- Email: [email protected]
Thermal skin injury: Effect of fluid therapy on the transcapillary colloid osmotic gradient
JOURNAL
OF SURGICAL
RESEARCH
50,272-278
(1991)
Thermal Skin Injury: Effect of Fluid Therapy on the Transcapillary Colloid Osmotic Gradient HENNING
ONARHEIM,
Department
M.D., AND ROLF K. REED, M.D.
of Physiology,
University
of Bergen, Norway
Submitted for publication February
2, 1990
The wick method was developed to collect interstitial fluid for measurements of protein concentration and colloid osmotic pressure (COP) [6]. The wick technique has also been used to investigate alterations in colloid osmotic pressure in interstitial fluid (COP,,) under pathological conditions like congestive heart failure and nephrotic syndrome and following extracorporeal circulation [ 7-91. After thermal injury colloid osmotic pressure has been determined directly in interstitial fluid in a few studies: analyses were performed on fluid collected by subcutaneous wicks [4, lo] or by a wick catheter technique [ll]. Lund and Reed [4] determined COP, following moderate-size and extensive burns in rats, but with no fluid resuscitation. In patients with extensive burns COP, was determined using wick fluid [lo] or a wick catheter [ll], but not until 6 hr after the injury. In the present study alterations in COP, in the early period after thermal injury were determined. The effects of different resuscitation fluids on the transcapillary oncotic gradient were studied by direct measurements in interstitial fluid, thereby circumventing the problem of possible lack of interstitial-lymphatic steady-state. Recently a marked negativity of the interstitial hydrostatic pressure (Pi3 was observed following thermal injury [12, 131. In order to fully outline the postburn changes in the interstitial pressures, we further addressed how infusion of lactated Ringer’s and plasma affected Pi,.
The effects of fluid therapy on interstitial colloid osmotic and hydrostatic pressures in thermally injured skin were investigated in anesthetized rats subjected to full-thickness scald burns to 40% of the body surface area and resuscitation for 3 hr by either lactated Ringer’s or plasma. Interstitial fluid hydrostatic pressure (PU) was reduced from -2 mm Hg to -20 to -40 mm Hg after injury, which will profoundly increase transcapillary filtration. Following the onset of fluid therapy, P,, increased to slightly positive values. In control, colloid osmotic pressure in plasma (COP,) was 20.6 + 0.4 mm Hg and in interstitial fluid (COP,,) 13.7 f 0.3 mm Hg (means’? SEM). The transcapillary oncotic pressure gradient (COP,,,, = COP, - COPif) was 6.9 k 0.4 mm Hg. Following nonresuscitated thermal injury, COP, declined to 18-19 mm Hg (P < 0.05) and COP,, was reduced to 10.4 f 0.5 mm Hg (P < 0.05). Fluid therapy by lactated Ringer’s markedly reduced COP, (12.3 + 0.3 mm Hg; P < 0.05), and COP,, was almost abolished (2.6 + 0.7 mm Hg; P < 0.05). In contrast, plasma infusion maintained COP,, whereas increased significantly (11.1 -t 1.2 mm Hg; P cop,, < 0.05). Noncolloid saline solutions have been preferred for the initial fluid therapy for burns. The present study provides evidence that this will reduce both a situation in which edema formaCOP, and COP,,, tion will be favored. o mu Academic PEWS. IIIC.
INTRODUCTION MATERIALS
Thermal injury rapidly increases vascular permeability and leads to marked edema formation [l-3]. Alterations in capillary permeability have been determined from accumulation of radiolabeled tracer proteins as well as changes in water content [l, 4, 51. Changes in lymph composition and flow have also been used to determine alterations in the composition of interstitial fluid [2,3]. Under steady-state conditions lymph may be representative of interstitial fluid composition. However, a burn injury followed by fluid resuscitation will probably not represent a steady state; in that case lymph may well be different from interstitial fluid. 0022-4804/91$1.50 Copyright 0 1991by AcademicPress,Inc. All rights of reproductionin any form reserved.
AND
METHODS
Nonfasted female Wistar rats (weight range 205-270 g) were anesthetized with pentobarbital (50 mg/kg, ip), with supplementary doses as needed to maintain anesthesia throughout the study period. Rectal temperature was maintained at 37°C by a servocontrolled heating pad. PE50 catheters were placed in the jugular vein and the carotid artery for fluid infusion and collection of blood samples. The back and abdomen were closely clipped. Scald burns were inflicted on 40% of the body surface area (BSA) (20% BSA on the back plus 20% BSA on the 272
ONARHEIM
AND
REED:
TRANSCAPILLARY
abdomen) [14]. Exposure of back skin to boiling hot water for 10 set and abdominal skin for 4 set results in uniform, full-thickness injuries [15]. Sham control animals were treated similarly to the animals in the burn groups except for immersion in water. Citrated rat plasma was produced by exsanguination of 500-g Wistar rats; one part of sodium citrate (3.13%) was added per seven parts of whole blood. The colloid osmotic pressure of this plasma was 13-14 mm Hg. Measurements Interstitial fluid was collected with Wick method. multifilamentous nylon wicks (Enkalon 3 X 3, Produktgroep Industrielle Garens, Arnheim, Holland). The wicks, each consisting of three strands made up of 300 filaments of about 25 pm in diameter, were rinsed in acetone, ethanol, and distilled water [ 161 and soaked in 0.9% saline prior to implantation [17]. Pairs of doublestranded wicks were placed subcutaneously on the back in lengths of 4-6 cm by means of a straight suture needle [ 171. To prevent evaporation, implantation of wicks and all handling of wicks in open air were performed inside an infant incubator, which served as a humidified chamber (100% relative humidity at room temperature). For each sampling of wick fluid, two double-stranded wicks were implanted. Wicks were separated by at least 1 cm. The larger part of the rat’s back was exposed to thermal injury; the small area of normal skin available precluded collection of wick fluid from noninjured areas. After an implantation time of 60 min, the wicks were swiftly pulled out and their central part transferred to 2-ml centrifuge tubes supplied with a funnel [ 161. Wicks that were visibly bloodstained were discarded [6]. After centrifugation, 5-10 ~1 of wick fluid could be pipetted up from the bottom of the centrifuge tubes. Colloid osmotic pressure. COP was measured in serum (COP,) and in wick fluid (COPJ with a colloid osmometer designed for sample volumes as low as 5 ~1 [18]. Diaflo YMlO ultrafiltration membranes (Amicon, Danvers, MA) were used on the osmometer. This membrane will reject 90% of globular solutes of mol wt 10,000 and >98% of albumin (manufacturer’s specifications). COP was expressed as the absolute value of the recorded pressure drop (see Fig. 2). COP measurements were performed on the day of the experiment and are expressed as the mean values of duplicate analyses. Parallel COP measurements usually deviated by less than 0.5 mm Hg. A house standard of rat serum (COP around 22 mm Hg) was analyzed daily as a control (day-to-day coefficient of variation: 5%). The wick technique was tested in vitro by soaking wicks in rat serum or rat serum diluted with saline and thereafter processed as described above. COP of the fluid obtained after centrifugation was identical to that of the control serum. Wick fluid from wicks soaked in normal saline consistently gave readings of zero pres-
ONCOTIC
GRADIENT
273
IN BURNS
