Favorable effects of hyperosmotic on myocardial edema and infarct
reperfusion size
D. GARCIA-DORADO, P. THEROUX, R. MUNOZ, J. ALONSO, J. ELIZAGA, F. FERNANDEZ-AVILES, J. BOTAS, J. SOLARES, J. SORIANO, AND J. M. DURAN Servicio de Cardiologia, Departamento de Medicina Experimental, Hospital General Gregorio-Maranon, Madrid 28007, Spain; and Montreal Heart Institute, Montreal, Quebec Hl T 1 C8, Canada Garcia-Dorado, D., P. Thbroux, R. Munoz, J. Alonso, J. Elizaga, F. Fernandez-Aviks, J. Botas, J. Solares, J. Soriano, and J. M. Duran. Favorable effects of hyperosmotic reperfusion on myocardial edema and infarct size. Am. J. Physiol. 262 (Heart Circ. PhysioL. 31): Hl7-H22, 1992.-Myocardial water content and infarct size were studied in 39 pigs randomly assigned to a nonintervention group, a group with an intracoronary infusion of a control solution, and a group with a hyperosmotic infusion to 450 mosM by the addition of Dmannitol. The intracoronary solutions were selectively infused into the left anterior descending coronary artery just distal to the occlusion site starting 48 min after occlusion. Reperfusion was performed 3 min later and the infusion rate progressively tapered off over the following 33 min. Multiple myocardial fragments were then obtained in nine pigs, from endocardial, mesocardial, and epicardial regions of the ischemic and control myocardium. Water content measured after 48 h of dessication was significantly greater in the reperfused [530 t 7 ml/100 (mean t SE) g dry wt] compared with control myocardium (374 t 3; P < 0.0001) and similar in reperfused control and isotonic infusion groups (556 t 7 and 543 t 8 ml/l00 g dry wt); it was 491 k 11 with intracoronary D-mannitol infusion, representing 35% less increase (P < 0.001). In the 30 remaining pigs, area at risk and infarct size were measured 24 h later by in vivo fluorescein and in vitro triphenyltetrazolium chloride. Infarct size was similar in control and in the isotonic reperfused hearts, 6.80 2 1.05 and 6.22 t 0.76% of ventricular weight, and smaller with D-mannitol, 4.46 & 0.46 (P < 0.05). The ratio of infarct size to area at risk was also smaller [0.415 t 0.029 vs. 0.543 t 0.052 and 0.547 t 0.045 (P c 0.02)]. Thus hyperosmotic reperfusion with mannitol significantly reduces early myocardial edema associated with coronary occlusion and reperfusion and also infarct size. mannitol; myocardial water content; area at risk; infarct coronary occlusion; coronary reperfusion
size;
RESTORES OXYGEN SUPPLY to the ischemic myocardium and, when applied early, results in significant myocardial salvage. However, some of the changes induced by reperfusion itself may be deleterious for the ischemic cells and precipitate cell necrosis (14). Various attempts have been made to protect the ischemic myocardium and increase the benefit derived from reperfusion. Calcium antagonist drugs administered during coronary occlusion and reperfusion possesssuch potential (7). Antifree radical agents have also been tested but have yielded controversial results (3, 4, 18). One of the mechanisms leading to cell death during ischemia is loss of water regulation, osmotic cell swelling, and disruption of the sarcolemma (16,20). These critical prelethal changes are further exaggerated on reperfusion (9) and may also contribute to the no-reflow phenomenon (22). Previous in vitro and in vivo studies have shown that increasing the osmolarity of the reperfusion REPERFUSION
0363-6135/92
$2.00 Copyright
media resulted in improved left ventricular function (15, 22), increased coronary blood flow (l&22), reduced myocardial water retention (15, 21), and better cell morphology (15, 21, 22). The purpose of this study was to test the hypothesis that a hyperosmotic reperfusion could also reduce infarct size, to indirectly document the role of myocardial cell edema per se in cell necrosis, and to determine whether its prevention could represent a rewarding approach to myocardial cell preservation during ischemia and reperfusion. Two previous studies designed at answering this question have yielded contradictory results (10, 11). In the present study, a pig heart model was used to obviate any confounding effect of collateral blood flow, and the hyperosmotic solution was selectively infused into the ischemic area to achieve high local intramyocardial osmotic pressure levels matching those observed in severely ischemic myocytes while avoiding systemic toxicity. METHODS Animal preparation. The animal preparation used has been previously described (5-8). Sixty-two large white pigs weighing 33-42 kg were premeditated with diazepam (5 mg) and azaperone (5 mg/kg), anesthetized with thiopental sodium (300-500 mg iv), intubated, and mechanically ventilated (Boyle Modular, Medishield, Hatfield, UK). Anesthesia was maintained with a continuous infusion of thiopental at the rate needed to prevent spontaneous respiratory movements. The right femoral artery and vein were dissected and 5-Fr Tygon catheters advanced into the inferior vena cava and descending aorta. A midline sternotomy was performed and the pericardium opened. The left anterior descending coronary artery was dissected free as close as possible to the midpoint of its total length and surrounded by a Ti Crown no. 2 filament. This filament was passed through a segment of Nelaton tubing to form a snare, which was used to close or open the artery. Heparin sodium (160 IU/kg) was then administered intravenously as a single bolus injection and lidocaine (1 mg/kg) was given. Afterwards, a Judkins 8-Fr guiding catheter for right coronary artery catheterization was immediately introduced into the main left coronary artery via the right carotid artery. A 4-Fr Rentrop intracoronary catheter was advanced through this guiding catheter into the distal left anterior descending artery. The guiding catheter was then pulled back to the central aorta, and the Rentrop catheter was slowly withdrawn to position its tip 5 mm distal to the dissection site, taking care that no collateral branch existed between its tip and the coronary occlusion site. This procedure, performed under visual inspection with continuous electrocardiographic and pressure monitoring, was usually completed in less than 30 s and in no cases required more than 3 min. Studyprotocol. The study design included three study groups. In the control group, reperfusion was performed after 48 min by releasing the snare surrounding the left anterior descending
0 1992 the American
Physiological
Society
H17
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
H18
HYPEROSMOTIC
coronary artery without any other intervention. In the two other groups, an infusion was administered directly into the area at risk using the Rentrop catheter for 3 min before releasing the coronary obstruction and for 33 min, thereafter, for a total perfusion time of 36 min. The perfusion rate was 9 ml/ min for 6 min (3 min during occlusion and 3 during reperfusion), 6 ml/min for 10 min, and 3 ml/min for the following 20 min. A Harvard pump was used to assure constant delivery. The solutions for intracoronary infusion were prepared 30 min beforehand and kept in a bath at 37°C. Bovine albumin (Fraction V, Boehringer Mannheim, Mannheim, FRG; 50 g/l) was added to a solution of lactate Ringer to adjust the colloid osmotic pressure to 16 mmHg. This resulted in a solution with normal osmotic and oncotic pressure. Calcium chloride was added to attain the physiological range of free calcium concentration. The pH was equilibrated to 7.0 with 0.1 M NaOH. In one of the study groups, the osmolarity of the perfusate was raised to 450 mosM by adding 30 g/l of D-mannitol (Sigma, St. Louis, MO). The actual concentrations of Na’, K’, free and total Ca2’ of the solution, and the osmotic and oncotic pressure were checked in every experiment. Measured values were 140.9 $- 1.4 Na’, 5.6 t 0.1 K’, 2.72 t 0.1 Ca2’, 16.0 t 3 mmHgoncotic pressure (Onkometer BMT 921, Boehringer Ingelheim, Berlin, FRG), 276 t 4 mosM osmolarity for the isotonic solution, and 455 k 5 for the mannitol group (Fiske, Needham Heights, MA). The homogeneity of the transmural distribution of the intracoronary infusion was studied by the intracoronary injection of 25,000 microspheres of 15-25 pm in size labeled with technetium-99m from a technetium molybdenum source (0.075 mCi) (Mallinckrodt, Petten, Holland). The microspheres were added to the 27 ml of the solutions infused during the occlusion period. In 30 animals, the sternotomy was closed 45 min after reperfusion, leaving in place a pleuropericardial drain. All catheters were withdrawn and the animals were returned to their cages. Nine other randomly selected animals, 3 in each study group, were killed after 33 min of reperfusion to determine myocardial water content. Study monitoring. Hematocrit and serum content of Na+, K’, Ca2+,glucose,and creatinine, measuredat the beginning and at the end of each experiment, were unchanged. Arterial pH, Pao, and Paco, were closely monitored to adjust the ventilation parametersto keeppH between7.35and 7.42 and PCO~ between36 and 46 mmHg. Aortic blood pressurewas continuously monitored with a crystal quartz transducer (HewlettPackard 1290A, Hewlett-Packard, Andover, MA) and recorded with lead II of the electrocardiogram on a Hewlett-Packard 78308Apolygraph (Hewlett-Packard, Waltham, MA) and SiemensElema-12recorder (Siemens,Erlangen, FRG) at a paper speedof 10 mm/s during the occlusionperiod and of 25 mm/s during the first 45 min of reperfusion.Thesetracings wereused to identify ventricular arrhythmias occurring during occlusion, intracoronary infusion, and reperfusion. The number of ventricular premature beats per minute wascounted. Exclusions. Twenty-three of the 62 animals were excluded from analysisfor various reasons:9 in the control group, 7 in the normosmoticintracoronary infusion, and 7 in the mannitol infusion group. Nine animals (4 in the control group, 2 in the normosmotic, and 3 in the mannitol group) died the night following the first surgicalprocedure.Eleven others (4 in each of the control group and normosmoticinfusion group and 3 in the mannitol group) had reocclusion of the left anterior descendingcoronary artery at the site of the previous dissection and were excluded. One animal in the control group had an infarct present in the territory of the left circumflex coronary artery. One pig in the mannitol infusion group was also excluded at the first surgery becauseof spontaneousreperfusion induced by accidental withdrawal of the Rentrop catheter. The
REPERFUSION
