Restriction of cell lysis by homologous complement: II. Protection of erythrocytes against lysis by newly activated complement

Our previous work revealed that homologous complement (C) was ineffective in lysing antibody-sensitized erythrocytes (EA) even at high concentrations. It was also shown that activation of complement on homologous EA resulted in the binding of C9 and the formation of EA bearing complement proteins C1 through C9 (EAC1-9), yet few hemolytic sites were formed. Instead, as shown here, the formation of homologous EAC1-9 caused the cells to become resistant to lysis even by heterologous complement during a second incubation. In contrast, when homologous EAC1-8 were produced by incubating EA with C9-depleted serum, such intermediates were not protected against lysis by heterologous complement during a second incubation. Furthermore, homologous C9 on EAC1-9 was able to reduce the hemolytic efficiency of heterologous complement without blocking C activation and the formation of new C5b-9 complexes. Protection was not modified when homologous EAC1-9 were produced in one step, by incubation of EA with serum, or sequentially by adding C9 to EAC1-8. The minimum number of 9-sites required to confer a protective effect on EAC1-9 was less than 200 per cell. Thus, in addition to its known effect in heterologous cell killing, homologous C9 is capable of protecting homologous cells against inadvertent complement lysis.     

Hemodialysis-associated platelet activation and thrombocytopenia

The interactions between platelets and dialysis membranes were studied prospectively in 10 patients undergoing long-term stable dialysis. Transient but significant thrombocytopenia and platelet activation were found during dialysis with the commonly used cuprophane membrane. Platelet counts decreased from 231 +/- 21 X 10(3)/mm3 before dialysis to 127 +/- 28 X 10(3)/mm3 at 90 minutes following initiation of dialysis (p less than or equal to 0.007). Thromboxane B2, an index of platelet activation, also increased from a baseline level of 1.06 +/- 0.2 pg/10(6) platelets to 7.3 +/- 3.0 pg/10(6) platelets at 90 minutes (p less than or equal to 0.04). Cuprophane membranes were also shown to induce complement activation with C3a desArg, the stable derivative of C3 activation, showing a threefold increase from baseline 15 minutes after initiation of dialysis. In contrast, during dialysis with a non-complement-activating dialyzer membrane, polymethylmethacrylate, thrombocytopenia and platelet activation were not observed. These data suggest that platelet activation and thrombocytopenia during hemodialysis are associated with complement activation during hemodialysis in a manner similar to dialysis-associated neutropenia.     

Hemodialysis-associated induction of cytokines

The interleukin hypothesis relates chronic pathology in long-term end-stage renal disease (ESRD) patients to repeated stimulation of mononuclear cell cytokine production during hemodialysis. In vitro experiments demonstrated different possible mechanisms involved in hemodialysis-associated cytokine induction: adherence of mononuclear cells to the dialyzer membrane; complement activation, and the passage of cytokine-inducing bacterial fragments from contaminated dialysate through the dialyzer membrane into the blood. Studies investigating mononuclear cells from ESRD patients ex vivo suggest that these cells become activated during hemodialysis with complement-activating membranes and that the type of dialyzer membrane may influence mononuclear cell cytokine production in response to endotoxin. According to biological assays, plasma levels of interleukin-1 but not interleukin-6 activity seem to be elevated in ESRD patients compared to normal subjects and may increase further during treatment depending on the choice of the dialyzer membrane. However, to date, partly due to insufficient assay sensitivity and circulating inhibitors, measurements of interleukin-1, interleukin-6 and tumor necrosis factor in plasma by specific immunoassays could not finally prove elevated plasma cytokine levels in ESRD patients. Since circulating mononuclear cells are a major source of cytokines, studying the activation of these cells ex vivo seems to be the best approach to study hemodialysis-associated cytokine induction.     

Effect of RBCs on the activation of human complement by heparin-protamine complexes

Complement activation on red cells by heparin-protamine complexes was studied by using whole human serum. C3 bound to red cells was measured by radiolabeled monoclonal antibody to C3, and fluid-phase C5a was determined by radioimmunoassay. Heparin and protamine in clinically relevant concentrations caused the binding of C3 to red cell membranes, and the measurement of C3 binding provided a sensitive indicator of complement activation produced by these complexes. Complement activation by these reagents occurred at concentration ratios of protamine and heparin at which protamine neutralized the anticoagulant effect of heparin. Heparin-protamine complexes appeared to bind to red cells and produce complement activation by the classic pathway. C5a generation with heparin-protamine complexes in serum was greatly enhanced in the presence of red cells and increased with increasing red cell concentration. This enhancement of complement activation in the presence of red cells was also seen as measured by depletion of available C3 hemolytic complement units in the fluid phase. Thus red cells seem to play an important role in activation of complement by heparin-protamine complexes.     

