W J um” lW/lllfll/WWW/lfl 3129300 989 8697 This is to certify that the thesis entitled Experimental Non-Lethal Endotoxemia in the Pony: Modification of the Pathophysiology by Polymixin B. presented by Dolores Johanna Kunze has been accepted towards fulfillment ‘ of the requirements for Master of Science degree in_L_a_m_e_Anj_mal Surgery and Medicine 7 (x I”. ‘\ L‘ka’Wé ‘/ 1K ¥:_\ Major professor .2( \\l Date Fabruary 1331980 07639 L OVERDUE FINES ARE 35¢ PER DAY PER ITEM it ,._-\~ ”'1" Return to buck drop to remove 'Qa‘; this checkout; iron;- your record. '.\l‘r~ «a 1'. .;'-_- WW s)' '-0 (rs. 9U L i’C .U , "c '3 ‘5 ‘ EXPERIMENTAL NON-LETHAL ENDOTOXEMIA IN THE PONY: MODIFICATION OF PATHOPHYSIOLOGY BY POLYMIXIN B SULFATE BY Dolores Johanna Kunze A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Large Animal Surgery and Medicine 1980 ABSTRACT EXPERIMENTAL NON-LETHAL ENDOTOXEMIA IN THE PONY: MODIFICATION OF THE PATHOPHYSIOLOGY BY POLYMIXIN B BY Dolores Johanna Kunze Two groups of five ponies were treated with very low doses of endotoxin; the first group received endotoxin only and the second received endotoxin plus polymixin 3 administered concurrently. Two other groups, two ponies per group, served as controls with one receiving polymixin B and the other receiving saline. Following the infusion, the ponies trembled and coughed. They also developed pyrexia, leucopenia and transient pulmonary hypertension resulting from increased pulmonary vascular resistance. In the animals receiving endotoxin plus polymixin B, the leucopenia, pulmonary hypertension and increased pulmo- nary vascular resistance were ameliorated. Polymixin B, which binds to the lipid A moiety of the endotoxin molecule, did not block the pyrexia, suggesting that some other part of the endotoxin molecule might be responsible for endogenous pyrogen release. An increase in circulating leucocytes, particularly lymphocytes, was observed in the ponies given polymixin B alone. ACKNOWLEDGEMENTS My deepest gratitude is extended to Dr. Robinson, whose patience, enthusiasm and encouragement were essential to the completion of this work. Such intelligence and empathy are rare and are very much appreciated. Special thanks to Drs. Mather and Riley and the Large Animal Faculty for their support. In particular, I am especially grateful to Drs. Scott and Stowe for their advice and help in the preparation of this thesis. My appreciation and thanks also belong to Ms. Bobbi Milar for not only her technical expertise and assistance, but also for her good nature, an invaluable asset. And, finally, to my husband, Morrow, my gratitude for his calm, quiet support which was a major sustaining force throughout this some- times hectic enterprise. ii TABLE OF CONTENTS ACRHOWl edgements O O O O O O O O O O O O O O O O O O I O O 0 List Of Tables 0 O O O O O O O O O O O O O O O O O O O O O 0 List Of Figures 0 O O O O O O O O O O O O O O O O O O O O O I. II. III. IV. IntrOduction O O O O O O O O O O O O O O O O O O O O 0 Literature R8View O O O O O O O O O O O O O O O O O 0 Origin, Composition and Preparation of Endotoxin Clinical Signs of Endotoxemia . . . . . . . . . . Hematology of Endotoxemia . . . . . . . . . . . . Hemodynamics of Endotoxemia . . . . . . . . . . . Structure and Molecular Interactions of Polymixin Modification of Endotoxin Pathophysiology by Polymixin B . . . . . . . . . . . . . . . . . Experimental Rationale . . . . . . . . . . . . . Materials andMethOdS 00 0.00.0.0. 0000. Experimental Animals . . . . . . . . . . . . . . Surgical Procedure . . . . . . . . . . . . . . . Experimental Protocol . . . . . . . . . . . . . . Measurement of Blood Pressure . . . . . . . . . . Measurement of Cardiac Function . . . . . . . . . Hematologic Parameters . . . . . . . . . . . . . Statistical Analysis . . . . . . . . . . . . . . Results 0 O O O O O O O O O O O O O I O O O O O O O 0 Clinical Observations . . . . . . . . . . . . . . Hematologic Effects . . . . . . . . . . . . . . . HemOdynamiC Ef feCtB C O O O O O O O O O O 0 O O 0 iii Page ii vi 22 22 22 24 25 25 26 27 28 28 33 48 TABLE OF CONTENTS (Cont.) Page v. DiSCUSSion O O O O O I O O O O O I O O O O O O O O I O I O O 64 VI. mary O O O O O O I O O O O O O O O O O I O O O O I O O O 71 VII.LiStOfReferenCQSOOOOOOO0000000000000072 iv TABLE 10 LIST OF TABLES Statistical comparisons between temperature.......... Statistical comparisons between white blood cell counts . . . . Statistical comparisons between segmented neutrophil counts . . Statistical comparisons between non-segmented neutrophil counts Statistical comparisons between lymphocyte counts 0 o o o o o 0 Statistical comparisons between v01‘me8 O O I O O O O O O O O 0 Statistical comparisons between arterial pressure . . . . . . . Cardiac outputs . . . . . . . . Statistical comparisons between peripheral resistance . . . . . Statistical comparisons between vascular resistance . . . . . . means of body PAGE 32 36 40 43 47 51 57 58 62 63 FIGURE 10 Overview of host defenses and endotoxin interactions Changes Changes Changes Changes counts Changes Changes Changes Changes Tbtal peripheral resistance and pulmonary vascular LIST OF FIGURES in body temperature . . . . . . . . . . in circulating white blood cell counts in circulating segmented neutrophil counts in circulating non-segmented neutrophil in circulating lymphocyte counts . . . in packed cell volume . . . . . . . . . in systemic arterial blood pressure . . in pulmonary arterial blood pressure . reSiStance O O O O O O O O O O O O O I O O O 0 vi PAGE 14 31 35 39 42 46 50 53 56 61 INTRODUCTION Endotoxemia is commonly recognized as a contributing factor in various pathologic states, including gastrointestinal disease, aspira- tion pneumonia and wounds with gram-negative sepsis. Normal intestinal microflora, largely composed of gram-negative bacteria, continuously produce and release endotoxins which are absorbed by the intestinal mucosa and are detoxified by the liver. Concurrent hepatic or gastroin- testinal disease may reduce detoxification or may increase endotoxin absorption.“2 Since endotoxemia occurs frequently and often poses a serious threat to a patient's survival, the role of endotoxin in the pathogenesis of naturally occurring and experimental disease has been extensively investigated. Because the biologic effects of endotoxin are so varied, the endotoxemia patient may exhibit a range of signs from mild depression and pyrexia to massive, fulminating shoclt:.3'8 In experimental models, most researchers have used lethal doses of endo- toxins and have based their conclusions on the documentation of fatal shock.9‘15 In contrast, the clinical endotoxemic patient may or may not exhibit signs of shock. Therefore, the standard experimental model depicts only one manifestation of endotoxemia. The hematologic and phy- siologic changes associated with clinical endotoxemia generally develop more gradually and are much more prolonged than those seen in experimen- tal models.13'14'15'20 Clinically, endotoxemia in the horse has been implicated or suggested as a complication in certain gastrointestinal disorders.19'21'23 And, as with other species, the experimental endo- toxemia studies in horses produced abnormalities not entirely consistent with those observed in clinical cases.9i13'14 Therefore, comparisons and conclusions made between experimental animals and clinical animals may not be entirely valid. The therapeutic plan for endotoxin shock includes fluid volume replacement and correction of acid-base imbalances.3I17I18I25 With septic shock, appropriate antimicrobial agents are often used, but the use of bactericidal drugs may actually exacerbate shock by accelerating lysis of gram-negative bacteria and release of endotoxins. This possi- bility is necessarily weighed against the potentially graver threat of gram-negative septicemia. While most antibiotics that are effective against gram-negative organisms will not interfere with the bioactivity of the endotoxins, the cyclic cationic polypeptides will interfere by interacting with the lipid components of the endotoxin molecule.26"28 Polymixin B sulfate, colistin and tyrocidine are members of this group of antibiotics but are unpopular because of the accompanying nephrotoxicity.3o Polymixin B sulfate is less nephrotoxic than others in this group and is more effective in neutralizing