sure. In separate in uiuo experiments COP in wick fluid obtained from eight parallel wicks gave a coefficient of variation of 9-13%. Interstitial fluid hydrostatic pressure. Pif was measured by micropipets [ 19,201. Sharpened glass micropipets (2- to 4-pm tip diameter) filled with 0.5 M NaCl were inserted into dermis. The pipets were connected to a servocontrolled counterpressure unit (“servonull”) designed to keep the electric conductance within the tip of the pipet constant. Any external hydrostatic pressure acting on the micropipet either by forcing “isotonic” fluid into the pipet or by sucking hypertonic pipet fluid out will alter the electrical conductance within the pipet tip. A two-way membrane pump connected to the pipet will generate a counterpressure to readjust the electrical conductance within the pipet tip. This counterpressure generated by the membrane pump is identical to the external pressure [ 201. Only pipet positions with no visible deformation of the skin surface, as judged through a stereomicroscope, were accepted for registration of Pi,. Pressure readings were accepted when the following criteria were met [20]: (A) recorded pressure was constant when increasing the feedback gain; (B) fluid communication between pipet and interstitial fluid was free (a suction pressure applied by the pump resulted in increased electrical resistance in the pipet tip); and (C) zero pressure reading before and after measurement of P, was reproducible. Experimental
Protocol
Wick study. The experimental procedure was divided into three phases: A. Control period: Following insertion of vascular lines, wicks were inserted subcutaneously on the back and removed after a 60-min equilibration period. The animals were then allocated to one of five groups: Group
Thermal
injury
1. Control 2. No fluid group 3. Lactated Ringer’s
None 40% BSA 40% BSA
4. Plasma group 5. Infusion only
40% BSA None
Fluid therapy None None Lactated Ringer’s 5 ml/hr Plasma 2.5 ml/hr Lactated Ringer’s 5 ml/hr
B. Induction of thermal injury: This procedure required 5-7 min, whereafter the animal was returned to the incubator. C. Postburn observation period: A second pair of subcutaneous wicks were implanted for a 60-min period immediately after thermal injury. Fluid infusion was started 15 min after thermal injury, correspondingly 15 min after sham burn. A third pair of wicks was inserted from 120 to 180 min postinjury. Carotid blood samples
214
JOURNAL
OF SURGICAL
RESEARCH:
were collected preburn and 1 and 3 hr postburn for determination of hematocrit (HCT) and COP,. Pi, in dermis Measurement of P, by micropuncture. was measured in control, whereafter a 40% BSA thermal injury was inflicted. Animals were resuscitated by lactated Ringer’s at 5 ml/hr iv (lactated Ringer’s group, n = 6) or by plasma at 2.5 ml/hr iv (plasma group, n = 6). Infusions were started 15 min after injury and continued for a 3-hr study period. For practical reasons acceptable measurements of Pif could usually not be obtained for the first 5-10 min postburn. During the recording of P, liquid mineral oil was applied to the skin to avoid evaporation from the skin surface. Measurements were continued up to 180 min postinjury. Statistical
Methods
Results are presented as means + SEM. Analysis of variance (ANOVA) with repeated measures over time, one-way ANOVA, and t tests (paired and unpaired) were employed to assess differences between groups or treatments. Differences were assumed significant if P < 0.05. Biomedical datapack (BMDP, version 1985) was used for statistical analysis. RESULTS
Results are presented in Figs. l-3 and in Table 1. Results from the group which received lactated Ringer’s but no burn are only presented in the table. Preburn COP, was 20.6 ? 0.4 mm Hg (grand mean). In the control group COP, remained stable for the entire study period. Within 1 hr after thermal injury COP, was reduced to 18.3 + 0.8 mm Hg in animals that were not resuscitated (P < 0.05). Following resuscitation by lactated Ringer’s, COP,, decreased markedly: within 3 hr COP, was reduced by 7-8 mm Hg to 12.3 +- 0.3 mm Hg (P < 0.05). In plasma-resuscitated animals COP, increased from 21.3 to 22.4 mm Hg over the study period (P < 0.05). In control COP, was 13.7 + 0.4 mm Hg (grand mean). There was a moderate reduction in COP, over the observation period even in the control group (P < 0.05). Following thermal injury there was a marked reduction in COP,, most evident following fluid replacement by lactated Ringer’s (P < 0.05 vs control group) (Table 1). In the untreated animals and following plasma infusion, COP, declined more than in the controls, although the combined effect of treatment and time did not reach a statistically significant difference (P > 0.05 compared to the control group, by repeated measures ANOVA). The transcapillary oncotic gradient (COP,,,) in thermally injured skin was calculated as the difference between colloid osmotic pressure in plasma and that in interstitial fluid (COP, - COP,3 (Table 1; Fig. 1). Differences in COPpad between groups were moderate in control (Table 1). With time COPwad increased slightly
VOL.
50, NO. 3, MARCH
1991
in the control group (P < 0.05). Following infusion of rat plasma, COPmad increased by 4-5 mm Hg (P < 0.05 vs control group). A different pattern was observed following resuscitation by lactated Ringer’s, which led to an almost abolished COPpad (P -C 0.05) (Fig. 1). Infusion of lactated Ringer’s to noninjured rats (the infusion only group) significantly reduced COP, (P < 0.05 vs control group) (Table 1). Still, COP, remained higher in the infusion only group compared to that after thermal injury followed by lactated Ringer’s, administered at the same infusion rate (P < 0.05) (Fig. 1). Contrary to results in the thermally injured rats, infusion of lactated Ringer’s to noninjured rats (infusion only group) did not alter COP,,,, since both COP,, and COP, were reduced to the same extent (Table 1). HCT was stable around 50-51% in the control group. Following nonresuscitated thermal injury, HCT increased from 51 to 59% within 1 hr and remained elevated at 3 hr after injury. HCT was transiently elevated (by around 3%) 1 hr postinjury in both the lactated Ringer’s and the plasma groups, but returned toward control values 3 hr postinjury in both groups. When COP was measured in sera or in wick fluid from noninjured skin, a stable equilibrium pressure was rapidly obtained (see Fig. 2). No “overshoot” (transient positive pressure peak) was observed after the membrane was washed with saline, indicating that normally, larger molecules did not leak through the membrane. Samples of wick fluid collected after thermal injury consistently gave an initial, negative pressure deflection before a flat equilibrium pressure was obtained (Fig. 2). Further, an “overshoot” was observed after the membrane was washed with saline (Fig. 2). Such changes were not observed in the plasma samples. The effect of osmometer membrane “pore size” (cutoff level) was evaluated for two different osmometer membranes. The Amicon YMlO membrane (used throughout the present study) was compared to the Amicon PM30 membrane (30,000 mol wt cutoff), which has been used in a number of previous studies [4,16,21]. The YMlO membrane gave l-2 mm Hg higher values for COP around 20 mm Hg and 1 mm Hg higher values for COP around 10 mm Hg. Equilibration of small ions over the osmometer membrane was evaluated in separate experiments: 0.5 M NaCl gave a very transient pressure response when applied on the osmometer membrane. Sodium citrate (citric acid: mol wt 192) gave a more pronounced, still transient pressure response (Fig. 2). Pifwas slightly subatmospheric (-2.0 f 0.2 mm Hg) in control (Fig. 3). Following thermal injury Pi, rapidly declined to -20 to -40 mm Hg (Fig. 3). The most negative values were observed in the first lo- to 20-min intervals. Following fluid infusion, Pi, returned toward control values and later remained slightly positive. Pi, became positive with plasma infusion sooner than it did following infusion of lactated Ringer’s (P G O.OEi)(Fig. 3).