Preliminary report on complement activating potential of polycarbonate membrane

Clinical as well as laboratory studies have been employed to assess the complement activating potential of polycarbonate membrane hemodialyzers. Blood samples from a group of patients undergoing sequential maintenance hemodialysis with cuprophane, polyacrylonitrile and polycarbonate devices were evaluated to define plasma levels of C3a antigen and leukocyte counts during the initial phases of hemodialysis. While polyacrylonitrile dialyzers did not activate complement to a significant extent, we did observe transient elevations in the plasma concentration of C3a and corresponding diminutions in the granulocyte counts of patients dialyzed with both cuprophane and polycarbonate dialyzers. However, polycarbonate devices appeared to activate complement to a lesser degree than cellulosic dialyzers. Laboratory evaluation of these three different types of dialyzers also provided evidence that polycarbonate membranes did not appear to activate human complement as readily as cuprophane. These observations suggest that polycarbonate membranes display complement-related biocompatibility properties that are intermediate between those of cuprophane and polyacrylonitrile.     

Biocompatibility of hemodialysis membranes: interrelations between plasma complement and cytokine levels

Hemodialysis (HD) membrane biocompatibility is defined as absence of complement activation. We have recently shown that circulating levels of interleukin (IL) 1 and IL-2 predict death and survival, respectively, of HD patients. Studies have assessed IL-1 in treatments with biocompatible and less biocompatible dialysis membranes, but no study has correlated circulating levels of all these immunoreactants. We assessed these immunoreactants, and temperature as an outcome, during HD in patients treated with different membranes. Twelve stable patients, receiving thrice-weekly chronic bicarbonate HD, were randomly dialyzed with three different types of membranes, composed of: Cuprophan, cuprammonium rayon modified cellulose, and Hemophan. Blood was drawn from the arterial line port before (Pre) and 15, 30, and 60 min during and after (Post) HD. Patients' temperatures were measured before and after each treatment. The plasma concentrations of IL-1 and IL-2 and factors C3a and C5a were assessed by ELISA. There were no differences between baseline levels of any of the immunoreactants in patients treated with different dialyzers. C3a, C5a, and IL-1 levels increased significantly during HD treatments with all three different membranes. C3a, C5a, and IL-1 levels during Cuprophan and Hemophan treatments were significantly higher than the levels during modified cellulose treatment at 30 and 60 min and Post (p < 0.01). For all the immunoreactants, however, the Post levels were higher than the Pre levels. In contrast to IL-1, there were no differences in mean IL-2 levels during treatments when different membranes were compared. There were few correlations of plasma C3a and C5a levels with plasma IL-1 levels, but there was only one treatment time in one dialyzer group during which IL-2 and any of the other factors were correlated. Pre and Post temperature values and percent change in temperature were not correlated with any of the immunoreactants measured. These data show that C3a, C5a, and IL-1 responses are similar, but not identical, during treatments with different membranes. The response of circulating IL-2 levels to treatments is quite different from that of plasma C3a, C5a and IL-1 levels and suggests that these changes are not solely due to treatment factors. Treatment with modified cellulose membranes is associated with a different immunoreactive profile as compared with patients dialyzed using other cellulose membranes. We suggest that circulating IL-1 levels are good biocompatibility markers.     

Animal models to study the cardiopulmonary effects of artificial kidney membranes

Some adverse reactions occurring during hemodialysis have been suggested to be caused by the bioincompatibility of dialyzer membranes. Experimental sheep and swine models have been developed to study the biological effects of these membranes. Infusion of cuprophane-activated plasma into the animals produces drastic hematologic and cardiopulmonary alterations. Available data suggest that these alterations are mediated by complement activation products and arachidonic acid metabolites. The consistent and sensitive nature of these models makes them useful tools to further study the interactions between the blood elements and artificial kidney membrane surfaces.     

Effects of complement activation in the isolated heart. Role of the terminal complement components.