last exclusion wasan animal in the normosmoticinfusion group with spontaneousventricular fibrillation at the time of reperfusion. Of the remaining 39 animals, 9 were usedfor analysis of water content, 3 from each of the study groups.The thirty other animalscompletedthe 24-h study period and wereevenly distributed among the three study groups. No animal was excluded after determination of area at risk and infarct size. The histopathological procedures were all performed blindly, and the investigators performing the surgical procedureswere alsoblinded to treatment allocation. Postmortem studies. Standard methodsfor determination of infarct size and areaat risk wereused.Twenty-four hours after the first surgical procedure, the animals were premeditated, anesthetized, and ventilated following the sameprocedure as for the initial surgery. The left anterior descendingcoronary artery was reoccludedand 5 ml of 10% fluorescein injected in the left atrium. Five secondslater, the heart was excised and immersedin cold saline. Patency of the ostium of the left main coronary artery, of the proximal segmentof the circumflex, and of left anterior descendingcoronary artery was assessed. After excision of the atria, great vessels,and epicardial fat, the heart was cut from base to apex into 6- to 8mm slices parallel to the atrioventricular groove. All the sliceswere individually weighedand photographedunder ultraviolet light at a peak wavelength of 3,560A using a Hewlett-Packard 187camera and Polaroid 687 black and white film (8). With this technique the normally perfused myocardium appearedbrilliant white, and the area at risk appearedblack, with sharply defined edges.To further define from the black backgroundthe epicardial and endocardial limits of the area at risk, a second photograph of each slice was obtained under room light. The sliceswere then incubated in a 1% triphenyltetrazolium chloride solution at 37°C until the necrotic area was sharply demarcated;the sliceswereagain photographedunder room light. The extent of the area at risk and of the infarct area were calculated by digital planimetry of the perfused, nonperfused, and necrotic areas of all slicesand expressedas a percent of total ventricular mass.The slice showing the most extensive area of necrosis was processedfor histological analysis and stained with hematoxylin eosin to validate in this study that negative tetrazolium chloride staining correspondedindeedto tissuenecrosis(8). Measurement of myocardial water content. In the group of nine animals allocated to water content measurement,the left anterior descending artery was reoccluded after 33 min of reperfusion. Fluoresceinwas then injected into the left atrium and the heart immediately excised, aspreviously described(5). The heart was then briefly immersedin melting salineice until arrested, and then cut in 8- to lo-mm slices.With the useof a surgical blade, transmural sectionswere obtained from two to three slicesof the area at risk and from one slice of the control area. These sectionswere cleaned of all epicardial and endocardial nonmyocardial tissueand rapidly subdividedinto three fragments of similar size representing the endocardial, mesocardial, and epicardial third of the left ventricular wall. All fragments obtained for a total of 110wereweighedimmediately before and again after 48 h of dessicationat 100°C (Mettler H3lAR, Mettler, Greifensee,Switzerland). Water content was calculated as percent of dry weight and assumingthe density of water to be 1, as gram per 100 gram dry weight using the formula [(fresh weight - dry weight)/dry weight] x 100 (9,20). Myocardial flow distribution. Transmural samplesof the reperfused myocardium and of control myocardium from the posterolateral wall of the left ventricle were obtained from all animals, including those killed after 33 min of reperfusion. These sampleswere also subdivided in endocardial, mesocardial, and epicardial thirds. Radioactivity counts (Kontron Gammamatic GMl) were expressedas counts per gram of tissue.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HYPEROSMOTIC
Statistical analyses. Hemodynamic parameters, data on blood flow distribution, and the average values for water content of all samples from the normal and ischemic areas of the same heart were analyzed by analysis of variance after having assessed the absence of any significant departure of the data from the normal distribution. For water content, a repeated measures factor considering the average of multiple samples in each of the three transmural layers was added to the analysis of variance to rule out any confounding effect of the site of the sampling. The significance of changes in heart rate and in blood pressure throughout the experiment was assessed by the Student’s t test for paired observations. The homogeneity of the area at risk in the three groups was assessed by the Kruskal-Wallis nonparametric test. Intergroup differences in infarct size and the ratio infarct size to area at risk and reperfusion arrhythmias were also investigated using nonparametric statistics. In a first step, the Jokhere-Sterptra free distribution test was performed to test the null hypothesis that the data were similar in the three study groups against the alternative hypothesis of a significant reduction in one or two treatment groups by ordered inequalities. In a second step, a free distribution test for multiple sample comparisons was used to assess the intergroup differences. Values are expressed as means of: SE. A P value ~0.05 was retained as significant. RESULTS
Hemodynamic data. Mean heart rate was similar in the three study groups before occlusion of the left anterior descending coronary artery: 91 t 11 beats/min in the control group, 85 t 5 in the normosmotic infusion group, and 86 t 5 in the mannitol infusion group. Mean blood pressure was also similar as 83 t 6, 72 t 2, and 73 t 11 mmHg, respectively. The only significant hemodynamic change observed throughout the study consisted in a 8.8 t 3.0 mmHg decrease in mean aortic blood pressure present in the three study groups during the occlusion period, before any intervention (P c 0.05, compared with preocclusion values). Heart rate was, however, unchanged at that time (0.2 t 4.9 beats/min variation). The intracoronary infusion before reperfusion resulted in a 0.4 t 4.5 beats/min increase in heart rate in the normosmotic group and a 4 t 4.4 beats/min decrease in the mannitol group [nonsignificant (NS)] with a nonsignificant fall of 7.1 t 4.0 and 0.1 t 1.6 mmHg, respectively, in blood pressure. The hemodynamic data were less stable during the reperfusion period because of the presence of reperfusion arrhythmias. When assessedduring periods of sinus rhythm, heart rate was 90.4 t 7.5 beats/min in the control group, 94.7 t 7.9 in the normosmotic group, and 82.9 t 3.6 in the mannitol group. Mean blood pressure was 73.5 t 4.0, 66.6 t 5.9, and 68.3 -+ 11.2 mmHg (NS), respectively. The intra-arterial catheters were removed after 45 min of reperfusion, and subsequent blood pressure readings were not obtained. Ventricular premature beats occurred at a rate of 1.3 -+ 0.7 beats/min during the late occlusion period. The intracoronary infusion during the occlusion period resulted in a significant increase in the rate of ventricular premature beats to 67 t 20 beats/min in the normosmotic group. No such increase was however observed with the mannitol infusion, and the number of premature beats remained 7 t 5 beats/min as in the control group
H19
REPERFUSION
during the last 3 min of occlusion before reperfusion. c oronary reperfusion triggered a similar number of ventricular premature beats in the three study groups (Fig. 1)
Intramyocardial blood flow distribution. Microspheres injected along with the intracoronary infusion during the occlusion period distributed almost exclusively within the area at risk with only 0.5% of the total activity outside this area. The distribution of the infusate within the area at risk was in the subendocardial third 0.83 t 0.37 ml min-l l g-l of tissue in the normosmotic infusion group and 0.65 t 0.40 in the mannitol group (NS); in midwall, 0.85 t 0.38 and 0.58 t 0.31, respectively (NS); and in the outer third 0.85 t 0.33 and 1.37 t 0.33 ml. min. g-l (NS). The se flgures refer to the infusate flow during the last 3 min of coronary occlusion. Myocardial water content. Water content was measured by dessication of multiple fragments of the normal and reperfused myocardium in nine animals, three from each study group. Water content of the normal nonischemic myocardium averaged 374 t 3 ml/l00 g dry wt and was the same in the control group (364 t 6 ml/100 g dry wt), in the normosmotic (381 t 2 ml/l00 g dry wt), and in the mannitol group (378 t 7 ml/100 g dry wt). The reperfused myocardium showed an excess water content with average values of 530 t 7 ml/100 g dry wt in the three groups (P c 0.0001) with significant intergroup difference (P c 0.001). This intergroup difference in reperfused animals reflected lower values in the mannitol group [491 t 11 ml/l00 g dry wt (P < O.OOl)] compared with each of the two other study groups, which showed similar water content (556 t 7 ml/l00 g dry wt) in control and (543 t 8 ml/100 g dry wt) in normosmotic group, P = NS (Fig. 2). Reperfusion thus resulted in a 47% excess in total water content; with mannitol the excess was reduced by 35%. No transmural gradient in water content was present in any of the groups. Mean water content in the normal myocardium was 383 t 6,372 t 5, and 369 t 6 ml/100 g l
T 150
1 0 Control
-3 -2 -1 intracoronarv
0
1
sinfusion reperfusion
Time
2
3
+ b
(minutes)
Fig. 1. Number of ventricular premature beats during late occlusion and early reperfusion in 3 study groups. In control group, no intracoronary infusion was performed and arrhythmias occurred during early reperfusion. Intracoronary infusion of normosmotic infusion triggered reperfusion arrhythmias before release of coronary occlusion. Mannitol infusion did not induce such arrhythmias, which were delayed until upon coronary reperfusion as in control group. *P < 0.05.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
H20
HYPEROSMOTIC 600
REPERFUSION
11.42 t 0.817 o in the normosmotic solution, and 10.59 t 0.67% in the hyperosmotic group. Infarct size measured after 24 h by the triphenyltetrazolium technique was also the same in the control and normosmotic groups (6.80 t 1.05 and 6.22 t 0.76% of ventricular mass, NS) but significantly less in the mannitol group (4.47 t 0.46; P < 0.02). Similarly, the ratio infarct size to area at risk was similar in control and normosmotic groups (0.543 t 0.052 and 0.547 t 0.045) and significantly less (0.415 t 0.029) in the mannitol group (P < 0.001) for a 24% reduction in predicted infarct size (Fig. 2).