ONARHEIM
AND
REED:
TRANSCAPILLARY
ONCOTIC
TABLE Colloid
Osmotic
Pressures
in Plasma (COP,), in Controls
GRADIENT
275
IN BURNS
1
Interstitial Fluid (COP,&, and Colloid and following Thermal Injury
Osmotic
Gradient
(COP,,)
Burn
Control COP, (mm Hg) Control 60 min 180 min COP, (mm Hg) Control 60 min 180 min COPgrad (mm
Lactated Ringer’s
No fluid
19.7 +- 0.5
18.3 f 0.8’.+
15.0 f 0.6*.+
20.9 + 0.9
19.0 + 0.7
12.3 + 0.3*-+
21.3 -r- 1.0 20.7 f 1.0 22.4 f 0.9*
12.7 k 0.7 11.5 f 0.7* 10.9 f 0.9*
13.7 +- 0.4 12.5 -t 0.6 10.4 + 0.5*
13.6 + 0.8 12.7 + 0.7
14.9 f 0.5+ 13.4 f 0.5*
9.7 zk 0.7*
11.3 ‘- 0.8*
8.4 of- 0.3*-+
8.4 + 0.6 9.2 f 0.4*
6.4 -c 0.7 5.8 + 0.6+ 8.5 + 0.8
6.4 f 0.9 7.3 iz 0.8
7.4 f 1.7 6.1 + 0.8+ 7.7 f 0.7
DISCUSSION
The wick method was developed to collect interstitial fluid for measurements of protein concentration and colloid osmotic pressure [6,17]. Wick fluid was found to give a correct estimate of COP, in subcutaneous tissue provided that sufficient proteins were added to the wick before or after insertion [ 171.
20
E
15
5
10
u 5 0
25
13.6 _’ 1.7 11.5 + 0.9
20.1 +- 0.7
20.8 k 0.9
10.0 zk 0.4*
0,
21.0 + 0.6 17.7 2 0.3*,+ 16.1 f 0.8*!+
21.1 + 0.8
6.0 k 0.8+ 2.3 3~ 0.9*,+ 2.6 f 0.7*~’
Note. Infusion only group: no thermal injury, lactated Ringer’s infused at the same rate as thermal Means k 1 SEM, n = 8. * P < 0.05 vs prehurn. + P < 0.05 vs control group (by ANOVA and t test).
I
Infusion only
I-k)
Control 60 min 180 min
25
Plasma
li!iL Lactated
Ringer’s
r
Plasma
C
Control
60
180
Control
60
180
FIG. 1. Colloid osmotic pressure in plasma (COP,) (open bars) and interstitial fluid (COP,) (solid bars) in control and 60 and 180 min after burn injury. Colloid osmotic gradient (COP,,) corresponds to the difference between COP, and COP,. Means f SEM, n = 8.
injury, resuscitated
by lactated Ringer’s.
Two recent modifications of the wick technique have suggested that saline-soaked wicks might give a certain underestimate of COP, [21, 221. However, “the crossover method” [22] requires insertion of at least three wicks for each determination of COP,,, which will limit its application for repeated measurements in a small animal. “Dry wicks” [21, 221 could have been used, still, saline-soaked wicks were preferred to allow for repeated measures and for comparisons to previous studies [4,
No fluid
Control group
11.1 f 1.2*
101.
Fluid replacement seems to have been sufficient since HCT was only transiently elevated in the resuscitated cases. On the basis of the alterations in HCT (corrected for blood samples), we predict that 3 hr postinjury plasma volume was unchanged from control in plasmaresuscitated animals and reduced by 5% in animals that received infusion of lactated Ringer’s. The Amicon YMlO membrane was used on the osmometer; this membrane has a pore diameter which should safely exclude all plasma proteins. Wick fluid from thermally injured skin gave an initial, negative pressure deflection on the osmometer prior to a stable equilibrium pressure (Fig. 2). A transient positive pressure peak occurred following washing of the membrane. Analogue observations were made following application of hypertonic saline, titrated rat plasma, or sodium citrate on the osmometer (Fig. 2), while no such phenomena were seen in wick fluid from noninjured skin. These observations suggest that after thermal injury molecules of molecular weights below the cutoff level for the membrane leaked through the intact osmometer membrane. This is in accordance with previous suggestions that
276
JOURNAL
OF SURGICAL
RESEARCH:
FIG. 2. Pressure recordings with colloid osmometer. (Top) Normal pressure tracing (flat equilibrium) obtained after plasma or wick fluid application. (Middle) Wick fluid from thermally injured skin area demonstrates an initial, brief pressure deflection following application of sample before a normal (flat) equilibrium was reached, a slight “overshoot” is seen following washing of membrane. (Bottom) Application of sodium citrate: transient pressure response; note marked “overshoot” following washing of membrane (due to small molecules leaking through membrane). D, drying of the membrane; A, application of sample; S, washing of membrane with 0.9% saline. For details, see Results and Discussion.
thermal injury (by breakdown of larger molecules) may generate “new” osmoles [2, 131. For control sera the YMlO membrane gave l-2 mm Hg higher COP values than the Amicon PM30 membrane (used in a number of previous studies) [4, 16, 21, 221. This may in part explain the somewhat higher COP, in control (13.7 + 0.4 mm Hg) compared to that in some previous studies (around 9-11 mm Hg; [4, 16, 21, 221). COP in plasma remained stable in the control group (Table 1). Nonresuscitated thermal injury was followed by a l-2 mm Hg reduction in COP, (Fig. l), which is somewhat less than a 2-6 mm Hg reduction observed in a previous study where no fluid therapy was provided [4]. A marked reduction of COP, was observed following lactated Ringer’s infusion: COP, was reduced to 12.3 + 0.3 mm Hg at 3 hr postinjury (P < 0.05) (Fig. 1). In thermally injured skin the extravasation rate of plasma proteins may increase to lo-50 times that of controls [ 1, 4,5], leading to a marked reduction of the intravascular pool of plasma proteins. COP, will be reduced when the remaining plasma proteins are diluted following massive saline infusion. Thus COP, was lower after thermal injury followed by infusion of lactated Ringer’s than following infusion of the same volume of Ringer’s to noninjured animals (Table 1).
VOL.
50, NO. 3, MARCH
1991
There was a certain reduction in COP,, with time even in the control experiments (12.7 vs 10.9 mm Hg; P < 0.05). When wick techniques are used to collect interstitial fluid it may therefore be relevant to specify the timing of wick fluid samples in relation to the induction of anesthesia. In the present study COP, was reduced by 3-4 mm Hg after thermal injury (P < 0.05). In extensively burned patients, resuscitated mainly by acetated Ringer’s, COP, in injured skin was reduced compared to that in normals [lo]. A significant reduction of COP, in severely burned patients was also found 4-8 hr after thermal injury, using a wick catheter technique [ 111. In contrast, in rats COP,, was not significantly altered following larger (40% BSA), nonresuscitated burns, whereas following smaller (10% BSA) burns even a certain increase of COP, was observed [4]. COP,r.¶d was 6-8 mm Hg in the control situation, which compares well to previous studies in the rat [4, 161. COPGad was not significantly altered after thermal injury in the nonresuscitated cases (Table 1). This contrasts previous observations that COP,,, was nullified or even reversed after nonresuscitated thermal injury [4]. However, in that study experiments were not performed within a humidified chamber [4]. Evaporative fluid losses from wicks handled in the open will rapidly increase COP, [ 161 and may lead to erroneously low values for COP,,,. In the present study handling of wicks solely within an incubator eliminated that problem: in recovery studies wick fluid from wicks soaked in serum gave the same COP as the primer. Following volume replacement by lactated Ringer’s, COP,,, declined to 2.6 +- 0.7 mm Hg (P < 0.05), which still is significantly different from a zero gradient (P
0”
A-A
-20.
,j
I’\{
,~.
Lactated Ringer’s
;-O,Unresuyitated, 8
I C
2 1 limo (hours) ofter thermal injury
3
FIG. 3. Interstitial fluid hydrostatic pressure (P,) measured by micropuncture in control (C) and following thermal injury. Fluid therapy by lactated Ringer’s or plasma was started 15 min postinjury. Means + SEM. For comparison data from Lund et al. [12] for Ps following a similar, nonresuscitated injury are also included.