endotoxin when admixed before administration.28r29 Polymixin B sulfate has been given before, during and after endotoxin administration with variable results, but some lessening of either the disease state or the mortality occurred with the use of the antibiotic concurrent with endotoxin administration.31‘43 It has even been used in a few clinical cases with limited positive results.44 The timing of the administration of the drug appears to be critical, which has made clinical application seem impractical. The objectives of these experiments were to investigate sublethal endotoxemia in ponies, and to determine if the subsequent pathophysiology could be blocked by polymixin B. LITERATURE REVIEW Origin, Composition and Preparation of Endotoxin Endotoxin, the portion of an enteric bacterium responsible for the toxic effects, is the outer part of the cell wall. This material, a complex of lipopolysaccharide (LPS) and protein is released from cells during active growth and upon cell lysis; the molecule encompasses both the endotoxin activity and the somatic antigens of the bacterium. Three distinct regions form the molecule: the outer hydrophilic region, a heteropolysaccharide containing the repeating O-specific antigenic units; the central acid core, linking the O-antigens and the lipid portion; and the internal lipid-rich region, or lipid A.15125I46 Lipid A has a high affinity for cell membranes, for other lipids and for protein. It represents the biologically active part of the wild strains and the com- mercially prepared endotoxin, and the polysaccharide portion facilitates the solubility of the lipid in aqueous media.16 Regardless of the bac- terial species of origin, the experimental effects of purified prepara- tions generally resemble the signs of naturally occurring gram-negative sepsis.25 Response varies dramatically with dose, and the method of LPS extraction may modify the toxicity of the preparation. The Boivin method involves extraction with ice-cold trichloroacetic acid, resulting in a preparation consisting primarily of LPS with some protein and lipid: this preparation is slightly more active than some others because not only does it contain large amounts of lipid A, but it also has a carbohydrate side chain which appears to enhance activity.46147 With the other commonly used preparation, the Westphal method, endotoxin is extracted with phenol which does not preserve some of the side chains.15l4S Different lots obtained from a single supplier and derived from the same bacterial strain may vary in their ability to elicit a given response, particularly from platelets.4'8‘52 Overall, however, the major trends are consistent enough to allow comparison of experiments using varying combinations of lots, extraction methods and bacterial species. Whereas individual animals of a given species respond similarly, there are species differences in response to endotoxin. Clinical Signs of Endotoxemia No one set of clinical signs fits every endotoxemia patient; with varying degrees of endotoxemia, a given patient of any species, including human or equine, could appear to be normal or to be in near- terminal shock. However, the shock cases usually have similar signs: pyrexia, sweating, increased heart rate, varying degrees of depression, cold extremities, shivering, shallow respirations or some degree of dyspnea, congested or cyanotic mucous membranes, prolonged capillary refill time, and, occasionally, evidence of coagulation disturbances such as prolonged bleeding times.3r9t17r53 Very mild cases are usually pyrexic and may have a mild secondary depression; fever is one of the most consistent signs of endotoxemia, regardless of species or degree of involvement; endotoxin is not only a very potent releaser of endogenous pyrogen, but also may have a direct cerebral effect.5'7 The greater the degree of intoxication, the more severe the signs become. Hematology of Endotoxemia The familiar hematologic picture of endotoxemia, both naturally occurring and experimentally induced, is characterized by an initial neutropenia and thrombycytopenia, a somewhat later appearing lymphopenia and an even later 1eucocytosis.9l15'20'25'35144'53'61I64’66 Platelets and leucocytes predominate as the peripheral agents for endotoxin clearance; granulocytes actually remove endotoxin, and platelets appear to facilitate removal by enhancing recognition of endotoxin by phagocytes.54 When endotoxin attaches to platelets, the combination forms the equivalent of a membrane-associated antigen. Through this interaction, both the classical and alternate complement pathways are activated with the subsequent generation of opsonins and chemotactic factors and the lysis of gram-negative bacteria.55"57 The mechanisms of the platelet-endotoxin interactions depend critically on the presence or absence of immune adherence sites on the platelet membrane. Primate platelets lack these receptors and respond differently than the plate- lets of rats, rabbits, dogs, and guinea pigs.58r59 The presence of these receptors has not been reported in the horse. In rabbits, rats, dogs, and guinea pigs, after platelet-endotoxin combination occurs, pla- telet responses are usually characterized by aggregation and release of platelet constituents, including ADP and vasoactive amines such as histamine and serotonin.60 The release response of platelets may occur through lysis, in which granules and cytoplasmic constituents are liberated, or by secretion, in which the platelet membrane remains intact and only granule constituents are released. Serum complement is probably involved in the lytic response.58 Horse platelets do respond to endotoxin, but the exact mechanism has not been determined.61 While little is known of the physiologic, pathophysiologic and pharmacologic characteristics of horse platelets, they respond to ADP with aggregation, and in most other species, aggregation occurs with secre- tion of dense granule contents.62I63 Endotoxin-platelet interactions may contribute to endotoxin shock through three mechanisms: through the release of vasoactive substances such as histamine and serotonin; through the formation of intravascular occlusive platelet aggregates; and through the induction of disseminated intravascular coagulation.16I67 The release reaction has already been briefly discussed and will be discussed further in the hemodynamics section. Platelet aggregation has been extensively investigated in its contributions to endotoxin pathophysiology and shock. Platelets are activated and become sticky in response to endotoxin; clumps of plate- lets and leucocytes consistently form in the capillary beds of the liver, spleen and particularly of the lung.4c1or12I25I61I63I59 The lungs are a very important site of endotoxin and host defense inter- actions. The aggregations that form in small vessels may contribute mechanically to the pulmonary signs observed: dyspnea with pulmonary congestion and edema and foamy mucoid material draining from the noses and mouths of moribund animals.10v14r53r69 However, these aggregates are not directly involved in the pulmonary hemodynamic alterations, but platelet granular constituents are.4l10'12I60155r55I69‘77 The release reaction is more important in the pathogenesis of endotoxin shock than is the aggregation reaction, but aggregation usually precedes release.67 The extraordinary ability of endotoxins to produce tissue injury through the initiation of coagulation changes, disseminated intravascular coagulation (DIC), has been recognized for over 50 years.72I73 Endotoxin is a solid phase activator of Hageman factor (Factor XII) which, in turn, initiates the intrinsic coagulation pathway.74 In addition to this, active Hageman factor can activate pre- kallikrein to form kallikrein which generates bradykinin, and active Hageman factor is capable of directly activating Factor VII, indicating the probable contribution of the extrinsic coagulation pathway to endo- toxin initiated coagulopathies.75'77 Endotoxin also damages endothelial cells and indirectly activates the extrinsic pathway through the release of tissue thromboplastins.56 The diagnosis of DIC is made in the laboratory, characterized by decreased platelet numbers, prolongation of the partial thromboplastin time (intrinsic coagulation pathway), pro- longation of the prothrombin time (extrinsic coagulation pathway), and elevation in fibrin degradation products. The development of coagulo- pathies in response to endotoxin has been reported in most species studied, including horses.61 Subsequently, the formation of microthrombi in many organs, particularly the kidneys and lungs, contri- butes to mortality. An endotoxemia patient may survive the initial shock, if provided with adequate medical support, only to die from secondary complications of renal failure or pneumonitis. The commonest hematologic event in endotoxemia is an extreme leuco- penia followed by leucocytosis. Neutrophils are the most affected, probably because they not only interact directly with endotoxin, but they are involved in endotoxin-induced platelet and complement interac- tions as well. When endotoxin attaches to platelets, complement is activated (alternate and classical pathways), and opsonins and chemotac- tic factors are generated.55‘57 Because of these, neutrophils and other granulocytes are attracted, and their enzymatic and phagocytic activity increases.54r57 With increasing levels of endotoxin, the cellular defenses are overcome: complement factors are depleted; plate- lets aggregate and degranulate; coagulation factors are consumed; neutrophils marginate and migrate into tissues, decline in function, degranulate and degenerate; damaged endothelial cells exfoliate to float freely in the bloodstream; and the reticulo-endothelial system (RES) is depressed.3t57I67I78 In neutrophils, endotoxin directly interacts with lysosomal membranes, disrupting membrane transport and causing enzyme leakage.54 Eventually, neutrophil chemotaxis is inhibited or overcome by humoral factors released from other cells triggered by endotoxin, but the major neutrophil functional defect stems from increased adhesiveness.57r79 In ‘gigg neutrophil adherence significantly increases and is maximal one hour after treatment with endotoxin. This increased adherence is asso- ciated with a reciprocal granulocytopenia that is followed by a progressive granulocytosis. This adhesiveness appears to be due to a heat-stable plasma component that requires a heat-labile cofactor that might be complement.79 Granulocyte adhesiveness i3 gitgg and margina- tion 33 give are closely associated, complement-dependent phenomena.80 Also, the i3 gigg granulocytopenia corresponds with histological evi- dence of sequestration of platelets and granulocytes in pulmonary capillary beds.4:10,12,67 10 Aside from the vascular occlusion and the resulting anoxic damage, the trapped granulocytes release their lysosomes and exacerbate the local tissue injury. Through the use of a Sanders ear chamber, the inflammatory response to endotoxin was directly observed in rabbits.81 Within three minutes of endotoxin injection, leucocytes were sticking to venule endothelium and would momentarily stick to arteriolar walls. After ten minutes, emboli composed of platelets, leucocytes and erythro- cytes appeared. By one hour, some venules and capillaries were occluded, and their endothelium was swollen. Three hours later, erythrocytes and leucocytes passed from the vessels into the surrounding connective tissue with accompanying edema formation. If the rabbit were dying, the microvascular circulation eventually stopped; this did not occur in surviving animals. In animals which recovered, the circulation did not return to normal for approximately two months when thrombi were replaced. Histopathologic examination of the same tissues from nonsur- vivors revealed similar pathology: congested microvasculature with focal fibrin thrombi; swelling of endothelial cells; margination and migration of leucocytes; and extravasation of erythrocytes. In animals surviving 72 hours, the thrombotic vessels were markedly necrotic, and the supporting connective tissue was degenerative. On histological examination, the microvascular pathology of the lung was identical to that just described for the ear. Concurrently, in these same animals, circulating leucocyte numbers decreased dramatically in ten minutes and remained low for six hours, suggesting leucocyte destruction or migration into tissues. In other ex- periments, when circulating leucocyte counts were followed over 24 hours, 11 the decrease in total leucocytes was mainly due to a drop in neutrophils which was maximal at about one hour.9t10'14'15137I51o54v57r80o82 Monocytes decreased at the same time, lymphocytes decreased later, and at about 24 hours, the leucocytes, particularly neutrophils, increased to normal levels or higher with an increase in immature forms. The total blood granulocyte pool is composed of the circulating granulocyte pool and the marginated granulocyte pool. Radioisotope labeling shows that the transient granulocytopenia caused by endotoxin is due to a shift of cells from the circulating pool to the marginated pool without an increase in total granulocytes. Later, in the granulocytosis phase, all three pools increase without altering the ratio between circulating and marginated cells.82 These events correlate well with the increased adhesiveness induced by endotoxin.79 Augmentation of adherence was maximal one hour after endotoxin administration and inhibition of adherence was marked by 24 hours. These findings explain the shifts between peripheral compartments but don't explain the overall number changes or the appearance of immature forms. This leucocytosis does not appear to be due to a direct effect on blood flow through the marrow, but rather due to a leucocytosis-inducing factor released in response to granulocytopenia.56 Electron microscopy of bone marrow events showed broad gaps in the nor- mally closed sinus endothelium through which immature cells entered the sinus lumina five minutes after endotoxin. By 15 minutes, the sinus hyperemia had increased, and by 60 minutes, sinus destruction was so extensive that, in some areas, sinus walls were no longer recognizable. The overall result was an increase in immature forms of all cell lines 12 in the general circulation.83 Thus, the granulocytosis is the sum of shifting of mature cells back from the marginated pool, and the early release of mature and immature cells from the bone marrow. The effect of endotoxin on lymphocytes is not well understood, but lymphopenia follows the early granulocytopenia by a few hours. This lymphopenia may be an indirect rather than a direct effect, for serum cortisol is known to increase in many species following endo- toxin, and corticosteroids can directly and indirectly affect lympho- cytes.25126r70I84 In some species, corticosteroids lyse lymphocytes (a dose-related phenomenon), in others initiate sequestration, probably in the bone marrow, and in yet others the reaction is probably mixed.25v84 Also, migratory function appears impaired, lymphokine pro- duction is antagonized, and recruitment of additional lymphoid cells is inhibited.34r85 Endotoxin appears to interact with receptors on lympho- cytes and macrophages; it is mitogenic for B cells but not T cells, transforms B lymphocytes, induces a macrophage chemotactic factor from B cells, activates macrophages, and causes them to release lysosomal enzymes.85 In the horse, specific interactions of endotoxin and lympho- cytes have not been defined, but the lymphopenia is believed to be a secondary effect of endogenous cortisol release.61 The major effects of endotoxin seem to occur mainly through induced hormones and humoral factors. One hematologic hallmark of endotoxin remains to be discussed: the marked elevation of packed cell volume (PCV) observed in endotoxin shock.14I25151r64 An endotoxin receptor on erythrocytes has been 13 isolated, but this is more important in complement activation.23r43 Epinephrine released during sympathicoadrenal activity in response to arterial hypotension is a potent initiator of splenic contraction, and increases in circulating levels of epinephrine are reported in endo- toxin shock.70'87'90 In species such as the horse with splenic capsule contractility, endotoxin has a marked effect on PCV, and when the spleen contracts, erythrocyte reserves are ejected into the general circulation.51'64'86'87 A secondary rise in blood viscosity follows the PCV elevation.87 Other contributions to the PCV increase are losses in plasma volume secondary to increased capillary permeability and increased lymph production.91 Increased blood viscosity may contribute to the coagulopathies and, therefore, may exacerbate the tissue hypoxia of shock. Figure 1 shows a schematic overview of the interplay of humoral and cellular components. Hemodynamics of Endotoxemia From the preceding discussion, it should be obvious that no endo- toxin induced event stands alone; the interactions are circuitous and interdependent. Similarly, the hemodynamic alterations and the initiating events of the patient in endotoxin shock are complex and well interrelated. Classically, the hemodynamic response to endotoxin can be divided into three categories on the basis of species differences. The first is characterized by early splanchnic venous pooling and is seen in the dog, rat, mouse, coyote, and bear; the second category is characterized by pulmonary venous pooling, and this group includes the cat, horse, sheep, calf, and rabbit; and lastly, in primates and presumably in man, multiple sites of venous pooling are 14 A« an ooumoflocfl ohm uoommo poouao m unmxo on nzocx ma aflxouooco scflnz um mouwmv mZOHBU€MMBZH ZHxOBOsz 92¢ mmmZMhMD Emom m0 3MH>mm>O .H MMDUHM .5635. E3232 tassel , Sufi—anon"... .