1
DISCUSSION l
Control
Normoosmotic infusion
Mannitol infusion
Fig. 2. Myocardial water content (top) and ratio of infarct size to area at risk (bottom) in 3 study groups. Water content and ratio are similar in control group and in normosmotic reperfusion groups. Mannitol added to intracoronary infusion significantly decreases both water content and predicted infarct size. *P < 0.02.
Table
1. Myocardial water content in control and reperfused myocardium Transmural
Sampling Site
Control
Control
Endocardial Mesocardial Epicardial Mean
372k3 362kll
Normosmotic Infusion
Mannitol Infusion
All
392t16 376tlO
383t6 372t5
rnyocardium 384k2 377t2
359t16
381G
367t9
369t6
364k6
381t2
378t7
374k3
Reperfused
rnyocardium
Endocardial 553t20 536t5 483k21 Mesocardial 56Okll 555&16 506t28 Epicardial 555t9 539&17 483k4 Mean 556t7 543t8 491Ik11* Values are means t SE. Myocardial water content expressed in ml of water per 100 g of dry weight. * P < 0.001 compared with reperfused myocardium of the control and normosmotic reperfusion groups. The transmural location of the myocardial fragments did not influence the results.
dry wt, respectively, in endo-, meso-, and epicardial sections. The mean of the reperfused myocardium of control and normoncotic infusion groups was 544 t 10, 557 t 9, and 547 t 9 (NS) ml/100 g dry wt compared with 483 t 21,506 t 28, and 483 t 4 ml/100 g dry w-t in the mannitol infusion group. Table 1 provides the water content calculated in transmural layers of the control nonischemic myocardium and of the reperfused myocardium in the three study groups. Infarct size, area at risk, and infarct size to area at risk ratio. The area at risk calculated from in vivo fluorescein
injection was of similar size in the three study groups: 12.3 t 1.19?’o o f ventricular mass in the control group,
This study demonstrates that coronary reperfusion with a hyperosmotic solution with mannitol can reduce early myocardial edema associated with coronary artery occlusion and reperfusion, favorably influence the ratio infarct size to area at risk and limit the ultimate infarct size. Comparison with previous studies. Reduction in cell swelling by a hyperosmotic solution has previously been demonstrated in experiments involving isolated hearts (15, 21). Studies have also shown that mannitol can improve blood flow to the ischemic area (15,22), preserve left ventricular function (15, 22), maintain the histological and ultrastructural aspect of the myocardium (15, 21, 22) and improve cellular electrophysiology (21). The consequences of these beneficial effects on infarct size per se have been tested, however, only in a few studies that have yielded controversial results. Kloner et al. (11) administered mannitol before and after a 40-min coronary occlusion in 29 dogs and reported a 50% reduction in necrosis measured in the posterior papillary muscle. More recently, Klein et al. (10) reperfused 21 pigs with an intracoronary infusion of mannitol at various concentrations after a 75-min coronary occlusion. Infarct size and myocardial water content, measured 2 h later, were not improved with the hyperosmotic solution. Nevertheless, higher ATP levels in the myocardium reperfused with higher mannitol concentration were found. A third study in dogs has shown a beneficial effect on recovery of function with administration of a hyperosmolar cardioplegic solution of mannitol and glucose; the concomitant use of diltiazem and several other substrates complicates the interpretation of the results of this study (15) In the present investigation, the hyperosmotic intracoronary infusion was administered directly into the ischemic area, starting 3 min before and maintained for 33 min during reperfusion. Myocardial edema observed during early reperfusion was reduced by 35%, infarct size after 24 h by 31%, and the ratio of infarct size to area at risk improved by 24% compared with control hearts with no intracoronary infusion and to hearts reperfused with a normosmotic infusion. The two groups with the intracoronary infusion differ from each other only by the presence or absence of mannitol in the intracoronary perfusate, ruling out a possible confounding effect of the infusion per se. Methodological considerations. The experimental model of coronary occlusion and reperfusion used in this
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HYPEROSMOTIC
study has previously been shown to produce infarcts of reproducible sizes and to be sensitive to interventions influencing the ratio of infarct size to area at risk (5, 7). Reproducibility of measures of infarct size is helped by the absence of collateral blood flow in the pig heart, the standard site of occlusion of the left anterior descending coronary artery, and also by the quantification of infarct size 24 h after reperfusion when the borders can be well delineated by the triphenyltetrazolium chloride technique. The correlation observed between this technique and histological analysis in this study was excellent. We previously also documented this correlation by a systematic analysis of the transmural extension of the infarct every millimeter in multiple myocardial slices (8). Results with the staining technique may be less reliable when performed earlier after coronary occlusion (17). The 45 to 55min occlusion period selected corresponds to a period of rapid growth of the infarct area within the area at risk and may represent a critical time to detect myocardial salvage (6, 19). This duration of occlusion also corresponds to the time required to achieve revascularization with the currently used fibrinolytic agents in humans. Direct intracoronary infusion was used to permit selective impregnation of the area at risk with a highly hyperosmolar solution corresponding to 1.5 times normal osmolarity, as opposed to previous studies, which have generally limited the hyperosmotic state to 20-60 mosM (1, 11). The intracoronary infusion of this hyperosmotic solution was well tolerated and did not induce any detectable hemodynamic change. This high concentration was selected to match previously described osmotic pressure of ischemic cells (20, 21) and the duration of the infusion to correspond to the duration of the transient ultrastructural changes described after reperfusion (9); the decreasing rate of the infusion was empirically selected to mimic the presumed progressive normalization of extracellular osmotic pressure (21). Estimation of flow by microspheres injected during coronary occlusion showed a rate of perfusion of 0.84 t 0.07 ml. min-’ . g tissue-’ in the normosmotic infusion group and 0.86 t 0.08 ml min-l . g tissue-l in the mannitol group (NS), near normal coronary blood flow. Reperfusion arrhythmias. Characterization of reperfusion arrhythmia was not a main goal of this study. However, the normosmotic infusion triggered reperfusion arrhythmias. These appeared before restoration of anterograde coronary blood flow. Such was not the use in the mannitol infusion group where arrhythmias occurred only after restoration of coronary blood flow as in the control group. Myocardial water content. The myocardial water content of the control group and of the group reperfused with the normosmotic infusion showed an excess water of 169 ml/100 g of dry wt or 48% compared with normal myocardium. Reperfusion with mannitol decreased this water excess by 35%. Mechanisms of action. The observations do not allow the extrapolation that reduction in infarct size is directly related to the decrease in myocardial edema. Water content was measured after 33 min of reperfusion and infarct size 24 h later, and the two observations could represent l
REPERFUSION
H21
independent findings. Nevertheless, water and sodium contents of the ischemic myocardium increase early after coronary occlusion (9) and tissue osmolarity measured after 75 min of ischemia ranges between 360 and 420 mosM (20, 21). The increase is initially more important in myocytes than in the extracellular space with an 8% excess of volume after 30 min of occlusion. Reperfusion results in a sudden restoration of extracellular osmolarity to normal values and generates an abrupt osmotic gradient that results in further cell swelling with additional water excess, which may represent 16% of cell volume. Cell swelling imposes mechanical overload to the sarcolemma leading to the formation of sarcolemma blebs, sarcolemmal damage, and disruption (20). Whereas aerobic myocytes can withstand increases up to 40% in cell volume without damage, an 8% increase in ischemic myocytes can result in sarcolemmal disruption (20). Other possible beneficial actions of mannitol could be washout of extracellular and intracellular toxic metabolites and prevention of the toxicity associated with rapid oxygen restoration. Mannitol possesses free radical scavenger properties inhibiting the hydroxyl radicals located extracellularly (13). A few studies have suggested that the benefits of mannitol could be explained by this scavenging mechanism (2, 12), based on the observation of an absence of an effect of hyperosmolar solutions other than mannitol. The agents used in these studies were however sodium and glucose (12) and both penetrate cell membrane, limiting the driving osmotic force, and possibly inducing other effects on cell physiology. Mannitol, on the other hand, acts extracellularly except in cells with altered permeability and irreversibly damaged (16, 20). In a study performed using the very same protocol, high doses of superoxide dismutase administered intracoronary from 3 min before to 33 min after