ONARHEIM
AND
REED:
TRANSCAPILLARY
< 0.05). This is a more moderate reduction than that observed in severely burned patients, where COP,, was found to be equal to or even higher than COP, when measured 6-12 hr postinjury [lo]. However, it corresponds well to previous estimates of COP,,, after thermal injury in sheep resuscitated by lactated Ringer’s, using lymph as a representative for interstitial fluid in injured skin [3]. Fluid transport across the capillary wall is described by the Starling equation, J, = CFC. (P, - Pi, - 0. (COP,, - COPif)),
where J, is the volume of fluid filtered across the capillary wall, CFC is the capillary filtration coefficient, PC is the capillary hydrostatic pressure, P, is the interstitial fluid hydrostatic pressure, and u is the capillary reflection coefficient for plasma proteins. Normally the net filtration pressure (the net pressure imbalance across the capillary wall) is on the order of 0.5 mm Hg. For methodological reasons Pi, could only be measured in dermis; Pi, in dermis may not necessarily be representative for whole skin (including both dermis and subcutis) [ 121. Still, following thermal injury P, decreased from a normal value of -2 mm Hg to values around -20 to -40 mm Hg (Fig. 3). The markedly negative Pi, in injured skin may increase the net filtration pressure up to two orders of magnitude and will be the main driving force for edema formation in the initial stages. After the first 30-min period postinjury, the negative Pi, subsided (Fig. 3). From 1 hr postinjury Pi, was around ambient pressure or became slightly positive, as is seen in other edematous states [ZO]. When compared to data from Lund et al. [ 121 of Pi, following nonresuscitated thermal injury (Fig. 3), normal Pi, in the present study was reached sooner with resuscitation, and more promptly following plasma infusion (P < 0.05 vs lactated Ringer’s). The gradient in colloid osmotic pressure between plasma and the interstitium will influence the net transcapillary fluid transport: a reduced effective . COP,,d) will favor edema formation. After COP,db thermal injury, u has been estimated to be reduced from 0.87 to 0.45 [23]; this may further contribute to a reduced effective COP,, . When Pif is positive, PC (capillary hydrostatic pressure) and COPsad will be the driving forces for fluid filtration. A reduced COP,,, may facilitate edema formation also in noninjured tissue. Actually, fluid replacement by infusion of lactated Ringer’s after thermal injury was found to increase fluid accumulation in noninjured tissue [3, 41. This is to our knowledge the first study where colloid osmotic pressure has been determined directly in interstitial fluid following different forms of parenteral fluid substitution. Plasma and lactated Ringer’s had opposite
ONCOTIC
GRADIENT
IN BURNS
277
effects on COP,,, after a standardized thermal injury. Plasma maintained, or actually increased, COP,,,,, whereas lactated Ringer’s almost annihilated COP,,, . Saline solutions (without colloids) have been preferred for the initial fluid therapy in burns. The present study underlines that this will reduce both COP, and COP,,, and may therefore favor edema formation. ACKNOWLEDGMENTS We are grateful to Eli Gunn Kjarlaug and Malvin Gismervik for technical assistance. The study received financial support from The Norwegian Research Council for Science and the Humanities and The Norwegian Council on Cardiovascular Diseases.
REFERENCES 1. Leape, L. L. Kinetics of burn edema formation in primates. Ann. Surg. 176:223,1972. 2. Arturson, G. Microvascular permeability to macromolecules in thermal
injury. Acta Physiol. Stand. Suppl. 463: 111, 1979. R. H., Kramer, G., and Harms, B. Role of thermal injury-induced hypoproteinemia on fluid flux and protein permeability in burned and nonburned tissue. Surgery 95: 136,1984. Lund, T., and Reed, R. K. Microvascular fluid exchange following thermal skin injury in the rat: Changes in extravascular colloid osmotic pressure, albumin mass, and water content. Circ. Shock 20: 91, 1986. Onarheim, H., Lund, T., and Reed, R. Thermal skin injury. II. Effects on edema formation and albumin extravasation of fluid resuscitation with lactated Ringer’s, plasma, and hypertonic saline (2,400 mosmol/l) in the rat. Circ. Shock 27: 25, 1989. Aukland, K., and Fadnes, H. 0. Protein concentration of interstitial fluid collected from rat skin by a wick method. Acta Physiol. Stand. 88: 350, 1973. Noddeland, H., Omvik, P., Lund-Johansen, P., Ofstad, J., and Aukland, K. Interstitial colloid osmotic and hydrostatic pressures in human subcutaneous tissue during early stages of heart failure. Clin. Physiol. 4: 283, 1984. Koomans, H. A., Kortlandt, W., Geers, A. B., and Mees, E. J. D. Lowered protein content of tissue fluid in patients with the nephrotic syndrome: Observations during disease and recovery. Nephron 40: 391, 1985. Rein, K. A., Semb, K., Myhre, H. O., Levang, 0. W., Christensen, O., Stenseth, R., and Sande, E. Transcapillary fluid balance in subcutaneous tissue of patients undergoing aortocoronary bypass with extracorporeal circulation. Stand. J. Thorac. Cardiovmc. Surg. 22: 267, 1988. Pitklnen, J., Lund, T., Aanderud, L., and Reed, R. K. Transcapillary colloid osmotic pressures in injured and non-injured skin of seriously burned patients. Burns Incl. Therm. Znj. 13: 198, 1987. Wachtel, T. L., Frank, H. A., Sanders, R., Hargens, A. R., and Peters, R. M. Definition of the Starling forces with wick catheter in burned patients. J. Burn Care Rehabil. 4: 331, 1983. Lund, T., Wiig, H., and Reed, R. K. Acute postburn edema: Role of strongly negative interstitial Auid pressure. Amer. J. Physiol. 255 (Heart Circ. Physiol. 24): H1069, 1988. Lund, T., Onarheim, H., Wiig, H., and Reed, R. K. Mechanisms behind increased dermal imbibition pressure in acute burn edema. Amer. J. Physiol. 256 (Heart Circ. Physiol. 25): H940, 1989.
3. Demling,
4.
5.
6.
7.
8.
9.
10.
11.
12. 13.
278 14. 15.
16.
JOURNAL
OF SURGICAL
RESEARCH:
Walker, H. L., and Mason, A. D. A standard animal burn. J. Trauma 8: 1049,1968. Herndon, D. N., Wilmore, D. W., and Mason, A. D. Development and analysis of a small animal model simulating the human postburn hypermetabolic response. J. Surg. Res. 25: 394, 1978. Heir, S., and Wiig, H. Subcutaneous interstitial fluid colloid osmotic pressure in dehydrated rats. Actu Physiol. Stand. 133:
20. 21.
365,1988. 17.
18. 19.
Fadnes, H. O., and Aukland, K. Protein concentration and colloid osmotic pressure of interstitial fluid collected by the wick technique. Microuasc. Res. 14: 11, 1977. Aukland, K., and Johnsen, H. M. A colloid osmometer for small 6uid samples. Acta Physiol. &and. SO: 485, 1974. Wiederhielm, C. A., Woodbury, J. W., Kirk, S., and Rushmer,
22.
23.
VOL.
50, NO. 3, MARCH
1991
R. F. Pulsatile pressures in the microcirculation of frog’s mesentery. Amer. J. Physiol. 207: 173, 1964. Wiig, H., and Reed, R. K. Compliance of the interstitial space in rats. II. Studies on skin. Acta Physiol. Stand. 113: 30'7, 1981. Wiig, H., Heir, S., and Aukland, K. Colloid osmotic pressure of interstitial fluid in rat subcutis and skeletal muscle: Comparison of various wick sampling techniques. Acta Physiol. &and. 133: 167, 1988. Kramer, G. C., Sibley, L., Aukland, K., and Renkin, E. M. Wick sampling of interstitial fluid in rat skin: Further analysis and modifications of the method. Microvase. Res. 32: 39, 1986. Pitt, R. M., Parker, J. C., Jurkovich, G. J., Taylor, A. E., and Curreri, P. W. Analysis of altered capillary pressure and permeability after thermal injury. J. Surg. Res. 42: 693, 1987.