“EBWM ._. Ffihfiflm 265.0235 , .23.... 2.3:. , , 353.23 353.123 mgfiomcm 32¢ 3.00 33005.2 330 3 >0 *meoo -an» .-o E» GHQ * -2502 a _ . E .2358ch * managements. mm ._ n L. 4 * 2.25. 228... 0.820....25 :>8.oon_A.ii.w8.oom coHoooI Lozenges”; *mofiwofi _ w emcEme. ........ or oficEcJ «5.8.2335 29:03:80 I Lacy—055$ e.flcozesocofl 15 suspected.10‘13l17r25r35v65t70187v91 Each category will be discussed individually using the responses of the dog, horse and monkey as typical of the Species in each group. In most species, the vascular reaction to endotoxin is systemic arterial hypotension which develops as a result of decreasing cardiac output due to low venous return, secondary to vascu- lar pooling. In experimental endotoxemia, initially the vasculature is constricted and vascular resistance increases, but later vascular resistance decreases and vessels dilate.7o The early reactions in the dog are transitory; portal vein pressure rises with a concurrent fall in central venous pressure. An increase in small intestine weight suggests a further contribution to the splanchnic venous pooling (the mechanisms involved are poorly understood). Also, pulmonary arterial pressure increases due to a rise in both pulmonary arterial and pulmonary venous resistances.1°v92 Pulmonary vascular changes occur rapidly and are over by 30 minutes. TOtal peripheral resistance rises only slightly, but the cardiac output drops precipi- tously with a resulting systemic hypotension.93 Hence, the fall in blood pressure is due to decreased total blood flow, the result of diminished venous return from blood pooling in the liver, and not due to peripheral dilation or myocardial weakness. This initial hypotension later yields to a return to normal pressure levels, reflecting the release of pooled blood from the liver. The final, prolonged hypoten- sion leads to fatal shock, but the exact site (or sites) of the causa- tive venous pooling is unknown; it does not appear to be the liver.94 The progressive, late-occurring fall in total peripheral resistance, coupled with low cardiac output, results in irreversible shock.93 16 Following slow intravenous injection of endotoxin in the horse, early changes include an immediate increase in pulmonary arterial pressure, a marked decrease in systemic arterial pressure, and only a slight rise in central venous pressure. Mesenteric pooling does not appear to occur, but there is a striking early rise in pulmonary vascu- lar resistance and subsequent pulmonary venous pooling. The systemic arterial pressure returns to normal levels and remains there for several hours. With the later occuring preterminal hypotension, again venous pooling probably occurs, but the exact site remains unknown. The total peripheral resistance may be low at this time.13r14 In monkeys, endotoxin shock is characterized by decreased cardiac output, decreased central arterial blood pressure, and decreased total peripheral resistance. In contrast to other species, the arterial blood pressure declines gradually and after relatively stable spells, drifts into shock levels. Portal venous pressure rises only slightly, but there is no evidence of hepatic or splanchnic pooling; mesenteric blood flow does not change at any time, but mesenteric vascular resistance declines at about 45 minutes after endotoxin injection and remains low for the remainder of the experiment. A probable deficient venous return would account for a decreased cardiac output and the hypotension.7ol95 The patterns differ, but the hemodynamic events of all species have similar features, as do the hematologic events of endotoxemia, and not surprisingly, the two areas are interrelated; white blood cell and platelet numbers decrease concurrently with the early pressure changes. The relationship is direct; a component of whole blood is responsible for the pressor effects of endotoxin, and more specifically, platelets 17 are essential for the hemodynamic pathophysiology.10I69'71 Concurrently with the thrombocytopenia characteristic of endotoxemia, serum concentrations of several vasoactive substances, particularly pressors, increase. Plasma serotonin reaches a maximum level 15 seconds after endotoxin injection and rapidly disappears.16'65I66196I97 Platelet aggregation is not essential for the pressor response, but release is; concurrent administration of a serotonin antagonist will not block the thrombocytopenia, i.e., will not block clumping but will block release.69 As previously mentioned, bradykinin increases through Hageman factor activation and may account for some of the hypotension.74r38v89 Histamine and the catecholamines also occur in higher than normal levels with endotoxin shock.69r70I88I89 Histamine is found in the platelets of most species and in the leucocytes of almost all species.69 There is not only an immediate histamine increase, but there is also a later rise, apparently ”induced" histamine, synthesized in or near vascular endothelial cells following endotoxin.98 The release of biogenic amines in endotoxemia is slower and more gradual than that in anaphylaxis, but similarities do exist: profound hemoconcentration, leucopenia, thrombocytopenia, and pulmonary arterial hypertension with decreased central venous pressure. In acute systemic anaphylaxis in the horse, histamine and serotonin concentrations appeared coupled with white blood cell and platelet numbers; very early in anaphylactic shock, whole blood histamine concentrations rise sharply, but later a much lower histamine level coincides with the most profound period of leucopenia and thrombocytopenia.99 Horse leucocytes are par— ticularly rich in histamine and serotonin, and while levels of biogenic 18 amines have not been measured in the horse, on the basis of the induced anaphylaxis data and the known equine platelet response data, it is not unreasonable to assume that the hemodynamic changes in horses in endo- toxin shock are attributable to vasoactive amines such as histamine and serotonin.62r63r99"'101 Histamine is a constrictor of smooth muscle of the trachea and pulmonary arteries and veins of horses.99v101 Histamine also produces hepatic vein constriction in the dog and is a potent systemic arteriolar dilator. Catecholamines increase in many species in endotoxin shock, possibly accounting for the transient increase in blood pressure following the initial decrease.12188 The effects are probably dose related, and the wide variability in vasoactive amine responses among species may explain the varied species responses to endotoxin. Structure and Molecular Interactions of Polymixin B As previously mentioned, polymixin B is a cyclic, cationic polypeptide; it has a molecular weight of approximately 1250 daltons and contains threonine, 0, Y-diaminobutyric acid, and a nine carbon saturated fatty acid.29r3° This family of drugs is nephrotoxic, and a relationship between intramolecular structure and toxicity has been suggested. While polymixin B is considerably less nephrotoxic than some of the other polymixins, the toxicity of these peptides may be partly due to their D-amino acid content.30 In levels approximating the usual therapeutic dose, polymixin