The mechanisms of the complement-mediated myocardial injury associated with ischemia and reperfusion have not been elucidated fully. Complement activation may directly mediate injury through actions of the anaphylatoxins C3a and C5a or generation of the membrane attack complex C5b-9. A model was developed to examine the direct effects of complement activation on heart function, assess myocardial tissue damage, and determine which complement components mediate tissue injury. Isolated rabbit hearts were perfused with Krebs-Henseleit buffer by using a modified Langendorff apparatus. Human plasma was added to the perfusate as a source of complement. Rabbit tissue activates human complement. Treatment with 6% normal plasma resulted in complement activation as assessed by the generation of Bb, C3a, C5a, and SC5b-9. Functional changes in cardiac performance became apparent 7-15 minutes after plasma addition and developed fully over the next 20-30 minutes. The effects were dependent on the complement titer and included 1) an increase in the end-diastolic pressure, 2) a decrease in the developed pressure, 3) an increase in the coronary perfusion pressure, and 4) an increase in lymphatic fluid formation. These effects were not elicited when an inhibitor of complement activation (FUT-175) was present or when heat-inactivated plasma was used. The effects of complement activation on myocardial function could not be reproduced by treatment with recombinant human C5a, zymosan-activated plasma, or plasma selectively depleted of C8. Myocardial tissue accumulated sodium and calcium and lost potassium as a result of complement activation. Activation caused the release of creatine kinase from myocytes and an increase in the radiolabeled albumin space of the hearts. The data demonstrate that complement activation caused decrements in myocardial function and increased the coronary perfusion pressure and lymphatic fluid flow rate. The effects were not mediated by the anaphylatoxins but were dependent on the distal complement component C8, suggesting that C5b-9 was responsible for the physiological changes. Complement activation directly mediated tissue injury in a manner consistent with plasmalemmal disruption as a result of C5b-9 formation. The data suggest that the C5b-9 complex, which is known to form under conditions of ischemia, may contribute directly to myocardial cell injury.     

Analysis of density changes and chemotactic receptors of leukocytes from chronic hemodialysis and peritoneal dialysis patients

Analysis of standard Ficoll-Hypaque (density = 1.077 g/ml) separation profiles of peripheral white blood cells (WBC) from patients undergoing hemodialysis (HD) demonstrated that dialysis caused a marked decrease in the density of polymorphonuclear leukocytes (PMN) resulting in about 50% of these cells separating with the mononuclear cells. In vitro exposure of normal control peripheral blood to HD membranes as well as to the purified chemotactic factors C5a, C5ades-Arg, and formyl-Met-Leu-Phe (fMLP) also resulted in PMN density changes which altered the Ficoll-Hypaque separation profiles of WBC. Therefore, these results imply that C5a generation, resulting from complement activation by the HD membrane, induced the density changes in the PMN from HD patients. Further studies using flow cytometry and fluorescein-labeled chemotactic factors (C5a, formyl-Met-Leu-Phe-Lys [fMLPL] and casein) indicated that HD patients had a significant reduction in the ability of their PMN and monocytes to bind C5a. This contrasted with the findings of no significant difference in the percentage or fluorescence intensity of HD patients' PMN or monocytes binding casein or fMLPL. Functional studies to analyze chemotactic-factor-mediated responses indicated that there was a decreased ability of HD patients' PMN and monocytes to generate superoxide anion, produce H2O2 and release myeloperoxidase in response to both C5a and fMLP. Additional studies evaluated the binding of chemotactic factors to PMN and monocytes from normal blood following passage through a hemodialyzer and from patients undergoing HD. Analysis of receptor binding by control cells passed through the dialyzer showed that there was a progressive decrease in the percentage of C5a-receptor-positive PMN and monocytes but no change with casein or fMLPL. In contrast, peripheral PMN and monocytes from chronic renal failure patients on HD showed no difference in C5a, casein or fMLPL receptors during the course of HD as compared to the predialysis period. This appears to be attributable to a difference in the regulation of the C5a that is generated as a result of the dialysis-membrane-induced activation of the complement system. Although C5a has been shown to be continuously generated during the course of HD, these patients show no modulation of their C5a receptors during the course of HD or when their whole blood is exposed to dialysis membrane fibers. These findings suggest that there are mechanisms functioning in chronically dialyzed patients to protect them from the effects of excessive C5a generation during HD.(ABSTRACT TRUNCATED AT 400 WORDS)     