reperfusion had no effect on infarct size (4). Mannitol is not expected to add much benefit to superoxide dismutase as an hydroxide radical scavenger, considering that the superoxide anion is required for the generation of the hydroxyl radical from hydrogen peroxide (17). A more direct demonstration that mannitol acts through a hyperosmotic mechanism rather than through an antioxidant effect would require the demonstration of the same effects with a pure hyperosmotic agent. In practice, such an agent is difficult to find: 1) it should be water soluble and have a relatively low molecular weight to allow important increase in osmotic pressure; 2) it must also be large enough to prevent its entry into the cell; and finally 3) such an agent should not accept electrons. Since most of these agents are maximally oxidized, they could behave as oxidants with paradoxically harmful effects. Clinical implications. Reperfusion with mannitol is associated with infarct size smaller than predicted from the area at risk, suggesting that the conditions of reperfusion could be important for true salvage of myocardial cells that would otherwise have died. This stresses the importance of continuous research to try to derive maximal benefit from pharmacological and mechanical recanalization of acutely occluded arteries in acute myocardial infarction in humans. Prevention of cell swelling may represent one of the promising approaches to pre-
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HZ2
HYPEROSMOTIC
serve ischemia-weakened sarcolemma, cell integrity, and cell-to-cell progression in myocardial necrosis. This approach could also be useful in other clinical conditions when cell preservation is important, such as complex coronary angioplasty, myocardial cardioplegia, and heart transplantation. The authors thank Lute Begin and Lise de Repentigny for the excellent secretarial work. This study was supported by Grant PA 86/0389 from the Comision Interministerial de Ciencia y Tecnologia and Grant l/89 of the Convenio Madrid-Quebec. Address for reprint requests: P. Theroux, Montreal Heart Institute, 5000 Belanger St., Montreal, Quebec HlT lC8, Canada. Received
2 November
1990; accepted
in final
form
14 August
1991.
REFERENCES 1. Apstein, C. S., F. N. Gravino, and C. C. Houdenschild. Determinants of a protective effect of glucose and insulin on the ischemic myocardium. Effects on contractile function, diastolic compliance, metabolism and ultrastructure during ischemia and reperfusion. Circ. Res. 52: 515-526, 1983. 2. England, M. D., C. C. Caravochi, J. F. O’Brien, E. Solis, J. R. Pluth, T. A. Orszulak, M. P. Kaye, and H. V. Schass. Influence of antioxidants (mannitol and allopurinol) on oxygen free radical generation during and after cardiopulmonary bypass. Circulation 74, Suppl. III: III-134-111137, 1986. 3. Engler, R., and E. Gilpin. Can superoxide dismutase alter myocardial infarct size? Circulation 79: 1137-1142, 1989. 4. Garcia-Dorado, D., P. Theroux, J. Alonso, J. Elizaga, J. Botas, F. Fernandez-AvilC?s, J. Soriano, R. Munoz, and J. Solares. Intracoronary infusion of superoxide dismutase and reperfusion injury in the pig. Basic Res. CardioZ. 85: 619-629, 1990. 5. Garcia-Dorado, D., P. Theroux, J. Elizaga, F. FernandezAviles, J. Alonso, and J. Solares. Influence of tachycardia and arterial hypertension on infarct size in the pig. Cardiouasc. Res. 22: 620-626,1988. 6. Garcia-Dorado, D., P. Theroux, J. Elizaga, M. Galinanes, J. Solares, M. Riesgo, M. J. Gomez, A. Garcia-Dorado, and F. Aviles. Myocardial reperfusion in the pig heart model: infarct size and duration of coronary occlusion. Cardiovasc. Res. 21: 537544, 1987. 7. Garcia-Dorado, D., P. Theroux, F. Fernandez-Aviles, J. Elizaga, J. Solares, and M. Galinanes. Diltiazem and progression of myocardial ischemic damage during coronary artery occlusion and reperfusion in porcine heart. J. Am. CoZl. Cardiol. 10: 906911, 1987. 8. Garcia-Dorado, D., P. Theroux, J. Solares, J. Alonso, F. Fernandez-Aviles, J. Elizaga, J. Soriano, J. Botas, and R. Munoz. Determinants of hemorrhagic infarcts. Histologic observations from experiments involving coronary occlusion, coronary reperfusion and reocclusion. Am. J. Pathol. 137: 301-311, 1990.
REPERFUSION 9. Jennings, R. B., J. Schaper, M. L. Hill, C. Steenbergen, and K. A. Reimer. Effect of reperfusion late in the phase of reversible ischemic injury. Changes in cell volume, electrolytes, metabolites and ultrastructure. Circ. Res. 56: 262-278, 1985. 10. Klein, H. H., K. Nebendahl, M. Schubothe, and H. Kreuzer. Intracoronary mannitol during reperfusion does not affect infarct size in ischemic, reperfused porcine hearts. Basic Res. Cardiol. 80: 251-259,1985. 11. Kloner, R. A., K. A. Reimer, J. T. Willerson, and R. B. Jennings. Reduction of experimental myocardial infarct size with hyperosmotic mannitol. Proc. Sot. Exp. Biol. A4ed. 151: 677-683, 1976. 12. Macgovern, G. J., S. F. Bolling, A. S. Carole, B. H. Bulkley, and T. J. Gardner. The mechanism of mannitol in reducing ischemic injury: hyperosmolarity or hydroxil scavenger. Circulation 70, Suppl. I: 1-91-I-95, 1984. 13. Maestro, R. F., H. H. Thaw, J. Bjork, M. Plonker, and K. E. Arfors. Free radicals as mediators of tissue injury. Acta Physiol. Stand. 492, Suppl. I: 43-57, 1980. 14. Nayler, W. G., and J. S. Elz. Reperfusion injury: laboratory artifact or clinical dilemma. Circulation 74: 215-221, 1986. 15. Okamoto, F., B. S. Allen, G. D. Buckberg, H. Young, H. Bugyi, and J. Leaf. Studies of controlled reperfusion after ischemia. Reperfusate composition: interaction of marked hyperglycemia and marked hyperosmolarity in allowing immediate contractile recovery after four hours of regional ischemia. J. Thorac. Cardiovasc. Surg. 92: 583-593, 1986. 16. Reimer, K. A., R. B. Jennings, and M. L. Hill. Total ischemia in dog hearts, in vitro. 2. High energy phosphate depletion and associated defects in energy metabolism cell volume regulation and sarcolemmal integrity. Circ. Res. 49: 901-911, 1981. 17. Reimer, K. A., C. E. Murry, and V. J. Richard. The role of neutrophils and free radicals in the ischemic-reperfused heart: why the confusion and controversy? J. Mol. Cell. Cardiol. 21: 12351239, 1989. 18. Richard, V. J., C. E. Murry, R. B. Jennings, and K. A. Reimer. Therapy to reduce free radicals during early reperfusion does not limit myocardial infarcts caused by 90 min of ischemia in dogs. Circulation 78: 473-480, 1988. 19. Schaper, W., K. Binz, S. Sass, and B. Winkler. Influence of collateral blood flow and variations in MV02 on tissue-ATP content in ischemic and infarcted myocardium. J. Mol. Cell. Cardiol. 19: 19-37, 1987. 20. Steenbergen, C., M. L. Hill, and R. B. Jennings. Volume regulation and plasma membrane injury in aerobic, anaerobic and ischemic myocardium in vitro. Effects of osmotic cell swelling on plasma membrane integrity. Circ. Res. 57: 864-875, 1985. 21. Tranum-Jensen, J., M. Janse, J. W. T. Fiolet, W. J. G. Krieger, C. N. D’Alnoncourt, and D. Durrer. Tissue osmolality, cell swelling and reperfusion in acute regional myocardial ischemia in the isolated porcine heart. Circ. Res. 49: 364-381, 1987. 22. Willerson, J. T., W. J. Powell, T. E. Guiney, J. J. Stark, C. A. Sanders, and A. Leaf. Improvement in myocardial function and coronary blood flow in ischemic myocardium after mannitol. J. Clin. Invest. 51: 2989-2998, 1972.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
reperfusion size
D. GARCIA-DORADO, P. THEROUX, R. MUNOZ, J. ALONSO, J. ELIZAGA, F. FERNANDEZ-AVILES, J. BOTAS, J. SOLARES, J. SORIANO, AND J. M. DURAN Servicio de Cardiologia, Departamento de Medicina Experimental, Hospital General Gregorio-Maranon, Madrid 28007, Spain; and Montreal Heart Institute, Montreal, Quebec Hl T 1 C8, Canada Garcia-Dorado, D., P. Thbroux, R. Munoz, J. Alonso, J. Elizaga, F. Fernandez-Aviks, J. Botas, J. Solares, J. Soriano, and J. M. Duran. Favorable effects of hyperosmotic reperfusion on myocardial edema and infarct size. Am. J. Physiol. 262 (Heart Circ. PhysioL. 31): Hl7-H22, 1992.-Myocardial water content and infarct size were studied in 39 pigs randomly assigned to a nonintervention group, a group with an intracoronary infusion of a control solution, and a group with a hyperosmotic infusion to 450 mosM by the addition of Dmannitol. The intracoronary solutions were selectively infused into the left anterior descending coronary artery just distal to the occlusion site starting 48 min after occlusion. Reperfusion was performed 3 min later and the infusion rate progressively tapered off over the following 33 min. Multiple myocardial fragments were then obtained in nine pigs, from endocardial, mesocardial, and epicardial regions of the ischemic and control myocardium. Water content measured after 48 h of dessication was significantly greater in the reperfused [530 t 7 ml/100 (mean t SE) g dry wt] compared with control myocardium (374 t 3; P < 0.0001) and similar in reperfused control and isotonic infusion groups (556 t 7 and 543 t 8 ml/l00 g dry wt); it was 491 k 11 with intracoronary D-mannitol infusion, representing 35% less increase (P < 0.001). In the 30 remaining pigs, area at risk and infarct size were measured 24 h later by in vivo fluorescein and in vitro triphenyltetrazolium chloride. Infarct size was similar in control and in the isotonic reperfused hearts, 6.80 2 1.05 and 6.22 t 0.76% of ventricular weight, and smaller with D-mannitol, 4.46 & 0.46 (P < 0.05). The ratio of infarct size to area at risk was also smaller [0.415 t 0.029 vs. 0.543 t 0.052 and 0.547 t 0.045 (P c 0.02)]. Thus hyperosmotic reperfusion with mannitol significantly reduces early myocardial edema associated with coronary occlusion and reperfusion and also infarct size. mannitol; myocardial water content; area at risk; infarct coronary occlusion; coronary reperfusion