OF SURGICAL
RESEARCH
50,272-278
(1991)
Thermal Skin Injury: Effect of Fluid Therapy on the Transcapillary Colloid Osmotic Gradient HENNING
ONARHEIM,
Department
M.D., AND ROLF K. REED, M.D.
of Physiology,
University
of Bergen, Norway
Submitted for publication February
2, 1990
The wick method was developed to collect interstitial fluid for measurements of protein concentration and colloid osmotic pressure (COP) [6]. The wick technique has also been used to investigate alterations in colloid osmotic pressure in interstitial fluid (COP,,) under pathological conditions like congestive heart failure and nephrotic syndrome and following extracorporeal circulation [ 7-91. After thermal injury colloid osmotic pressure has been determined directly in interstitial fluid in a few studies: analyses were performed on fluid collected by subcutaneous wicks [4, lo] or by a wick catheter technique [ll]. Lund and Reed [4] determined COP, following moderate-size and extensive burns in rats, but with no fluid resuscitation. In patients with extensive burns COP, was determined using wick fluid [lo] or a wick catheter [ll], but not until 6 hr after the injury. In the present study alterations in COP, in the early period after thermal injury were determined. The effects of different resuscitation fluids on the transcapillary oncotic gradient were studied by direct measurements in interstitial fluid, thereby circumventing the problem of possible lack of interstitial-lymphatic steady-state. Recently a marked negativity of the interstitial hydrostatic pressure (Pi3 was observed following thermal injury [12, 131. In order to fully outline the postburn changes in the interstitial pressures, we further addressed how infusion of lactated Ringer’s and plasma affected Pi,.
The effects of fluid therapy on interstitial colloid osmotic and hydrostatic pressures in thermally injured skin were investigated in anesthetized rats subjected to full-thickness scald burns to 40% of the body surface area and resuscitation for 3 hr by either lactated Ringer’s or plasma. Interstitial fluid hydrostatic pressure (PU) was reduced from -2 mm Hg to -20 to -40 mm Hg after injury, which will profoundly increase transcapillary filtration. Following the onset of fluid therapy, P,, increased to slightly positive values. In control, colloid osmotic pressure in plasma (COP,) was 20.6 + 0.4 mm Hg and in interstitial fluid (COP,,) 13.7 f 0.3 mm Hg (means’? SEM). The transcapillary oncotic pressure gradient (COP,,,, = COP, - COPif) was 6.9 k 0.4 mm Hg. Following nonresuscitated thermal injury, COP, declined to 18-19 mm Hg (P < 0.05) and COP,, was reduced to 10.4 f 0.5 mm Hg (P < 0.05). Fluid therapy by lactated Ringer’s markedly reduced COP, (12.3 + 0.3 mm Hg; P < 0.05), and COP,, was almost abolished (2.6 + 0.7 mm Hg; P < 0.05). In contrast, plasma infusion maintained COP,, whereas increased significantly (11.1 -t 1.2 mm Hg; P cop,, < 0.05). Noncolloid saline solutions have been preferred for the initial fluid therapy for burns. The present study provides evidence that this will reduce both a situation in which edema formaCOP, and COP,,, tion will be favored. o mu Academic PEWS. IIIC.
INTRODUCTION MATERIALS
Thermal injury rapidly increases vascular permeability and leads to marked edema formation [l-3]. Alterations in capillary permeability have been determined from accumulation of radiolabeled tracer proteins as well as changes in water content [l, 4, 51. Changes in lymph composition and flow have also been used to determine alterations in the composition of interstitial fluid [2,3]. Under steady-state conditions lymph may be representative of interstitial fluid composition. However, a burn injury followed by fluid resuscitation will probably not represent a steady state; in that case lymph may well be different from interstitial fluid. 0022-4804/91$1.50 Copyright 0 1991by AcademicPress,Inc. All rights of reproductionin any form reserved.
AND
METHODS
Nonfasted female Wistar rats (weight range 205-270 g) were anesthetized with pentobarbital (50 mg/kg, ip), with supplementary doses as needed to maintain anesthesia throughout the study period. Rectal temperature was maintained at 37°C by a servocontrolled heating pad. PE50 catheters were placed in the jugular vein and the carotid artery for fluid infusion and collection of blood samples. The back and abdomen were closely clipped. Scald burns were inflicted on 40% of the body surface area (BSA) (20% BSA on the back plus 20% BSA on the 272
ONARHEIM
AND
REED:
TRANSCAPILLARY
abdomen) [14]. Exposure of back skin to boiling hot water for 10 set and abdominal skin for 4 set results in uniform, full-thickness injuries [15]. Sham control animals were treated similarly to the animals in the burn groups except for immersion in water. Citrated rat plasma was produced by exsanguination of 500-g Wistar rats; one part of sodium citrate (3.13%) was added per seven parts of whole blood. The colloid osmotic pressure of this plasma was 13-14 mm Hg. Measurements Interstitial fluid was collected with Wick method. multifilamentous nylon wicks (Enkalon 3 X 3, Produktgroep Industrielle Garens, Arnheim, Holland). The wicks, each consisting of three strands made up of 300 filaments of about 25 pm in diameter, were rinsed in acetone, ethanol, and distilled water [ 161 and soaked in 0.9% saline prior to implantation [17]. Pairs of doublestranded wicks were placed subcutaneously on the back in lengths of 4-6 cm by means of a straight suture needle [ 171. To prevent evaporation, implantation of wicks and all handling of wicks in open air were performed inside an infant incubator, which served as a humidified chamber (100% relative humidity at room temperature). For each sampling of wick fluid, two double-stranded wicks were implanted. Wicks were separated by at least 1 cm. The larger part of the rat’s back was exposed to thermal injury; the small area of normal skin available precluded collection of wick fluid from noninjured areas. After an implantation time of 60 min, the wicks were swiftly pulled out and their central part transferred to 2-ml centrifuge tubes supplied with a funnel [ 161. Wicks that were visibly bloodstained were discarded [6]. After centrifugation, 5-10 ~1 of wick fluid could be pipetted up from the bottom of the centrifuge tubes. Colloid osmotic pressure. COP was measured in serum (COP,) and in wick fluid (COPJ with a colloid osmometer designed for sample volumes as low as 5 ~1 [18]. Diaflo YMlO ultrafiltration membranes (Amicon, Danvers, MA) were used on the osmometer. This membrane will reject 90% of globular solutes of mol wt 10,000 and >98% of albumin (manufacturer’s specifications). COP was expressed as the absolute value of the recorded pressure drop (see Fig. 2). COP measurements were performed on the day of the experiment and are expressed as the mean values of duplicate analyses. Parallel COP measurements usually deviated by less than 0.5 mm Hg. A house standard of rat serum (COP around 22 mm Hg) was analyzed daily as a control (day-to-day coefficient of variation: 5%). The wick technique was tested in vitro by soaking wicks in rat serum or rat serum diluted with saline and thereafter processed as described above. COP of the fluid obtained after centrifugation was identical to that of the control serum. Wick fluid from wicks soaked in normal saline consistently gave readings of zero pres-
ONCOTIC
GRADIENT
273
IN BURNS