B does not produce clinical signs of depression of renal function in normal dogs.102 Also, the polymixin antibiotics are inactivated ig‘gitrg by tissues because they bind to the phospholipids of cell membranes; polymixin B persists in liver, kidney, brain, heart, muscle, and lung for as long as 72 hours.103 The binding 19 is through electrostatic attraction to the negatively charged phospho- lipids of membranes.37v103 Another interesting example of the binding of the drug is in XEEEQ degranulation of mast cells by polymixin 8.104 While the exact significance of this is unknown, in one study the control animals given polymixin B sulfate alone developed a leuco— cytosis, possibly secondary to in gizg mast cell degranulation, and the release of chemotactic factors.37 Overall, the tendency of the polymixins to bind the mammalian membranes may be an important feature of their toxicity, but it does not appear to account for their benefi- cial effects. Polymixin B is bactericidal and is generally more active against gram-negative than gram-positive organisms. Susceptible bacteria absorb considerable amounts of the drug, which is a surface active agent, much like a cationic detergent. Detergents disorganize cell membranes and denature certain proteins, and similarly, polymixin combines with and disorganizes cell structures that are responsible for osmotic equili- brium; combination of polymixin B and sensitive bacteria results in leakage of low molecular weight cell constituents and thereby kills the cells. More specifically, polymixin B combines with the inner layer or membrane of the bacterial cell wall, the region of the cell responsible for endotoxin activity.30 The bacterial action of the polymixins is probably due to complexing of the drug to bacterial lipids; very signi- ficantly, polymixin B binds to the lipid A portion of endotoxin.27129 The stoichiometric relationship appears to be one to one with the for- mation of complexes with higher molecular weights.29 These complexes are reversible, and because the bond forms at the lipid A region, all types of endotoxin should respond similarly to polymixin 8.29130 20 Modification of Endotoxin Pathophysiology by_Polymixin B The cyclic cationic polypeptides, eSpecially polymixin B, directly neutralized endotoxin in a series of experiments in which the antibiotic was admixed with endotoxin prior to administration.28r31 In various other experiments, primarily in laboratory animals, polymixin B decreased the lethality, neutralized the coagulation derangements, and abrogated the hematologic and hemodynamic aberrations induced by endotoxin.“I3:’-I34"'43I105 In most of these experiments, the polymixin B and the endotoxin were combined before injection. However, in one experiment, the antibiotic was given before and after endotoxin, and survival rate still improved significantly.43 In another trial, poly- mixin B admininstered after endotoxin injection ameliorated the endotoxin-induced coagulopathy.41 And, during a clinical trial, poly- mixin B did not improve the survival rate in patients with coagulo- pathies secondary to liver cirrhosis, but it did partially ease the coagulation disorders.44 An ideal endotoxin-blocking dose of polymixin B has not been derived in any species, but most investigators have used a dose of 2.5 mg/kg of body weight. To be effective, polymixin B has to combine with the lipid A moiety of endotoxin before it becomes fixed in tissues. With some bacteria, the parent bacterial susceptibility and the suscep- tibility of the derived endotoxin to polymixin B are related; however, these two qualities can also be quite divergent.106 Whereas sensitive bacteria absorb much larger amounts of the antibiotic than do resistant strains, and polymixin B usually neutralizes endotoxin derived from sen- sitive microorganisms more efficiently than that derived from resistant bacteria, it is still effective against the endotoxins of nonsusceptible 21 bacteria.28r33 The beneficial mechanism of action of polymixin B on endotoxin-induced pathophysiology occurs through combination of the drug with the lipid A portion of the molecule, i.e., the portion of the mole- cule that is constant between bacterial species and responsible for the bioactivity of endotoxin. Experimental Rationale The spectrum of clinical endotoxin syndromes ranges from clinically normal animals with pyrexia and/or hematologic alterations (absolute leucopenias) to massive, lethal shock. But the majority of experimental studies, particularly in horses, has been done with lethal doses of endotoxin.9‘15 Thus, the pathophysiology of nonlethal endotoxemia is not well defined. Similarly, in the treatment of endotoxemia and endo- toxin shock, most of the agents or techniques used are directed at various manifestations of endotoxin pathology and not specifically at the endotoxin molecule.8v17l18'25 There is a direct interaction between polymixin B and the lipid A portion of the endotoxin molecule.23v29l31r33i106 Lipid A is the bioactive portion of the endo- toxin molecule and is present in all commercially prepared and naturally occurring forms of endotoxin.16I25I46 With these factors in mind, this study was designed to investigate sublethal endotoxemia, and to attempt to block or modify the resulting pathophysiology. Specifically, the febrile, hematologic and hemodynamic responses were measured and eva- luated in ponies given very low doses of endotoxin, and comparisons of these parameters with and without polymixin B were made in an attempt to evaluate the potential of this antibiotic as a therapeutic adjunct in the treatment of clinical endotoxemia. MATERIALS AND METHODS Experimental Animals Mature, clinically normal Shetland ponies (half were females and half were castrated males), weighing 130 to 240 kg, were brought indoors and allowed to acclimate to their surroundings for at least 24 hours before being readied as experimental preparations. All animals were examined and were found to be free of clinically detectable disease. A total of 14 experiments was performed. Surgical Procedure Twenty-four hours prior to the experimental period, the ponies were anesthetized with 0.2% thiamylal sodium in a 5% glyceryl guaiacolate plus 5% dextrose solution. Following intubation, anesthesia was main- tained with halothane and oxygen. The right side of the neck was clipped and surgically prepared, the skin and cutaneous colli muscle were incised, and with blunt dissection, the external jugular vein and the carotid artery were exposed. After freeing these structures from the surrounding tissues, the carotid artery was cannulated with a No. 7 French teflon end-hole catheter (United States Catheter and Instruments Corporation, Glen Falls, New Ybrk) which was filled with heparinized saline (0.05% sodium heparin in physiological saline, beef lung extraction, Upjohn Company, Kalamazoo, Michigan) and secured with catgut ligatures (American Cyanamid Company, Pearl River, New YOrk). This catheter was used to measure systemic arterial pressure. Two catheters 22 23 were introduced into the jugular vein; a No. 7 French balloon-tipped Swan-Ganz triple lumen catheter (Columbus Instruments International Corporation, Columbus, Ohio) was positioned in the pulmonary artery, and a second No. 7 French end-hole teflon catheter was positioned in the right atrium. Both were filled with heparinized saline prior to placement. The Swan-Ganz catheter, which has a thermistor near the tip, was used to measure body temperature, pulmonary arterial pressure and the changes in blood temperature during cardiac output determination. The second teflon catheter was used to record right atrial pressure and served as the injection port for the cold saline bolus used in the car- diac output determination by thermodilution. Both of these catheters were secured with catgut ligatures, and the skin was closed with a nonabsorbable suture material (Suprylon, Pfrimmer, West Germany) in a simple continuous pattern. All catheters were connected to pressure transducers (Statham P23 Db, Statham, Hato Rey, Puerto Rico), and all catheter placements were verified by pressure measurements (PDV-22 pressure preamplifier, Electronics for Medicine, Incorporated, White Plains, New Ybrk) and by the shape of pressure tracings recorded on a two-channel recorder (Model 2M-SB, MFE