativetissue oximetry in the human gastrointestinal tract. Am JSurg 1990; 159: 314-9.

21 Dawson KL, Geller ER, Kirkpatrick JR. Enhancement ofmitochondrial function in sepsis. Arch Surg 1988; 123:241-4.

22 Mela L, Bachalzo LV, Miller LD. Defective oxidativemetabolism of rat liver mitochondria in hemorrhagic andendotoxin shock. Amjf Physiol 197 1; 220: 571-7.

23 Astiz M, Rackow EC, Weil MH, Schumer W. Early impair-ment of oxidative metabolism and energy production insevere sepsis. Circ Shock 1988; 26: 311-20.

24 Hotchkiss RS, Rust RS, Dence CS, Wasserman TH, SongS-K, Hwang D-R, et al. Evaluation of the role ofcellular hypoxia in sepsis by the hypoxic marker[18F]fluoromisonidazole. Am J Physiol 1991; 261:R965-72.

25 Schaefer CF, Biber B, Lerner MR, Jobsis-VanderVliet FF,Fagraeus L. Rapid reduction of intestinal cytochromea,a3 during lethal endotoxemia. JT Surg Res 1991; 51:382-9 1.

26 Schaefer CF, Lerner MR, Biber B. Dose-related reductionof intestinal cytochrome a,a3 induced by endotoxin inrats. Circ Shock 1991; 33: 17-25.

27 Schaefer CF, Biber B. Effects of endotoxemia on the redoxlevel of brain cytochrome a,a3 in rats. Circ Shock 1993; 40:1-8.

on 16 March 2019 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.6.791 on 1 D

ecember 1995. D

ownloaded from