sure. In separate in uiuo experiments COP in wick fluid obtained from eight parallel wicks gave a coefficient of variation of 9-13%. Interstitial fluid hydrostatic pressure. Pif was measured by micropipets [ 19,201. Sharpened glass micropipets (2- to 4-pm tip diameter) filled with 0.5 M NaCl were inserted into dermis. The pipets were connected to a servocontrolled counterpressure unit (“servonull”) designed to keep the electric conductance within the tip of the pipet constant. Any external hydrostatic pressure acting on the micropipet either by forcing “isotonic” fluid into the pipet or by sucking hypertonic pipet fluid out will alter the electrical conductance within the pipet tip. A two-way membrane pump connected to the pipet will generate a counterpressure to readjust the electrical conductance within the pipet tip. This counterpressure generated by the membrane pump is identical to the external pressure [ 201. Only pipet positions with no visible deformation of the skin surface, as judged through a stereomicroscope, were accepted for registration of Pi,. Pressure readings were accepted when the following criteria were met [20]: (A) recorded pressure was constant when increasing the feedback gain; (B) fluid communication between pipet and interstitial fluid was free (a suction pressure applied by the pump resulted in increased electrical resistance in the pipet tip); and (C) zero pressure reading before and after measurement of P, was reproducible. Experimental
Protocol
Wick study. The experimental procedure was divided into three phases: A. Control period: Following insertion of vascular lines, wicks were inserted subcutaneously on the back and removed after a 60-min equilibration period. The animals were then allocated to one of five groups: Group
Thermal
injury
1. Control 2. No fluid group 3. Lactated Ringer’s
None 40% BSA 40% BSA
4. Plasma group 5. Infusion only
40% BSA None
Fluid therapy None None Lactated Ringer’s 5 ml/hr Plasma 2.5 ml/hr Lactated Ringer’s 5 ml/hr
B. Induction of thermal injury: This procedure required 5-7 min, whereafter the animal was returned to the incubator. C. Postburn observation period: A second pair of subcutaneous wicks were implanted for a 60-min period immediately after thermal injury. Fluid infusion was started 15 min after thermal injury, correspondingly 15 min after sham burn. A third pair of wicks was inserted from 120 to 180 min postinjury. Carotid blood samples
214
JOURNAL
OF SURGICAL
RESEARCH:
were collected preburn and 1 and 3 hr postburn for determination of hematocrit (HCT) and COP,. Pi, in dermis Measurement of P, by micropuncture. was measured in control, whereafter a 40% BSA thermal injury was inflicted. Animals were resuscitated by lactated Ringer’s at 5 ml/hr iv (lactated Ringer’s group, n = 6) or by plasma at 2.5 ml/hr iv (plasma group, n = 6). Infusions were started 15 min after injury and continued for a 3-hr study period. For practical reasons acceptable measurements of Pif could usually not be obtained for the first 5-10 min postburn. During the recording of P, liquid mineral oil was applied to the skin to avoid evaporation from the skin surface. Measurements were continued up to 180 min postinjury. Statistical
Methods
Results are presented as means + SEM. Analysis of variance (ANOVA) with repeated measures over time, one-way ANOVA, and t tests (paired and unpaired) were employed to assess differences between groups or treatments. Differences were assumed significant if P < 0.05. Biomedical datapack (BMDP, version 1985) was used for statistical analysis. RESULTS
Results are presented in Figs. l-3 and in Table 1. Results from the group which received lactated Ringer’s but no burn are only presented in the table. Preburn COP, was 20.6 ? 0.4 mm Hg (grand mean). In the control group COP, remained stable for the entire study period. Within 1 hr after thermal injury COP, was reduced to 18.3 + 0.8 mm Hg in animals that were not resuscitated (P < 0.05). Following resuscitation by lactated Ringer’s, COP,, decreased markedly: within 3 hr COP, was reduced by 7-8 mm Hg to 12.3 +- 0.3 mm Hg (P < 0.05). In plasma-resuscitated animals COP, increased from 21.3 to 22.4 mm Hg over the study period (P < 0.05). In control COP, was 13.7 + 0.4 mm Hg (grand mean). There was a moderate reduction in COP, over the observation period even in the control group (P < 0.05). Following thermal injury there was a marked reduction in COP,, most evident following fluid replacement by lactated Ringer’s (P < 0.05 vs control group) (Table 1). In the untreated animals and following plasma infusion, COP, declined more than in the controls, although the combined effect of treatment and time did not reach a statistically significant difference (P > 0.05 compared to the control group, by repeated measures ANOVA). The transcapillary oncotic gradient (COP,,,) in thermally injured skin was calculated as the difference between colloid osmotic pressure in plasma and that in interstitial fluid (COP, - COP,3 (Table 1; Fig. 1). Differences in COPpad between groups were moderate in control (Table 1). With time COPwad increased slightly
VOL.
50, NO. 3, MARCH
1991
in the control group (P < 0.05). Following infusion of rat plasma, COPmad increased by 4-5 mm Hg (P < 0.05 vs control group). A different pattern was observed following resuscitation by lactated Ringer’s, which led to an almost abolished COPpad (P -C 0.05) (Fig. 1). Infusion of lactated Ringer’s to noninjured rats (the infusion only group) significantly reduced COP, (P < 0.05 vs control group) (Table 1). Still, COP, remained higher in the infusion only group compared to that after thermal injury followed by lactated Ringer’s, administered at the same infusion rate (P < 0.05) (Fig. 1). Contrary to results in the thermally injured rats, infusion of lactated Ringer’s to noninjured rats (infusion only group) did not alter COP,,,, since both COP,, and COP, were reduced to the same extent (Table 1). HCT was stable around 50-51% in the control group. Following nonresuscitated thermal injury, HCT increased from 51 to 59% within 1 hr and remained elevated at 3 hr after injury. HCT was transiently elevated (by around 3%) 1 hr postinjury in both the lactated Ringer’s and the plasma groups, but returned toward control values 3 hr postinjury in both groups. When COP was measured in sera or in wick fluid from noninjured skin, a stable equilibrium pressure was rapidly obtained (see Fig. 2). No “overshoot” (transient positive pressure peak) was observed after the membrane was washed with saline, indicating that normally, larger molecules did not leak through the membrane. Samples of wick fluid collected after thermal injury consistently gave an initial, negative pressure deflection before a flat equilibrium pressure was obtained (Fig. 2). Further, an “overshoot” was observed after the membrane was washed with saline (Fig. 2). Such changes were not observed in the plasma samples. The effect of osmometer membrane “pore size” (cutoff level) was evaluated for two different osmometer membranes. The Amicon YMlO membrane (used throughout the present study) was compared to the Amicon PM30 membrane (30,000 mol wt cutoff), which has been used in a number of previous studies [4,16,21]. The YMlO membrane gave l-2 mm Hg higher values for COP around 20 mm Hg and 1 mm Hg higher values for COP around 10 mm Hg. Equilibration of small ions over the osmometer membrane was evaluated in separate experiments: 0.5 M NaCl gave a very transient pressure response when applied on the osmometer membrane. Sodium citrate (citric acid: mol wt 192) gave a more pronounced, still transient pressure response (Fig. 2). Pifwas slightly subatmospheric (-2.0 f 0.2 mm Hg) in control (Fig. 3). Following thermal injury Pi, rapidly declined to -20 to -40 mm Hg (Fig. 3). The most negative values were observed in the first lo- to 20-min intervals. Following fluid infusion, Pi, returned toward control values and later remained slightly positive. Pi, became positive with plasma infusion sooner than it did following infusion of lactated Ringer’s (P G O.OEi)(Fig. 3).