size;
RESTORES OXYGEN SUPPLY to the ischemic myocardium and, when applied early, results in significant myocardial salvage. However, some of the changes induced by reperfusion itself may be deleterious for the ischemic cells and precipitate cell necrosis (14). Various attempts have been made to protect the ischemic myocardium and increase the benefit derived from reperfusion. Calcium antagonist drugs administered during coronary occlusion and reperfusion possesssuch potential (7). Antifree radical agents have also been tested but have yielded controversial results (3, 4, 18). One of the mechanisms leading to cell death during ischemia is loss of water regulation, osmotic cell swelling, and disruption of the sarcolemma (16,20). These critical prelethal changes are further exaggerated on reperfusion (9) and may also contribute to the no-reflow phenomenon (22). Previous in vitro and in vivo studies have shown that increasing the osmolarity of the reperfusion REPERFUSION
0363-6135/92
$2.00 Copyright
media resulted in improved left ventricular function (15, 22), increased coronary blood flow (l&22), reduced myocardial water retention (15, 21), and better cell morphology (15, 21, 22). The purpose of this study was to test the hypothesis that a hyperosmotic reperfusion could also reduce infarct size, to indirectly document the role of myocardial cell edema per se in cell necrosis, and to determine whether its prevention could represent a rewarding approach to myocardial cell preservation during ischemia and reperfusion. Two previous studies designed at answering this question have yielded contradictory results (10, 11). In the present study, a pig heart model was used to obviate any confounding effect of collateral blood flow, and the hyperosmotic solution was selectively infused into the ischemic area to achieve high local intramyocardial osmotic pressure levels matching those observed in severely ischemic myocytes while avoiding systemic toxicity. METHODS Animal preparation. The animal preparation used has been previously described (5-8). Sixty-two large white pigs weighing 33-42 kg were premeditated with diazepam (5 mg) and azaperone (5 mg/kg), anesthetized with thiopental sodium (300-500 mg iv), intubated, and mechanically ventilated (Boyle Modular, Medishield, Hatfield, UK). Anesthesia was maintained with a continuous infusion of thiopental at the rate needed to prevent spontaneous respiratory movements. The right femoral artery and vein were dissected and 5-Fr Tygon catheters advanced into the inferior vena cava and descending aorta. A midline sternotomy was performed and the pericardium opened. The left anterior descending coronary artery was dissected free as close as possible to the midpoint of its total length and surrounded by a Ti Crown no. 2 filament. This filament was passed through a segment of Nelaton tubing to form a snare, which was used to close or open the artery. Heparin sodium (160 IU/kg) was then administered intravenously as a single bolus injection and lidocaine (1 mg/kg) was given. Afterwards, a Judkins 8-Fr guiding catheter for right coronary artery catheterization was immediately introduced into the main left coronary artery via the right carotid artery. A 4-Fr Rentrop intracoronary catheter was advanced through this guiding catheter into the distal left anterior descending artery. The guiding catheter was then pulled back to the central aorta, and the Rentrop catheter was slowly withdrawn to position its tip 5 mm distal to the dissection site, taking care that no collateral branch existed between its tip and the coronary occlusion site. This procedure, performed under visual inspection with continuous electrocardiographic and pressure monitoring, was usually completed in less than 30 s and in no cases required more than 3 min. Studyprotocol. The study design included three study groups. In the control group, reperfusion was performed after 48 min by releasing the snare surrounding the left anterior descending
0 1992 the American
Physiological
Society
H17
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
H18
HYPEROSMOTIC
coronary artery without any other intervention. In the two other groups, an infusion was administered directly into the area at risk using the Rentrop catheter for 3 min before releasing the coronary obstruction and for 33 min, thereafter, for a total perfusion time of 36 min. The perfusion rate was 9 ml/ min for 6 min (3 min during occlusion and 3 during reperfusion), 6 ml/min for 10 min, and 3 ml/min for the following 20 min. A Harvard pump was used to assure constant delivery. The solutions for intracoronary infusion were prepared 30 min beforehand and kept in a bath at 37°C. Bovine albumin (Fraction V, Boehringer Mannheim, Mannheim, FRG; 50 g/l) was added to a solution of lactate Ringer to adjust the colloid osmotic pressure to 16 mmHg. This resulted in a solution with normal osmotic and oncotic pressure. Calcium chloride was added to attain the physiological range of free calcium concentration. The pH was equilibrated to 7.0 with 0.1 M NaOH. In one of the study groups, the osmolarity of the perfusate was raised to 450 mosM by adding 30 g/l of D-mannitol (Sigma, St. Louis, MO). The actual concentrations of Na’, K’, free and total Ca2’ of the solution, and the osmotic and oncotic pressure were checked in every experiment. Measured values were 140.9 $- 1.4 Na’, 5.6 t 0.1 K’, 2.72 t 0.1 Ca2’, 16.0 t 3 mmHgoncotic pressure (Onkometer BMT 921, Boehringer Ingelheim, Berlin, FRG), 276 t 4 mosM osmolarity for the isotonic solution, and 455 k 5 for the mannitol group (Fiske, Needham Heights, MA). The homogeneity of the transmural distribution of the intracoronary infusion was studied by the intracoronary injection of 25,000 microspheres of 15-25 pm in size labeled with technetium-99m from a technetium molybdenum source (0.075 mCi) (Mallinckrodt, Petten, Holland). The microspheres were added to the 27 ml of the solutions infused during the occlusion period. In 30 animals, the sternotomy was closed 45 min after reperfusion, leaving in place a pleuropericardial drain. All catheters were withdrawn and the animals were returned to their cages. Nine other randomly selected animals, 3 in each study group, were killed after 33 min of reperfusion to determine myocardial water content. Study monitoring. Hematocrit and serum content of Na+, K’, Ca2+,glucose,and creatinine, measuredat the beginning and at the end of each experiment, were unchanged. Arterial pH, Pao, and Paco, were closely monitored to adjust the ventilation parametersto keeppH between7.35and 7.42 and PCO~ between36 and 46 mmHg. Aortic blood pressurewas continuously monitored with a crystal quartz transducer (HewlettPackard 1290A, Hewlett-Packard, Andover, MA) and recorded with lead II of the electrocardiogram on a Hewlett-Packard 78308Apolygraph (Hewlett-Packard, Waltham, MA) and SiemensElema-12recorder (Siemens,Erlangen, FRG) at a paper speedof 10 mm/s during the occlusionperiod and of 25 mm/s during the first 45 min of reperfusion.Thesetracings wereused to identify ventricular arrhythmias occurring during occlusion, intracoronary infusion, and reperfusion. The number of ventricular premature beats per minute wascounted. Exclusions. Twenty-three of the 62 animals were excluded from analysisfor various reasons:9 in the control group, 7 in the normosmoticintracoronary infusion, and 7 in the mannitol infusion group. Nine animals (4 in the control group, 2 in the normosmotic, and 3 in the mannitol group) died the night following the first surgicalprocedure.Eleven others (4 in each of the control group and normosmoticinfusion group and 3 in the mannitol group) had reocclusion of the left anterior descendingcoronary artery at the site of the previous dissection and were excluded. One animal in the control group had an infarct present in the territory of the left circumflex coronary artery. One pig in the mannitol infusion group was also excluded at the first surgery becauseof spontaneousreperfusion induced by accidental withdrawal of the Rentrop catheter. The
REPERFUSION