Recorders, Salem, New Hampshire). Transducers were calibrated prior to each experiment against a mercury manometer. After catheter placement was verified, all catheters were secured by a bandage, and the ponies were allowed to recover from anesthesia. Following recovery, each pony was returned to a stall and was given free choice of water and feed. The next day, the pony was placed in a specially designed restraining box and was allowed to become accustomed 24 to the box and equipment for 60 to 90 minutes before control values were taken. Catheter patency was maintained with periodic flushes of hepari- nized saline, and the pressure transducers were placed at the level of the base of the heart (approximately the point of the shoulder). Experimental Protocol Control measurements were made of the following variables: syste- mic arterial blood pressure (systolic, diastolic and mean), mean right atrial pressure, pulmonary arterial pressure (systolic, diastolic and mean), body temperature, cardiac output, packed cell volume, total white blood cell count, and differential count of white blood cells. After control measurements were taken, each pony was given one of the following solutions at the following level: 1 pg/kg body weight Escherichia coli 026:36 Boivin extracted endotoxin, lot 639746 (Difco Laboratories, Detroit, Michigan) in physiological saline at a con- centration of 1 ug/ml; 2.5 mg/kg body weight polymixin B sulfate U.S.P. (Pfizer, Incorporated, New Ybrk) prepared by dissolving the micronized powder in 5% dextrose solution and sterilizing by microfiltration (Nalgene Labware Division, Sybron corporation, Rochester, New Ybrk); 2.5 mg/kg body weight polymixin B sulfate and 1 pg/kg body weight endotoxin; or sterile physiological saline. All solutions were given by slow infusion (average infusion time approximately five minutes) through the right atrial catheter, with the exception of those experiments in which both endotoxin and polymixin B were given. In those, the two agents were given through separate catheters simultaneously: the poly- mixin B through the right atrial catheter and the endotoxin through the 25 proximal hole of the Swan-Ganz catheter. A total of 14 experiments was performed: 5 endotoxin only, 5 endotoxin plus polymixin B, 2 polymixin B only, and 2 physiological saline only. Measurement of Blood Pressure Blood pressure was measured directly using the heparinized saline filled indwelling catheters connected to the transducers, amplifier and recorder. Catheters were assumed to be accurately measuring blood pressure when blood could easily be aspirated through them, and they could readily be flushed. With the electronic damping of the recorder, the mean pressure values were recorded. Measurement of Cardiac Function Cardiac output was measured by thermodilution;107"109 a ten ml volume of cold physiologic saline (0-2°C) was injected manually as rapidly as possible into the right atrium through the teflon catheter. The blood temperature change in the pulmonary artery was sensed by the thermistor at the tip of the Swan-Ganz catheter, the other end of which was connected to the cardiac output computer (Cardiotherm -500, columbus Instruments International Corp., Cblumbus, Ohio). The temperature of the injectate was detected by a second thermistor connected to the computer. At least five cardiac output determinations were made, and an average of these was recorded for each time period. Cardiac output and blood pressure readings were taken prior to infusion, at 15 minutes, at 30 minutes, and hourly from one to four hours following infusion. Times were measured from the finish of the infusion. Additionally, in the endotoxin experiments, the first noti- ceable blood pressure rise was 10 minutes after injection, and in the 26 endotoxin-polymixin B experiments, the first noticeable pressure increase was at 50 minutes. At these times, additional recordings of cardiac output and pressures were made. Hematologic Parameters All blood samples were collected in evacuated glass tubes con- taining EDTA for anticoagulation (EDTA vacutainer; Becton, Dickinson and Co., Rutherford, N.J.). The blood samples were drawn from the carotid artery catheter immediately prior to infusion, at 15 minutes, at 30 minutes, at 1 hour, and at hourly intervals from 2 hours to 10 hours, and at 24 hours after infusion. Again, times were measured from the finish of the infusion. The packed cell volumes were determined by the microhematocrit method using a microcapillary centrifuge (International Equipment Co.; Boston, Massachusetts), and plasma solids were determined by refractometer (American Optical Corporation; Buffalo, New Ybrk). White blood cell counts were determined manually by using a hemocy- tometer (Spencer Bright-Line, American Optical Co., Buffalo, New YOrk) charged to contain a 0.9 mm3 volume. The hemocytometer was filled from a Unopette dilution vial (Unopette Test 5855; Becton, Dickinson and Co., Rutherford, New Jersey) which has a buffered ammonium oxalate solution that hemolyzes mature red blood cells and preserves the platelets, leucocytes and reticulocytes. The Unopette system has a capillary pipette that fills automatically to a 20 pl volume which, when mixed with the vial's contents, becomes a 1:100 dilution. After allowing 10 minutes for the red cells to hemolyze, the hemocytometer was charged, and the number of white blood cells in its block of nine squares was 27 counted. Ten percent of the count was added to the number of cells counted and the sum multiplied by 100 to arrive at the number of leucocytes/mm3. The differential leucocyte count was determined by microscopic examination of 100 cells of a wright's stained blood smear. All smears were made immediately after the collection of blood. Statistical Analysis Data were analyzed in a two-way analysis of variance, each animal as its own control, in a completely randomized block design. Mean dif- ferences were compared by the Student-Newman-Keuls' test. The Student's t test was used to detect any significant difference between means in the endotoxin experiments and the endotoxin and polymixin B experiments when a significant change occurred in one of these sets of experiments. The coefficient of variability was used to determine the relative variability of the sets of experiments. All comparisons were made at the p=0.05 level. The use of the coefficient of variability allows com- parison of variability, thereby indicating the presence or lack of uni- formity within the data for a given set of experiments.11or111 RESULTS Clinical Observations In all of the experiments, the ponies remained bright and alert, and their oral mucous membrane color and capillary refill time remained normal. At about 10 or 15 minutes after endotoxin injection, the ponies' pulse and respiration rates increased moderately, and they began to tremble. The trembling decreased gradually and stopped after about 2 hours. Also, at about 15 to 30 minutes after injection, the ponies ”rattled” as they breathed, as if they had fluid in their tracheas. This lasted only a few minutes, and during this time, several of the ponies coughed a few times. No exudate was ever seen at the noses or mouths of these ponies. Of the group given endotoxin plus polymixin B, three also began trembling, but in these the trembling began about 50 to 60 minutes after injection. The trembling stopped within 20 minutes in all but one pony which continued to tremble for about 1 hour. Also, one of these trembling ponies coughed several times after the trembling began. The body temperatures of both the endotoxin group and the endotoxin plus polymixin B group were significantly elevated 1 hour after infusion (Figure 2, Table 1). The time courses of the temperature rises of these two groups were parallel out to the 3 hour sampling time. The mean maximum temperature of the endotoxin group was 38.8°C at 4 hours post-infusion. 28 29 By 24 hours, the body temperature returned to normal. In the endotoxin plus polymixin B group, the mean maximum temperature was 39°C, which was reached at 2 hours. 30 .mcmofi consmcHIoum 0>Huoommou scum acouommwo hausmoauecmam mum mucwom omen» ha oousomoummu mcmofi noon u hav .mumou .mHsoMIcmssoziucoosum was moccaum> mo memwamcm xooan ouoamaoc oouaeoocou mcams .ocam> coamcwcaloum Bonn Amo.oumv omcmno ucmowmacmwm haamoaumaumum u a .Nnc .ocaamm Hashes 3m mxxss P spa: eoamsmca mouaoaeea mzHaam .mue .m eaxaamaoa 3m mx\ms m.