ONARHEIM
AND
REED:
TRANSCAPILLARY
ONCOTIC
TABLE Colloid
Osmotic
Pressures
in Plasma (COP,), in Controls
GRADIENT
275
IN BURNS
1
Interstitial Fluid (COP,&, and Colloid and following Thermal Injury
Osmotic
Gradient
(COP,,)
Burn
Control COP, (mm Hg) Control 60 min 180 min COP, (mm Hg) Control 60 min 180 min COPgrad (mm
Lactated Ringer’s
No fluid
19.7 +- 0.5
18.3 f 0.8’.+
15.0 f 0.6*.+
20.9 + 0.9
19.0 + 0.7
12.3 + 0.3*-+
21.3 -r- 1.0 20.7 f 1.0 22.4 f 0.9*
12.7 k 0.7 11.5 f 0.7* 10.9 f 0.9*
13.7 +- 0.4 12.5 -t 0.6 10.4 + 0.5*
13.6 + 0.8 12.7 + 0.7
14.9 f 0.5+ 13.4 f 0.5*
9.7 zk 0.7*
11.3 ‘- 0.8*
8.4 of- 0.3*-+
8.4 + 0.6 9.2 f 0.4*
6.4 -c 0.7 5.8 + 0.6+ 8.5 + 0.8
6.4 f 0.9 7.3 iz 0.8
7.4 f 1.7 6.1 + 0.8+ 7.7 f 0.7
DISCUSSION
The wick method was developed to collect interstitial fluid for measurements of protein concentration and colloid osmotic pressure [6,17]. Wick fluid was found to give a correct estimate of COP, in subcutaneous tissue provided that sufficient proteins were added to the wick before or after insertion [ 171.
20
E
15
5
10
u 5 0
25
13.6 _’ 1.7 11.5 + 0.9
20.1 +- 0.7
20.8 k 0.9
10.0 zk 0.4*
0,
21.0 + 0.6 17.7 2 0.3*,+ 16.1 f 0.8*!+
21.1 + 0.8
6.0 k 0.8+ 2.3 3~ 0.9*,+ 2.6 f 0.7*~’
Note. Infusion only group: no thermal injury, lactated Ringer’s infused at the same rate as thermal Means k 1 SEM, n = 8. * P < 0.05 vs prehurn. + P < 0.05 vs control group (by ANOVA and t test).
I
Infusion only
I-k)
Control 60 min 180 min
25
Plasma
li!iL Lactated
Ringer’s
r
Plasma
C
Control
60
180
Control
60
180
FIG. 1. Colloid osmotic pressure in plasma (COP,) (open bars) and interstitial fluid (COP,) (solid bars) in control and 60 and 180 min after burn injury. Colloid osmotic gradient (COP,,) corresponds to the difference between COP, and COP,. Means f SEM, n = 8.
injury, resuscitated
by lactated Ringer’s.
Two recent modifications of the wick technique have suggested that saline-soaked wicks might give a certain underestimate of COP, [21, 221. However, “the crossover method” [22] requires insertion of at least three wicks for each determination of COP,,, which will limit its application for repeated measurements in a small animal. “Dry wicks” [21, 221 could have been used, still, saline-soaked wicks were preferred to allow for repeated measures and for comparisons to previous studies [4,
No fluid
Control group
11.1 f 1.2*
101.
Fluid replacement seems to have been sufficient since HCT was only transiently elevated in the resuscitated cases. On the basis of the alterations in HCT (corrected for blood samples), we predict that 3 hr postinjury plasma volume was unchanged from control in plasmaresuscitated animals and reduced by 5% in animals that received infusion of lactated Ringer’s. The Amicon YMlO membrane was used on the osmometer; this membrane has a pore diameter which should safely exclude all plasma proteins. Wick fluid from thermally injured skin gave an initial, negative pressure deflection on the osmometer prior to a stable equilibrium pressure (Fig. 2). A transient positive pressure peak occurred following washing of the membrane. Analogue observations were made following application of hypertonic saline, titrated rat plasma, or sodium citrate on the osmometer (Fig. 2), while no such phenomena were seen in wick fluid from noninjured skin. These observations suggest that after thermal injury molecules of molecular weights below the cutoff level for the membrane leaked through the intact osmometer membrane. This is in accordance with previous suggestions that
276
JOURNAL
OF SURGICAL
RESEARCH:
FIG. 2. Pressure recordings with colloid osmometer. (Top) Normal pressure tracing (flat equilibrium) obtained after plasma or wick fluid application. (Middle) Wick fluid from thermally injured skin area demonstrates an initial, brief pressure deflection following application of sample before a normal (flat) equilibrium was reached, a slight “overshoot” is seen following washing of membrane. (Bottom) Application of sodium citrate: transient pressure response; note marked “overshoot” following washing of membrane (due to small molecules leaking through membrane). D, drying of the membrane; A, application of sample; S, washing of membrane with 0.9% saline. For details, see Results and Discussion.
thermal injury (by breakdown of larger molecules) may generate “new” osmoles [2, 131. For control sera the YMlO membrane gave l-2 mm Hg higher COP values than the Amicon PM30 membrane (used in a number of previous studies) [4, 16, 21, 221. This may in part explain the somewhat higher COP, in control (13.7 + 0.4 mm Hg) compared to that in some previous studies (around 9-11 mm Hg; [4, 16, 21, 221). COP in plasma remained stable in the control group (Table 1). Nonresuscitated thermal injury was followed by a l-2 mm Hg reduction in COP, (Fig. l), which is somewhat less than a 2-6 mm Hg reduction observed in a previous study where no fluid therapy was provided [4]. A marked reduction of COP, was observed following lactated Ringer’s infusion: COP, was reduced to 12.3 + 0.3 mm Hg at 3 hr postinjury (P < 0.05) (Fig. 1). In thermally injured skin the extravasation rate of plasma proteins may increase to lo-50 times that of controls [ 1, 4,5], leading to a marked reduction of the intravascular pool of plasma proteins. COP, will be reduced when the remaining plasma proteins are diluted following massive saline infusion. Thus COP, was lower after thermal injury followed by infusion of lactated Ringer’s than following infusion of the same volume of Ringer’s to noninjured animals (Table 1).
VOL.
50, NO. 3, MARCH
1991
There was a certain reduction in COP,, with time even in the control experiments (12.7 vs 10.9 mm Hg; P < 0.05). When wick techniques are used to collect interstitial fluid it may therefore be relevant to specify the timing of wick fluid samples in relation to the induction of anesthesia. In the present study COP, was reduced by 3-4 mm Hg after thermal injury (P < 0.05). In extensively burned patients, resuscitated mainly by acetated Ringer’s, COP, in injured skin was reduced compared to that in normals [lo]. A significant reduction of COP, in severely burned patients was also found 4-8 hr after thermal injury, using a wick catheter technique [ 111. In contrast, in rats COP,, was not significantly altered following larger (40% BSA), nonresuscitated burns, whereas following smaller (10% BSA) burns even a certain increase of COP, was observed [4]. COP,r.¶d was 6-8 mm Hg in the control situation, which compares well to previous studies in the rat [4, 161. COPGad was not significantly altered after thermal injury in the nonresuscitated cases (Table 1). This contrasts previous observations that COP,,, was nullified or even reversed after nonresuscitated thermal injury [4]. However, in that study experiments were not performed within a humidified chamber [4]. Evaporative fluid losses from wicks handled in the open will rapidly increase COP, [ 161 and may lead to erroneously low values for COP,,,. In the present study handling of wicks solely within an incubator eliminated that problem: in recovery studies wick fluid from wicks soaked in serum gave the same COP as the primer. Following volume replacement by lactated Ringer’s, COP,,, declined to 2.6 +- 0.7 mm Hg (P < 0.05), which still is significantly different from a zero gradient (P
0”
A-A
-20.
,j
I’\{
,~.
Lactated Ringer’s
;-O,Unresuyitated, 8
I C
2 1 limo (hours) ofter thermal injury
3
FIG. 3. Interstitial fluid hydrostatic pressure (P,) measured by micropuncture in control (C) and following thermal injury. Fluid therapy by lactated Ringer’s or plasma was started 15 min postinjury. Means + SEM. For comparison data from Lund et al. [12] for Ps following a similar, nonresuscitated injury are also included.