last exclusion wasan animal in the normosmoticinfusion group with spontaneousventricular fibrillation at the time of reperfusion. Of the remaining 39 animals, 9 were usedfor analysis of water content, 3 from each of the study groups.The thirty other animalscompletedthe 24-h study period and wereevenly distributed among the three study groups. No animal was excluded after determination of area at risk and infarct size. The histopathological procedures were all performed blindly, and the investigators performing the surgical procedureswere alsoblinded to treatment allocation. Postmortem studies. Standard methodsfor determination of infarct size and areaat risk wereused.Twenty-four hours after the first surgical procedure, the animals were premeditated, anesthetized, and ventilated following the sameprocedure as for the initial surgery. The left anterior descendingcoronary artery was reoccludedand 5 ml of 10% fluorescein injected in the left atrium. Five secondslater, the heart was excised and immersedin cold saline. Patency of the ostium of the left main coronary artery, of the proximal segmentof the circumflex, and of left anterior descendingcoronary artery was assessed. After excision of the atria, great vessels,and epicardial fat, the heart was cut from base to apex into 6- to 8mm slices parallel to the atrioventricular groove. All the sliceswere individually weighedand photographedunder ultraviolet light at a peak wavelength of 3,560A using a Hewlett-Packard 187camera and Polaroid 687 black and white film (8). With this technique the normally perfused myocardium appearedbrilliant white, and the area at risk appearedblack, with sharply defined edges.To further define from the black backgroundthe epicardial and endocardial limits of the area at risk, a second photograph of each slice was obtained under room light. The sliceswere then incubated in a 1% triphenyltetrazolium chloride solution at 37°C until the necrotic area was sharply demarcated;the sliceswereagain photographedunder room light. The extent of the area at risk and of the infarct area were calculated by digital planimetry of the perfused, nonperfused, and necrotic areas of all slicesand expressedas a percent of total ventricular mass.The slice showing the most extensive area of necrosis was processedfor histological analysis and stained with hematoxylin eosin to validate in this study that negative tetrazolium chloride staining correspondedindeedto tissuenecrosis(8). Measurement of myocardial water content. In the group of nine animals allocated to water content measurement,the left anterior descending artery was reoccluded after 33 min of reperfusion. Fluoresceinwas then injected into the left atrium and the heart immediately excised, aspreviously described(5). The heart was then briefly immersedin melting salineice until arrested, and then cut in 8- to lo-mm slices.With the useof a surgical blade, transmural sectionswere obtained from two to three slicesof the area at risk and from one slice of the control area. These sectionswere cleaned of all epicardial and endocardial nonmyocardial tissueand rapidly subdividedinto three fragments of similar size representing the endocardial, mesocardial, and epicardial third of the left ventricular wall. All fragments obtained for a total of 110wereweighedimmediately before and again after 48 h of dessicationat 100°C (Mettler H3lAR, Mettler, Greifensee,Switzerland). Water content was calculated as percent of dry weight and assumingthe density of water to be 1, as gram per 100 gram dry weight using the formula [(fresh weight - dry weight)/dry weight] x 100 (9,20). Myocardial flow distribution. Transmural samplesof the reperfused myocardium and of control myocardium from the posterolateral wall of the left ventricle were obtained from all animals, including those killed after 33 min of reperfusion. These sampleswere also subdivided in endocardial, mesocardial, and epicardial thirds. Radioactivity counts (Kontron Gammamatic GMl) were expressedas counts per gram of tissue.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HYPEROSMOTIC
Statistical analyses. Hemodynamic parameters, data on blood flow distribution, and the average values for water content of all samples from the normal and ischemic areas of the same heart were analyzed by analysis of variance after having assessed the absence of any significant departure of the data from the normal distribution. For water content, a repeated measures factor considering the average of multiple samples in each of the three transmural layers was added to the analysis of variance to rule out any confounding effect of the site of the sampling. The significance of changes in heart rate and in blood pressure throughout the experiment was assessed by the Student’s t test for paired observations. The homogeneity of the area at risk in the three groups was assessed by the Kruskal-Wallis nonparametric test. Intergroup differences in infarct size and the ratio infarct size to area at risk and reperfusion arrhythmias were also investigated using nonparametric statistics. In a first step, the Jokhere-Sterptra free distribution test was performed to test the null hypothesis that the data were similar in the three study groups against the alternative hypothesis of a significant reduction in one or two treatment groups by ordered inequalities. In a second step, a free distribution test for multiple sample comparisons was used to assess the intergroup differences. Values are expressed as means of: SE. A P value ~0.05 was retained as significant. RESULTS
Hemodynamic data. Mean heart rate was similar in the three study groups before occlusion of the left anterior descending coronary artery: 91 t 11 beats/min in the control group, 85 t 5 in the normosmotic infusion group, and 86 t 5 in the mannitol infusion group. Mean blood pressure was also similar as 83 t 6, 72 t 2, and 73 t 11 mmHg, respectively. The only significant hemodynamic change observed throughout the study consisted in a 8.8 t 3.0 mmHg decrease in mean aortic blood pressure present in the three study groups during the occlusion period, before any intervention (P c 0.05, compared with preocclusion values). Heart rate was, however, unchanged at that time (0.2 t 4.9 beats/min variation). The intracoronary infusion before reperfusion resulted in a 0.4 t 4.5 beats/min increase in heart rate in the normosmotic group and a 4 t 4.4 beats/min decrease in the mannitol group [nonsignificant (NS)] with a nonsignificant fall of 7.1 t 4.0 and 0.1 t 1.6 mmHg, respectively, in blood pressure. The hemodynamic data were less stable during the reperfusion period because of the presence of reperfusion arrhythmias. When assessedduring periods of sinus rhythm, heart rate was 90.4 t 7.5 beats/min in the control group, 94.7 t 7.9 in the normosmotic group, and 82.9 t 3.6 in the mannitol group. Mean blood pressure was 73.5 t 4.0, 66.6 t 5.9, and 68.3 -+ 11.2 mmHg (NS), respectively. The intra-arterial catheters were removed after 45 min of reperfusion, and subsequent blood pressure readings were not obtained. Ventricular premature beats occurred at a rate of 1.3 -+ 0.7 beats/min during the late occlusion period. The intracoronary infusion during the occlusion period resulted in a significant increase in the rate of ventricular premature beats to 67 t 20 beats/min in the normosmotic group. No such increase was however observed with the mannitol infusion, and the number of premature beats remained 7 t 5 beats/min as in the control group
H19
REPERFUSION
during the last 3 min of occlusion before reperfusion. c oronary reperfusion triggered a similar number of ventricular premature beats in the three study groups (Fig. 1)
Intramyocardial blood flow distribution. Microspheres injected along with the intracoronary infusion during the occlusion period distributed almost exclusively within the area at risk with only 0.5% of the total activity outside this area. The distribution of the infusate within the area at risk was in the subendocardial third 0.83 t 0.37 ml min-l l g-l of tissue in the normosmotic infusion group and 0.65 t 0.40 in the mannitol group (NS); in midwall, 0.85 t 0.38 and 0.58 t 0.31, respectively (NS); and in the outer third 0.85 t 0.33 and 1.37 t 0.33 ml. min. g-l (NS). The se flgures refer to the infusate flow during the last 3 min of coronary occlusion. Myocardial water content. Water content was measured by dessication of multiple fragments of the normal and reperfused myocardium in nine animals, three from each study group. Water content of the normal nonischemic myocardium averaged 374 t 3 ml/l00 g dry wt and was the same in the control group (364 t 6 ml/100 g dry wt), in the normosmotic (381 t 2 ml/l00 g dry wt), and in the mannitol group (378 t 7 ml/100 g dry wt). The reperfused myocardium showed an excess water content with average values of 530 t 7 ml/100 g dry wt in the three groups (P c 0.0001) with significant intergroup difference (P c 0.001). This intergroup difference in reperfused animals reflected lower values in the mannitol group [491 t 11 ml/l00 g dry wt (P < O.OOl)] compared with each of the two other study groups, which showed similar water content (556 t 7 ml/l00 g dry wt) in control and (543 t 8 ml/100 g dry wt) in normosmotic group, P = NS (Fig. 2). Reperfusion thus resulted in a 47% excess in total water content; with mannitol the excess was reduced by 35%. No transmural gradient in water content was present in any of the groups. Mean water content in the normal myocardium was 383 t 6,372 t 5, and 369 t 6 ml/100 g l
T 150
1 0 Control
-3 -2 -1 intracoronarv
0
1
sinfusion reperfusion
Time
2
3
+ b
(minutes)
Fig. 1. Number of ventricular premature beats during late occlusion and early reperfusion in 3 study groups. In control group, no intracoronary infusion was performed and arrhythmias occurred during early reperfusion. Intracoronary infusion of normosmotic infusion triggered reperfusion arrhythmias before release of coronary occlusion. Mannitol infusion did not induce such arrhythmias, which were delayed until upon coronary reperfusion as in control group. *P < 0.05.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
H20
HYPEROSMOTIC 600
REPERFUSION
11.42 t 0.817 o in the normosmotic solution, and 10.59 t 0.67% in the hyperosmotic group. Infarct size measured after 24 h by the triphenyltetrazolium technique was also the same in the control and normosmotic groups (6.80 t 1.05 and 6.22 t 0.76% of ventricular mass, NS) but significantly less in the mannitol group (4.47 t 0.46; P < 0.02). Similarly, the ratio infarct size to area at risk was similar in control and normosmotic groups (0.543 t 0.052 and 0.547 t 0.045) and significantly less (0.415 t 0.029) in the mannitol group (P < 0.001) for a 24% reduction in predicted infarct size (Fig. 2).