~ :uw3 scamsmca moumoaoca mm .mu: .m swxflahaom 3m mx\ma m.N nuaa hausoocmuassam saxou uoecm 3m mx\m: F spa; eoamsmea mmumoaeea mmuoozm .muc .eaxouoecm 3m mx\ma P nus: coemSMCH moumoapca 002m .00 ca mosam> scamswsaaonm Scum ousumuomEou moon cw momcmno N HMDOHW 31 . cgghc... «won 950—..— vm .I w m a L. m m. I‘Vru Imwr. .0 Imw. \\ -v. ./ . .m. .o . 0.20.888. . 6. seam .. . t _ .o._ a .m._ * mznmm ll .3 * 2.82“. l... oozu l .m._ N mmDUHh 32 TABLE 1. Statistical comparisons between means of body temperature (0C). Endotoxin pre-inf. 24 hr. 15 min. 30 min. 1 hr. 3 hr. 2 hr. 4 hr. 37.4 37.5 37.7 37.9 38.3 38.6 38.7 38.8 Endotoxin plus Polymixin B pre-inf. 24 hr. 15 min. 30 min. 4 hr. 1 hr. 3 hr. 2 hr. 37.5 37.5 37.6 37.8 38.2 38.4 38.7 39 (Means underscored by the same line are not significantly different at the 0.05 level.) Coefficients of variability - body temperature CVENDO = 0 o 86% CVENDO+PB = 1 o 21% 0.40% CVSAL 33 In this group, the mean body temperature at 4 hours approached the pre- infusion temperature mean. In the polymixin B treated and the saline treated groups, no significant differences in temperature were seen. Student's t test performed on pre-treatment, 1 hour, and 24 hour means showed no significant difference between the body temperatures of the ponies given endotoxin and those given endotoxin and polymixin B. Hematologic Effects Following endotoxin infusion, the total white count began to decline sharply; by 15 minutes, a significant decrease in circulating white blood cell count appeared (Figure 3, Table 2). This decline con- tinued until it reached its lowest point at 1 hour. From there, the mean count began to rise and approached the pre-infusion level at 8 hours. A significant increase had developed by 24 hours. Similarly, in the endotoxin plus polymixin B group, the total white count was also lowest at 1 hour but remained within the normal range for horses. Unlike the endotoxin group, there was a significant increase by 8 hours that persisted to 10 hours. At 24 hours, the total white count had returned to pre-infusion levels. In the polymixin B treated group, total white counts fluctuated only slightly with a significant increase 4 hours after the infusion of the antibiotic. That increase did not persist. The saline treated group showed even less variation in white count over the experimental period, and at no time did a significant difference from the pre-injection mean appear. Student's t test per- formed on the pre-treatment means, the 1 hour means and the 24 hour means of the ponies treated with endotoxin and the ponies treated with endotoxin plus polymixin showed a significant difference between the two treatment groups only at the 1 hour comparison. 34 .oo:mwum> mo mwmmHmcm hmslosu an mosHm> coamsmcaloum Eoum occouomuwo unmoamacmwm mucomoumwu a .Nuc .ocaamm HmEuoc 3m mx\HE 9 one: concmca moumowoca mquam .mu: .m caxasxaom 3m mx\ma m.~ cua3 scamsmcw moumcaoca mm .mu: .m :Hxaamsom 3m waxes m.~ nus: sawsomcauassam saxouoeao 3m mx\mn P nus: aoamsmea moumoaoca mmuoozm .muc .Conuooco 3m mx\m: P spas cowmsmcw moumoaoca oozm .mafi\mHHoc mop cw rm .mosHm> scamSMCHIoum some mucsoo Haoo vocab ouas3 mcepmasouao ca mmmcmco m mmeHm 35 532. .. Son. 23... vm O. m w v N _ a n J? (I b n r h b b 0.... 9.... mz_._u am.m u mm>o wv.mp mm+oozm>o mm..~ u oozm>o mason eooan moan: : suasanmana> «0 mucoaoammmoo .mEE\mHHoo omh.~p was some 0:8 .moficom powwow» ocHHMm on» you mucoswuomxo ecu usonmsounu mucsoo Haoo ocean ouaga 0:» ca oocouomuao ucmoamasmam 0: mm: whose A.Ho>oa mo.c on» um ucououueo maucmoamacmam no: mum mafia each can an omuocmuooss mcmozv cos.o. oov.vp oom.m. omm.m. oom.mp ooe.n. ocm.m. oom.~. oo..~. own... .ua v .ucauoua .caa m. .un N .ue o .ue . .aas on .HB op .ue m .ue an m caxHESHom omm.a. omv.mF cam.ms cmo.m. omm.m. omm.m. omn.FP com..P cqm.o. cv~.m .un m .ua c. .ucauoua .u: am .ue m .aaa m. .un v .eaa on .un N .un . m sflstxaom cam saxouoosm omv.mF ooF.~_ coo... oo~.o. amp.“ oom.o oo~.m oms.a cmm.m cam.~ .u: «a .un o. .mcaumum .un m .ue 0 .cas mp .u: a .caa on .un N .nn P saxouoosm .AmEE\mHHoov mussoo Haoo woods ouas3 mcaumascuao mo mcmofi coozuon mcomwummEOo Hmoaumaumum .N manta 37 Most of the endotoxin-induced changes in the total leucocyte counts were attributable mainly to the changes in segmented neutrophils (Figure 4, Table 3). Again, the greatest decrease in cell numbers occurred 1 hour post-infusion. This decrease was profound with a pre-treatment mean of 6,801 cells/mm3 going to a 1 hour mean of of 202 cells/mm3, a 97% reduction. Mature neutrophils began returning to circulation after 1 hour but did not approach the pre-infusion level until 8 hours. An absolute increase of 11,510 cells/mm3 was present at 24 hours, representing an increase of 69% over the pre-treatment level. In contrast, the 1 hour mean for the endotoxin plus polymixin B group was 5,139 cells/mm3, and was 56% of the pre-treatment count. By 4 hours post-infusion, the mean count approached the pre-treatment mean and con- tinued to increase to a maximum of 12,030 cells/mm3 at 8 hours, a 31% increase over the pre-treatment level. There was no significant dif- ference in the segmented neutrophil means throughout the experiments for either the polymixin B-treated or the saline-treated ponies. The coef- ficients of variability show mild to moderate heterogeneity of the data. Student's t tests performed on pre-infusion, 1 hour and 24 hour means between pony groups treated with endotoxin and those treated with endo- toxin plus polymixin 8 showed a significant difference only at the 1 hour comparison. Non-segmented neutrophil counts basically followed the patterns of the segmented neutrophil counts (Figure 5, Table 4). In those animals given endotoxin, the few non-segmented neutrophils initially present disappeared by 1 hour. At 4 hours, the non-segmented neutrophils were increased over pre-infusion levels, and at 10 hours, a mean of 1,605 cells/mm3 represented 13% of the total white count. At 24 hours, 38 .wUGMfium> mo mflmhamcm hmslo3u an meHm> coflmSMGwlmum Eoum muconMMHv ucmowuflamfim mucmmoummu a .Nnc .mcHme Huauoc 3m mx\HE p nuwz conSMGA wouMUAUCH mqudm .Nu: .m cwxashaom 3m mx\mfi m.N nuw3 :OHmSMGa moumoaocfi mm .mu: .m caxafihaom 3m mx\ms m.~ nuH3 aamsoocmuasswm saxouowco 3m mx\m: P £Ufi3 GodeMGH moumowoca mmloozm .mus .cwaUOUcm 3m mx\m: P nuw3 scamamca houseflonw 092m .mEE\mHH00 mop ca .m .mo5am> coamSMcfilmum Eouu mucdoo Hanmouusw: coucwsmom mcfiumasouwo cw mwmcmno v mmDUHh 39 .8635 .. .31 m m b 23: v b Q. C O. O . COOOOOOOI ..... v mmDOHm ........ o .000... ...0...0.00...000....oo0...n.000 .ooo 0.’...0....‘ wz_._o & —. P " mm>0 &MN .II. mm+ODZW>U arm n oozm>o ”maflnmouuswc kucmsmom muaaanwaudb mo mucmwodumoou .aa0>wuommmmu mEE\mHHoo hmm.m can moh.h mums mecca on? .mwwcom owummuu ocaamm on» no oouwwuu m cwxflahaom on» Honuao new mucmsuuomxm on» usonmsounu nucdoo Hanmouuso: voucoamwm onu cw mosmuommav ucmowuaamwm o: was mums? A.Ho>oa mo.o on» an usmu0MMfio haucmoqufiamam no: can mafia mama mnu ha concomuovcs mammzv omo.~F omn.o. o~P.oP n~m.m h...m mom.m Ppm.m .eo.m sup.h mm..m .us m .us o. .u: o .n: v .mcaumum .u: «a .cas on .csa m. .un N .un . m fixdfinaom madam 5389065 oPm.F. mmm.n .cm.m mmm.m no..e hmm.~ m-.~ mmm mew «an .un an .u: as .mcaumum .ug m .u: o .un w .cHe m. .cws on .us m .ug P :axouoocm .AmsE\mHHmov mucdoo Adamouusoc Uwucofimom mcauwasouwo mo memos cmmzumn mcomwummfioo Hmowumwuwum .m wands 41 .oocmfium> mo mamaamcd %M3 :03» up mosHm> coamsmcanoum Bonn oocmuowmwc ucmonflcmHm mucommummu a .Nua .ocaamm Hmauoc 3m 3).: F :33 .8335 33035 $35. .Nu: 5 5x333 en mi? m.~ fie. c3265 nonwowcca mm .mu: .m :waahaom 3m mx\mE m.~ :ufl3 hamsoocmuaaawm saxouoocm 3m mx\mn F nua3 COwwomcH moumoflvcw mmloozm .mu: .zm mx\m: p nus: coamSMGH mmumoflvsfi oczm .msE\mHHwo mop ca .m .modam> coamsmcfiumum Bonn mucsoo Hanmouuso: voucmsmmmlco: mcwumasouao ca momccnu m mmDGHm 42 533:. .. Boa £3: m m w w _ 00000 Ki ...... ..,... ......‘............O‘.O.Ol.. 88.03838. . m23o mmm n mm>o «no. u mm+oozm>o mom u oozm>o maflnmouusm: Umucwamomucoc I auwafinwwuw> mo mucwwowmmooo .mao>quowmmou m8§\maawo as can he? .mnm mums memos was .mwdcom powwow» ocwamm can so moacom powwow» m cwstaaom «nu .mmficom wwumouu m :HXfiamaom mafia caxouovco may you magma Iauwmxm on» usocosousu munsoo Hasmouusoc voucmsmomlcoc 0:» ca moocmuommwv ucmoauasmam on who: awash A.Hm>oa mo.o on» no acoumMMHo haucmowmwcmfim no: mum mafia 080m on» an couoomuwccd mammzv m~m.