ONARHEIM
AND
REED:
TRANSCAPILLARY
< 0.05). This is a more moderate reduction than that observed in severely burned patients, where COP,, was found to be equal to or even higher than COP, when measured 6-12 hr postinjury [lo]. However, it corresponds well to previous estimates of COP,,, after thermal injury in sheep resuscitated by lactated Ringer’s, using lymph as a representative for interstitial fluid in injured skin [3]. Fluid transport across the capillary wall is described by the Starling equation, J, = CFC. (P, - Pi, - 0. (COP,, - COPif)),
where J, is the volume of fluid filtered across the capillary wall, CFC is the capillary filtration coefficient, PC is the capillary hydrostatic pressure, P, is the interstitial fluid hydrostatic pressure, and u is the capillary reflection coefficient for plasma proteins. Normally the net filtration pressure (the net pressure imbalance across the capillary wall) is on the order of 0.5 mm Hg. For methodological reasons Pi, could only be measured in dermis; Pi, in dermis may not necessarily be representative for whole skin (including both dermis and subcutis) [ 121. Still, following thermal injury P, decreased from a normal value of -2 mm Hg to values around -20 to -40 mm Hg (Fig. 3). The markedly negative Pi, in injured skin may increase the net filtration pressure up to two orders of magnitude and will be the main driving force for edema formation in the initial stages. After the first 30-min period postinjury, the negative Pi, subsided (Fig. 3). From 1 hr postinjury Pi, was around ambient pressure or became slightly positive, as is seen in other edematous states [ZO]. When compared to data from Lund et al. [ 121 of Pi, following nonresuscitated thermal injury (Fig. 3), normal Pi, in the present study was reached sooner with resuscitation, and more promptly following plasma infusion (P < 0.05 vs lactated Ringer’s). The gradient in colloid osmotic pressure between plasma and the interstitium will influence the net transcapillary fluid transport: a reduced effective . COP,,d) will favor edema formation. After COP,db thermal injury, u has been estimated to be reduced from 0.87 to 0.45 [23]; this may further contribute to a reduced effective COP,, . When Pif is positive, PC (capillary hydrostatic pressure) and COPsad will be the driving forces for fluid filtration. A reduced COP,,, may facilitate edema formation also in noninjured tissue. Actually, fluid replacement by infusion of lactated Ringer’s after thermal injury was found to increase fluid accumulation in noninjured tissue [3, 41. This is to our knowledge the first study where colloid osmotic pressure has been determined directly in interstitial fluid following different forms of parenteral fluid substitution. Plasma and lactated Ringer’s had opposite
ONCOTIC
GRADIENT
IN BURNS
277
effects on COP,,, after a standardized thermal injury. Plasma maintained, or actually increased, COP,,,,, whereas lactated Ringer’s almost annihilated COP,,, . Saline solutions (without colloids) have been preferred for the initial fluid therapy in burns. The present study underlines that this will reduce both COP, and COP,,, and may therefore favor edema formation. ACKNOWLEDGMENTS We are grateful to Eli Gunn Kjarlaug and Malvin Gismervik for technical assistance. The study received financial support from The Norwegian Research Council for Science and the Humanities and The Norwegian Council on Cardiovascular Diseases.
REFERENCES 1. Leape, L. L. Kinetics of burn edema formation in primates. Ann. Surg. 176:223,1972. 2. Arturson, G. Microvascular permeability to macromolecules in thermal
injury. Acta Physiol. Stand. Suppl. 463: 111, 1979. R. H., Kramer, G., and Harms, B. Role of thermal injury-induced hypoproteinemia on fluid flux and protein permeability in burned and nonburned tissue. Surgery 95: 136,1984. Lund, T., and Reed, R. K. Microvascular fluid exchange following thermal skin injury in the rat: Changes in extravascular colloid osmotic pressure, albumin mass, and water content. Circ. Shock 20: 91, 1986. Onarheim, H., Lund, T., and Reed, R. Thermal skin injury. II. Effects on edema formation and albumin extravasation of fluid resuscitation with lactated Ringer’s, plasma, and hypertonic saline (2,400 mosmol/l) in the rat. Circ. Shock 27: 25, 1989. Aukland, K., and Fadnes, H. 0. Protein concentration of interstitial fluid collected from rat skin by a wick method. Acta Physiol. Stand. 88: 350, 1973. Noddeland, H., Omvik, P., Lund-Johansen, P., Ofstad, J., and Aukland, K. Interstitial colloid osmotic and hydrostatic pressures in human subcutaneous tissue during early stages of heart failure. Clin. Physiol. 4: 283, 1984. Koomans, H. A., Kortlandt, W., Geers, A. B., and Mees, E. J. D. Lowered protein content of tissue fluid in patients with the nephrotic syndrome: Observations during disease and recovery. Nephron 40: 391, 1985. Rein, K. A., Semb, K., Myhre, H. O., Levang, 0. W., Christensen, O., Stenseth, R., and Sande, E. Transcapillary fluid balance in subcutaneous tissue of patients undergoing aortocoronary bypass with extracorporeal circulation. Stand. J. Thorac. Cardiovmc. Surg. 22: 267, 1988. Pitklnen, J., Lund, T., Aanderud, L., and Reed, R. K. Transcapillary colloid osmotic pressures in injured and non-injured skin of seriously burned patients. Burns Incl. Therm. Znj. 13: 198, 1987. Wachtel, T. L., Frank, H. A., Sanders, R., Hargens, A. R., and Peters, R. M. Definition of the Starling forces with wick catheter in burned patients. J. Burn Care Rehabil. 4: 331, 1983. Lund, T., Wiig, H., and Reed, R. K. Acute postburn edema: Role of strongly negative interstitial Auid pressure. Amer. J. Physiol. 255 (Heart Circ. Physiol. 24): H1069, 1988. Lund, T., Onarheim, H., Wiig, H., and Reed, R. K. Mechanisms behind increased dermal imbibition pressure in acute burn edema. Amer. J. Physiol. 256 (Heart Circ. Physiol. 25): H940, 1989.
3. Demling,
4.
5.
6.
7.
8.
9.
10.
11.
12. 13.
278 14. 15.
16.
JOURNAL
OF SURGICAL
RESEARCH:
Walker, H. L., and Mason, A. D. A standard animal burn. J. Trauma 8: 1049,1968. Herndon, D. N., Wilmore, D. W., and Mason, A. D. Development and analysis of a small animal model simulating the human postburn hypermetabolic response. J. Surg. Res. 25: 394, 1978. Heir, S., and Wiig, H. Subcutaneous interstitial fluid colloid osmotic pressure in dehydrated rats. Actu Physiol. Stand. 133:
20. 21.
365,1988. 17.
18. 19.
Fadnes, H. O., and Aukland, K. Protein concentration and colloid osmotic pressure of interstitial fluid collected by the wick technique. Microuasc. Res. 14: 11, 1977. Aukland, K., and Johnsen, H. M. A colloid osmometer for small 6uid samples. Acta Physiol. &and. SO: 485, 1974. Wiederhielm, C. A., Woodbury, J. W., Kirk, S., and Rushmer,
22.
23.
VOL.
50, NO. 3, MARCH
1991
R. F. Pulsatile pressures in the microcirculation of frog’s mesentery. Amer. J. Physiol. 207: 173, 1964. Wiig, H., and Reed, R. K. Compliance of the interstitial space in rats. II. Studies on skin. Acta Physiol. Stand. 113: 30'7, 1981. Wiig, H., Heir, S., and Aukland, K. Colloid osmotic pressure of interstitial fluid in rat subcutis and skeletal muscle: Comparison of various wick sampling techniques. Acta Physiol. &and. 133: 167, 1988. Kramer, G. C., Sibley, L., Aukland, K., and Renkin, E. M. Wick sampling of interstitial fluid in rat skin: Further analysis and modifications of the method. Microvase. Res. 32: 39, 1986. Pitt, R. M., Parker, J. C., Jurkovich, G. J., Taylor, A. E., and Curreri, P. W. Analysis of altered capillary pressure and permeability after thermal injury. J. Surg. Res. 42: 693, 1987.