1
DISCUSSION l
Control
Normoosmotic infusion
Mannitol infusion
Fig. 2. Myocardial water content (top) and ratio of infarct size to area at risk (bottom) in 3 study groups. Water content and ratio are similar in control group and in normosmotic reperfusion groups. Mannitol added to intracoronary infusion significantly decreases both water content and predicted infarct size. *P < 0.02.
Table
1. Myocardial water content in control and reperfused myocardium Transmural
Sampling Site
Control
Control
Endocardial Mesocardial Epicardial Mean
372k3 362kll
Normosmotic Infusion
Mannitol Infusion
All
392t16 376tlO
383t6 372t5
rnyocardium 384k2 377t2
359t16
381G
367t9
369t6
364k6
381t2
378t7
374k3
Reperfused
rnyocardium
Endocardial 553t20 536t5 483k21 Mesocardial 56Okll 555&16 506t28 Epicardial 555t9 539&17 483k4 Mean 556t7 543t8 491Ik11* Values are means t SE. Myocardial water content expressed in ml of water per 100 g of dry weight. * P < 0.001 compared with reperfused myocardium of the control and normosmotic reperfusion groups. The transmural location of the myocardial fragments did not influence the results.
dry wt, respectively, in endo-, meso-, and epicardial sections. The mean of the reperfused myocardium of control and normoncotic infusion groups was 544 t 10, 557 t 9, and 547 t 9 (NS) ml/100 g dry wt compared with 483 t 21,506 t 28, and 483 t 4 ml/100 g dry w-t in the mannitol infusion group. Table 1 provides the water content calculated in transmural layers of the control nonischemic myocardium and of the reperfused myocardium in the three study groups. Infarct size, area at risk, and infarct size to area at risk ratio. The area at risk calculated from in vivo fluorescein
injection was of similar size in the three study groups: 12.3 t 1.19?’o o f ventricular mass in the control group,
This study demonstrates that coronary reperfusion with a hyperosmotic solution with mannitol can reduce early myocardial edema associated with coronary artery occlusion and reperfusion, favorably influence the ratio infarct size to area at risk and limit the ultimate infarct size. Comparison with previous studies. Reduction in cell swelling by a hyperosmotic solution has previously been demonstrated in experiments involving isolated hearts (15, 21). Studies have also shown that mannitol can improve blood flow to the ischemic area (15,22), preserve left ventricular function (15, 22), maintain the histological and ultrastructural aspect of the myocardium (15, 21, 22) and improve cellular electrophysiology (21). The consequences of these beneficial effects on infarct size per se have been tested, however, only in a few studies that have yielded controversial results. Kloner et al. (11) administered mannitol before and after a 40-min coronary occlusion in 29 dogs and reported a 50% reduction in necrosis measured in the posterior papillary muscle. More recently, Klein et al. (10) reperfused 21 pigs with an intracoronary infusion of mannitol at various concentrations after a 75-min coronary occlusion. Infarct size and myocardial water content, measured 2 h later, were not improved with the hyperosmotic solution. Nevertheless, higher ATP levels in the myocardium reperfused with higher mannitol concentration were found. A third study in dogs has shown a beneficial effect on recovery of function with administration of a hyperosmolar cardioplegic solution of mannitol and glucose; the concomitant use of diltiazem and several other substrates complicates the interpretation of the results of this study (15) In the present investigation, the hyperosmotic intracoronary infusion was administered directly into the ischemic area, starting 3 min before and maintained for 33 min during reperfusion. Myocardial edema observed during early reperfusion was reduced by 35%, infarct size after 24 h by 31%, and the ratio of infarct size to area at risk improved by 24% compared with control hearts with no intracoronary infusion and to hearts reperfused with a normosmotic infusion. The two groups with the intracoronary infusion differ from each other only by the presence or absence of mannitol in the intracoronary perfusate, ruling out a possible confounding effect of the infusion per se. Methodological considerations. The experimental model of coronary occlusion and reperfusion used in this
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HYPEROSMOTIC
study has previously been shown to produce infarcts of reproducible sizes and to be sensitive to interventions influencing the ratio of infarct size to area at risk (5, 7). Reproducibility of measures of infarct size is helped by the absence of collateral blood flow in the pig heart, the standard site of occlusion of the left anterior descending coronary artery, and also by the quantification of infarct size 24 h after reperfusion when the borders can be well delineated by the triphenyltetrazolium chloride technique. The correlation observed between this technique and histological analysis in this study was excellent. We previously also documented this correlation by a systematic analysis of the transmural extension of the infarct every millimeter in multiple myocardial slices (8). Results with the staining technique may be less reliable when performed earlier after coronary occlusion (17). The 45 to 55min occlusion period selected corresponds to a period of rapid growth of the infarct area within the area at risk and may represent a critical time to detect myocardial salvage (6, 19). This duration of occlusion also corresponds to the time required to achieve revascularization with the currently used fibrinolytic agents in humans. Direct intracoronary infusion was used to permit selective impregnation of the area at risk with a highly hyperosmolar solution corresponding to 1.5 times normal osmolarity, as opposed to previous studies, which have generally limited the hyperosmotic state to 20-60 mosM (1, 11). The intracoronary infusion of this hyperosmotic solution was well tolerated and did not induce any detectable hemodynamic change. This high concentration was selected to match previously described osmotic pressure of ischemic cells (20, 21) and the duration of the infusion to correspond to the duration of the transient ultrastructural changes described after reperfusion (9); the decreasing rate of the infusion was empirically selected to mimic the presumed progressive normalization of extracellular osmotic pressure (21). Estimation of flow by microspheres injected during coronary occlusion showed a rate of perfusion of 0.84 t 0.07 ml. min-’ . g tissue-’ in the normosmotic infusion group and 0.86 t 0.08 ml min-l . g tissue-l in the mannitol group (NS), near normal coronary blood flow. Reperfusion arrhythmias. Characterization of reperfusion arrhythmia was not a main goal of this study. However, the normosmotic infusion triggered reperfusion arrhythmias. These appeared before restoration of anterograde coronary blood flow. Such was not the use in the mannitol infusion group where arrhythmias occurred only after restoration of coronary blood flow as in the control group. Myocardial water content. The myocardial water content of the control group and of the group reperfused with the normosmotic infusion showed an excess water of 169 ml/100 g of dry wt or 48% compared with normal myocardium. Reperfusion with mannitol decreased this water excess by 35%. Mechanisms of action. The observations do not allow the extrapolation that reduction in infarct size is directly related to the decrease in myocardial edema. Water content was measured after 33 min of reperfusion and infarct size 24 h later, and the two observations could represent l
REPERFUSION
H21