~ moo.. Pq~.F mom «mm mmw mm mm mm o .u: an .un as .un m .un m .un v .mcfinoum .cHa mp .u: m .c.5 an .un F saxouovcm .Amfia\maaoov mucsoo Hanmouusmc kucmfimmmlco: mcwumasouflo mo mammfi cmm3uon mcomflummfioo HMOHumHumum .v mqmda 44 non-segmented neutrophils made up 14% of the total white blood cell count. The 10 hour and the 24 hour means for the endotoxin treated ponies were significantly greater than the pre-treatment means, but on Student's t test comparisons between this group and the endotoxin plus polymixin B treated animals, there was no significant difference between the pre-infusion, 1 hour and 24 hour means. The coefficients of variability for these experiments were extremely large, reflecting the tremendous variability in the non-segmented neutrophil numbers (range within the endotoxin experiments: 0 to 2,528 cells/mm3). The changes in circulating lymphocytes were not as dramatic as were those in the neutrophils (Figure 6, Table 5). In the endotoxin group, a decreasing trend was maximal at 6 hours with a mean of 1,797 cells/mm3, which represented 52% of the pre-infusion mean. At 24 hours, the mean of 3,372 cells/mm3 was only slightly above the pre-infusion mean of 3,435 cells/mm3. Over the course of the endotoxin plus polymixin B experiments, lymphocytes declined to 2,469 cells/mm3 at 4 hours, a 37% decrease. At 24 hours, they rose to 4,993 cells/mm3, a 27% increase. There were no significant differences in lymphocyte means in the poly- mixin B or saline experiments, but at 4 hours a sharp increase in lymphocyte numbers was noted in the polymixin B treated animals. The means were 5,420 and 7,010 cells/mm3 respectively. Student's t test comparisons between means of endotoxin and endotoxin plus polymixin B experiments on pre-infusion, 6 hour and 24 hour means indicated no significant difference between these experiments at the 0.05 level. There were no statistically significant differences in the monocyte counts in any of the four treatment categories. The means for the 45 .Nuc .ocfiamm Hesse: 3m mx\Hs F nus: :oamnmcs nonsusccs mquam .Nuc .m :Hxasmaom 3m mx\as m.~ nus: cosmaucs mouMOflUca mm .mns .m :flxaahaom 3m mx\mE m.~ nuaz hamsoocmuasawm caxOpocaw 3m mx\mn F sues cosm5mcw woumowocfl mmloozm .mu: .caxouopsm 3m mx\mn P nua3 coamsucw moumoaucfi 092m .mES\mHHwo mo— cw .m .mmzam> cowmswcwlmum Scum mucsoo mphoonmsha mcwumasoufio ca mwmcmco w EDUHW c232. .. Ben. «.52.. e m .0 .00 . . .. ... . . .. ... a 6 0.0 00. 8. ’ . 0 4 . 3...... . O...- ..-..$"..h: :0. . n .0. .0. 0 .0 .0 O 0 “g ' -' ‘. .. 0“. . . . .. .. 0 0 .. 0.. 0 . coo 0. 0 . . 0 .. .. .. .. o MMDUHW m23o mm. u mm>o flmfl fl mm+ODZH>U arm n oozm>o moumoonmsha I muflawn0H00> mo mucmwowmwwoo .%Hm>wuoomwmu msfi\maamo opo.h can oov.m who: mamme 0:9 .moacom woumouu ocwamw man no Uwuwwuu m cwaE Ihaom on» now macoswuomxm on» unonmsounu mcmofi ouhoosmsha 0:» ca wwocmH0MMHU 0:00amacmwm on 0003 whose A.Hm>oa mo.o on» an acoHomMflo hausmowmwcmwm no: mum mafia 080m 029 an concomuovcs mcmozv mam.¢ mnm.e mov.e mmm.m mmm.~ msm.~ hvm.~ mum.~ vm~.~ mmv.~ .ua o. .un m .0: cm .ucsuoum .csa m. .ua 6 .un u .:45 on .un P .un v m :Axwahaom msdm :onuoocm ~me.m .mw.m mm¢.m nmo.m mme.~ ~mm.~ omv.~ vm..~ mmo.~ emn.. .ua em .caa cm .ucnumnm .cfia m. .un N .ua P .ua o. .u: 4 .us m .ua o anxouoecu .AmEE\mHHon mundoo muaoonmaha mcaumaaouwo mo memos newsman mcomaummaoo HMOflumHumpm .m Manda 48 endotoxin, the endotoxin plus polymixin B, the polymixin B, and the saline groups were 239, 176, 205, and 228 cells/mm3 respectively. The variability of this data was enormous; the coefficients of variability for the endotoxin, the endotoxin plus polymixin B, the polymixin B, and the saline groups were 118%, 90%, 83%, and 50% respectively. A slight change was noted in the packed cell volume in the animals treated with endotoxin without any accompanying increase in total plasma solids. The mean PCV started to increase immediately after infusion, and at one hour was 34%, an increase of 31% over the pre-treatment mean of 26% (Figure 7, Table 6). The endotoxin plus polymixin B group had no significant change in PCV until 24 hours when it increased. The Student's t test comparison showed a significant difference between the 1 hour means but not the pre-infusion or the 24 hour means of the endo- toxin and the endotoxin plus polymixin B experiments. In the polymixin B groups, one pony had an unexplained increase in PCV at 24 hours, and in the saline group, one pony had an unexplained decrease in PCV at 24 hours. The coefficients of variability show mild heterogeneity of the PCV data. Hemodynamic Effects Systemic arterial blood pressure did not change significantly in any of the four experiments (Figure 8). However, some trends appeared; at 10 minutes in the endotoxin treated animals, systemic arterial pressure increased sharply, at 1 hour fell to its lowest point and at 4 hours increased again. In the endotoxin plus polymixin B animals, systemic arterial pressure decreased very slightly, then rose above pre-infusion levels at 1 hour and returned to pre-infusion levels at 49 .oocmwum> mo mammamam mmslozu ha modam> downswcanonm Bonn oocouomuwo unmowwacmflm mucmmoummu a .muc .333 358 3m 9Q? P :33 5335 $3365 335 .Nu: .m :Hxsafiom 3m 33: m.~ an? 8335 ”"3835 mm .mua .m fifiaflom .5 mi? m.~ fin. sagomafiasfim saxouoocw 3m mx\mn F cuss GOHm5m:H mmHMOflocH mmlonzm .mu: .cwxouovcm 3m mx\m: P spa: coamsmcfl moumoaucw 002w .m as .m .mmsam> coamsmcwnmum scum mesao> Hamo omxomm cw momcmnu h MEDOH .m c232. .. 30¢ 23: SO .8 my 4. .w _ n. J. W F . . i 40... :01 um.- ........ ..0’... .0...0.00. . °\° mE:_o> :00 screen uz..._ Haoo voxomm I auaafinoaum> mo mucoHOfimmooo .wmm mm: >Um cums ommumbm any .msoum man» cH an n ao as u mm>o #0 an mm+82W>U a_— u oazm>o .mowcom poumouu m :wxfifihaom on» now mucosauwmxw on» usonmsounu moasao> Have voxomm 029 :w mocwuommao ucmowuasmwm o: mm3 whose A.Ho>ma mo.o 0:0 um unonMMflp handmOHMflcmHm nos mum mafia 050m way an noncomuopcs mcmozv cm on mm mm mu mm hm hm 0N mm .H: o— .u: v .HS 0 .wanoum .swa mu .un m .un F .Hn N .cwa on .u: cm ocmem mm um um mm mm mm mm mm «N mm .u: vm .Hn o— .Mn 0 .u: N .un v .HA 9 .GHE om .mcHIon .Hn m .caa mp m caxfiahaom maam saxouowsm em Pm Pm on on on mm mm mm mm .un P .u: N .:H5 on .u: 6N .H: v .H: m .caa m— .Hn or .u: o .mcHIon :«xouoocm .va moESHo> Haoo Umxomm mo memos cwmzuon mcomfiummsoo HMOfiumflumum .w wands 52 .Nu: .mcfiamm Hmauoc 3m 932 P and. .8332: ”“3835 mzaqam tans 5 55530.“ .5 3me m.~ :33 8335 mmumowvcw mm .mu: .m cHXHEmHom 3m mx\mE m.~ nuwz hamsowcmuasaam :onuoosm 3m mx\m1 F 092m :0w3 cowmsmcw mmumowocu mmloczm .mn: .saxouovco 3m mx\mn — nufiz cosmsmca mouonocH .mm 58 ca .m .mmsam> cosmsmcfiImHm Scum musmmmum vocab Hmwumuum oasmumhm ca momcmno w MmDOHh c222. .. Boa use: 4 m w L. a i 53 ‘ 0 EE 0 I ........-... chimp—Q .m .6223. III 2.5 m» mz_._ mo mflmhamcm hmsIozu ha wsam> coamsmcHIoum muH scum pcmHoNMfiw “anamowmwcmfim ma .cHE on um osHm> m casthaom msam saxouoocw on» xaco .Aav vmxume mamas on» no .mo:manm> mo mflmwawcm hmzlosu ha mosam> scamsusanum scum oocoumwmac guacamacmam mucomoummu 1 .Nu: .ocaawm Hague: 3m mxxaa P nuns cosmsucs mmumoncca manam .Nua .m :Hxaassom 3m mx\ma m.~ nuns scam IDMGH moumowvcfi mm .mus .cfixwahaom 3m mx\ma m.~ :uHB hamsoocmuasswm cfixouoccm 3m mx\m: P can: :onsmcn mmumonecn mmuoazm .mu: .caxououcm 3m mxxm: P and: cosmsuca mmumoscca oozm .mm as cw .m .mosam> cosmsuconum Scum whammwum vocab Hmauouum mumcoaasm ca momcmno m mmDGHm c232. I amen. £3: 56 . m a Fin. . uO-I ..................<..u1..u1:w..u. ........u.u.u........ O OISE O_ 053mmOLQ 3:23 * .ON 36:68.5 mz_._ mo mflmhamcm aszozu an mosam> GOfimSMCHIon Bonn oocoummwflv unmodmacmwm uncommummu « .mfl: .ocwamm Hesse: 3m mx\HB P nua3 cowm5mca moumowcca mqudm .mu: .m :«xwahaom 3m mx\mE m.~ cpwz cowmswcw moumowccfl mm .mu: .m :HXaahaom 3m mx\ma m.N nua3 hamsoocmuHSEHm saxouocco 3m mx\mn p cues consMCfi mounowocfi mmIoazm .muc .caxouovco 3m mx\mn P nus: conSMGH woumo IHocH 002m .CHE\A\mm SE cw wocmumamou Haasomm> mumcofiasm can mocmumflmou Hmumzmfluom dance or mmDGHh 61 £033.:- I “won. 950—... N m N _. p P _ _ . no 8838338...“ ,3..- .833... -13... ...............0... e . / {.1 wz_n_