independent findings. Nevertheless, water and sodium contents of the ischemic myocardium increase early after coronary occlusion (9) and tissue osmolarity measured after 75 min of ischemia ranges between 360 and 420 mosM (20, 21). The increase is initially more important in myocytes than in the extracellular space with an 8% excess of volume after 30 min of occlusion. Reperfusion results in a sudden restoration of extracellular osmolarity to normal values and generates an abrupt osmotic gradient that results in further cell swelling with additional water excess, which may represent 16% of cell volume. Cell swelling imposes mechanical overload to the sarcolemma leading to the formation of sarcolemma blebs, sarcolemmal damage, and disruption (20). Whereas aerobic myocytes can withstand increases up to 40% in cell volume without damage, an 8% increase in ischemic myocytes can result in sarcolemmal disruption (20). Other possible beneficial actions of mannitol could be washout of extracellular and intracellular toxic metabolites and prevention of the toxicity associated with rapid oxygen restoration. Mannitol possesses free radical scavenger properties inhibiting the hydroxyl radicals located extracellularly (13). A few studies have suggested that the benefits of mannitol could be explained by this scavenging mechanism (2, 12), based on the observation of an absence of an effect of hyperosmolar solutions other than mannitol. The agents used in these studies were however sodium and glucose (12) and both penetrate cell membrane, limiting the driving osmotic force, and possibly inducing other effects on cell physiology. Mannitol, on the other hand, acts extracellularly except in cells with altered permeability and irreversibly damaged (16, 20). In a study performed using the very same protocol, high doses of superoxide dismutase administered intracoronary from 3 min before to 33 min after reperfusion had no effect on infarct size (4). Mannitol is not expected to add much benefit to superoxide dismutase as an hydroxide radical scavenger, considering that the superoxide anion is required for the generation of the hydroxyl radical from hydrogen peroxide (17). A more direct demonstration that mannitol acts through a hyperosmotic mechanism rather than through an antioxidant effect would require the demonstration of the same effects with a pure hyperosmotic agent. In practice, such an agent is difficult to find: 1) it should be water soluble and have a relatively low molecular weight to allow important increase in osmotic pressure; 2) it must also be large enough to prevent its entry into the cell; and finally 3) such an agent should not accept electrons. Since most of these agents are maximally oxidized, they could behave as oxidants with paradoxically harmful effects. Clinical implications. Reperfusion with mannitol is associated with infarct size smaller than predicted from the area at risk, suggesting that the conditions of reperfusion could be important for true salvage of myocardial cells that would otherwise have died. This stresses the importance of continuous research to try to derive maximal benefit from pharmacological and mechanical recanalization of acutely occluded arteries in acute myocardial infarction in humans. Prevention of cell swelling may represent one of the promising approaches to pre-
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
HZ2
HYPEROSMOTIC
serve ischemia-weakened sarcolemma, cell integrity, and cell-to-cell progression in myocardial necrosis. This approach could also be useful in other clinical conditions when cell preservation is important, such as complex coronary angioplasty, myocardial cardioplegia, and heart transplantation. The authors thank Lute Begin and Lise de Repentigny for the excellent secretarial work. This study was supported by Grant PA 86/0389 from the Comision Interministerial de Ciencia y Tecnologia and Grant l/89 of the Convenio Madrid-Quebec. Address for reprint requests: P. Theroux, Montreal Heart Institute, 5000 Belanger St., Montreal, Quebec HlT lC8, Canada. Received
2 November
1990; accepted
in final
form
14 August
1991.
REFERENCES 1. Apstein, C. S., F. N. Gravino, and C. C. Houdenschild. Determinants of a protective effect of glucose and insulin on the ischemic myocardium. Effects on contractile function, diastolic compliance, metabolism and ultrastructure during ischemia and reperfusion. Circ. Res. 52: 515-526, 1983. 2. England, M. D., C. C. Caravochi, J. F. O’Brien, E. Solis, J. R. Pluth, T. A. Orszulak, M. P. Kaye, and H. V. Schass. Influence of antioxidants (mannitol and allopurinol) on oxygen free radical generation during and after cardiopulmonary bypass. Circulation 74, Suppl. III: III-134-111137, 1986. 3. Engler, R., and E. Gilpin. Can superoxide dismutase alter myocardial infarct size? Circulation 79: 1137-1142, 1989. 4. Garcia-Dorado, D., P. Theroux, J. Alonso, J. Elizaga, J. Botas, F. Fernandez-AvilC?s, J. Soriano, R. Munoz, and J. Solares. Intracoronary infusion of superoxide dismutase and reperfusion injury in the pig. Basic Res. CardioZ. 85: 619-629, 1990. 5. Garcia-Dorado, D., P. Theroux, J. Elizaga, F. FernandezAviles, J. Alonso, and J. Solares. Influence of tachycardia and arterial hypertension on infarct size in the pig. Cardiouasc. Res. 22: 620-626,1988. 6. Garcia-Dorado, D., P. Theroux, J. Elizaga, M. Galinanes, J. Solares, M. Riesgo, M. J. Gomez, A. Garcia-Dorado, and F. Aviles. Myocardial reperfusion in the pig heart model: infarct size and duration of coronary occlusion. Cardiovasc. Res. 21: 537544, 1987. 7. Garcia-Dorado, D., P. Theroux, F. Fernandez-Aviles, J. Elizaga, J. Solares, and M. Galinanes. Diltiazem and progression of myocardial ischemic damage during coronary artery occlusion and reperfusion in porcine heart. J. Am. CoZl. Cardiol. 10: 906911, 1987. 8. Garcia-Dorado, D., P. Theroux, J. Solares, J. Alonso, F. Fernandez-Aviles, J. Elizaga, J. Soriano, J. Botas, and R. Munoz. Determinants of hemorrhagic infarcts. Histologic observations from experiments involving coronary occlusion, coronary reperfusion and reocclusion. Am. J. Pathol. 137: 301-311, 1990.
REPERFUSION 9. Jennings, R. B., J. Schaper, M. L. Hill, C. Steenbergen, and K. A. Reimer. Effect of reperfusion late in the phase of reversible ischemic injury. Changes in cell volume, electrolytes, metabolites and ultrastructure. Circ. Res. 56: 262-278, 1985. 10. Klein, H. H., K. Nebendahl, M. Schubothe, and H. Kreuzer. Intracoronary mannitol during reperfusion does not affect infarct size in ischemic, reperfused porcine hearts. Basic Res. Cardiol. 80: 251-259,1985. 11. Kloner, R. A., K. A. Reimer, J. T. Willerson, and R. B. Jennings. Reduction of experimental myocardial infarct size with hyperosmotic mannitol. Proc. Sot. Exp. Biol. A4ed. 151: 677-683, 1976. 12. Macgovern, G. J., S. F. Bolling, A. S. Carole, B. H. Bulkley, and T. J. Gardner. The mechanism of mannitol in reducing ischemic injury: hyperosmolarity or hydroxil scavenger. Circulation 70, Suppl. I: 1-91-I-95, 1984. 13. Maestro, R. F., H. H. Thaw, J. Bjork, M. Plonker, and K. E. Arfors. Free radicals as mediators of tissue injury. Acta Physiol. Stand. 492, Suppl. I: 43-57, 1980. 14. Nayler, W. G., and J. S. Elz. Reperfusion injury: laboratory artifact or clinical dilemma. Circulation 74: 215-221, 1986. 15. Okamoto, F., B. S. Allen, G. D. Buckberg, H. Young, H. Bugyi, and J. Leaf. Studies of controlled reperfusion after ischemia. Reperfusate composition: interaction of marked hyperglycemia and marked hyperosmolarity in allowing immediate contractile recovery after four hours of regional ischemia. J. Thorac. Cardiovasc. Surg. 92: 583-593, 1986. 16. Reimer, K. A., R. B. Jennings, and M. L. Hill. Total ischemia in dog hearts, in vitro. 2. High energy phosphate depletion and associated defects in energy metabolism cell volume regulation and sarcolemmal integrity. Circ. Res. 49: 901-911, 1981. 17. Reimer, K. A., C. E. Murry, and V. J. Richard. The role of neutrophils and free radicals in the ischemic-reperfused heart: why the confusion and controversy? J. Mol. Cell. Cardiol. 21: 12351239, 1989. 18. Richard, V. J., C. E. Murry, R. B. Jennings, and K. A. Reimer. Therapy to reduce free radicals during early reperfusion does not limit myocardial infarcts caused by 90 min of ischemia in dogs. Circulation 78: 473-480, 1988. 19. Schaper, W., K. Binz, S. Sass, and B. Winkler. Influence of collateral blood flow and variations in MV02 on tissue-ATP content in ischemic and infarcted myocardium. J. Mol. Cell. Cardiol. 19: 19-37, 1987. 20. Steenbergen, C., M. L. Hill, and R. B. Jennings. Volume regulation and plasma membrane injury in aerobic, anaerobic and ischemic myocardium in vitro. Effects of osmotic cell swelling on plasma membrane integrity. Circ. Res. 57: 864-875, 1985. 21. Tranum-Jensen, J., M. Janse, J. W. T. Fiolet, W. J. G. Krieger, C. N. D’Alnoncourt, and D. Durrer. Tissue osmolality, cell swelling and reperfusion in acute regional myocardial ischemia in the isolated porcine heart. Circ. Res. 49: 364-381, 1987. 22. Willerson, J. T., W. J. Powell, T. E. Guiney, J. J. Stark, C. A. Sanders, and A. Leaf. Improvement in myocardial function and coronary blood flow in ischemic myocardium after mannitol. J. Clin. Invest. 51: 2989-2998, 1972.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on September 6, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