Accentuated formation of the terminal C5b-9 complement complex in patient plasma precedes development of the adult respiratory distress syndrome

Extensive studies have been conducted to determine the pathogenesis of the adult respiratory distress syndrome (ARDS) by investigating the role of complement, a mediator of inflammation. Complement activation products have been detected in blood samples from patients during ARDS. However, the individual complement components that have been assessed only indicated generalized inflammation, and none could unequivocally discriminate the onset of this acute inflammatory lung injury. In this two-year prospective study of 87 septic patients, 22 of whom developed ARDS (25%), we determined complement activation by quantifying the terminal complement complex (TCC), C5b-9. The TCC is a stable complement by-product formed following activation of either the classical or alternative pathways. Our results show that plasma TCC concentrations increased an average of 110% (p = 0.002) two days prior to the onset of ARDS and also transiently increased an average of 45% (p = 0.01) immediately preceding its resolution. Furthermore, plasma TCC concentrations were a more sensitive measure of this acute inflammatory lung injury than levels of C3a desarginine, C4a desarginine, C5a desarginine, and total hemolytic complement activity. We conclude that a temporal association exists between accentuated formation of plasma TCC and the development and also resolution of septic ARDS. Therefore, we suggest that researchers include plasma TCC concentrations in clinical studies when they could use a potential early indicator for ARDS.     

The complement system in ischemic heart disease

The mechanisms by which tissue injury after acute myocardial infarction (AMI) occurs has not been fully elucidated. Recent evidence in experimental models has suggested involvement of the complement system in microvascular and macrovascular injury subsequent to AMI. With respect to angina pectoris, whether or not the complement system is activated is not clear. The present study assessed the role of complement as a mediator of myocardial inflammation by quantifying products of complement activation, including C3d, C4d, Bb, and SC5b-9 complexes, in 31 patients with AMI, 17 patients with unstable angina pectoris, 19 patients with stable angina pectoris, and 20 normal volunteers. The plasma C3d levels increased in patients with AMI and in those with unstable angina pectoris (p less than 0.01). The plasma levels of C4d, Bb, and SC5b-9 increased only in patients with AMI (p less than 0.01). The plasma SC5b-9 level was related to peak creatine phosphokinase (r = 0.71) and inversely related to the ejection fraction (r = -0.71). The plasma SC5b-9 level of patients with congestive heart failure was higher than that of patients without congestive heart failure in AMI. These results show that activation of complement system occurs after AMI and show an association of myocardial damage with complement activation. With respect to angina pectoris, the complement system is mildly activated in patients with unstable angina pectoris; however, the cardiac function of patients with unstable angina pectoris is not damaged. The complement system of patients with stable angina pectoris is not activated.     

The complement system in the acute phase of myocardial infarction

Although some studies suggest that complement activation is involved in the development of acute myocardial infarction, there has been little convincing evidence of a change in the complement system in patients suffering from myocardial infarction. In this study circulating levels of C3a, C3, C4 and the total hemolytic complement titer (CH50) were serially measured in 12 patients with acute myocardial infarction up to 10 days after an attack. The plasma C3a level was greatly elevated throughout the post-attack observation period. The C3, C4 and CH50 levels were significantly increased above those controls on days 8, 9 and 10 after infarction. These findings indicate that there is complement activation in patients with acute myocardial infarction, and suggest a pathogenetic role for complement activation in the development of myocardial damage after infarction.     

Haemodialysis-induced activation of complement. Effects of different membranes

The ability of cellulose-based (Cuprophan, saponified cellulose ester) and synthetic (polyacrylonitrile, polycarbonate, polymethylmethacrylate) haemodialysis membranes to activate complement during treatment was compared, using functional, immunochemical, radioimmunoassays, and fast centrifugal analysis assay techniques. Cellulosic and synthetic membranes show a striking similarity in their complement activation when measured by immunochemical assays. Functional haemolytic assays for alternate pathway and CH50 demonstrate no significant differences from predialysis values. C3d levels were also unable to demonstrate differences between the membranes. C3a levels, on the other hand, demonstrated significant differences between cuprophan, polyacrylonitrile, and polymethylmethacrylate but not between Cuprophan and polycarbonate membranes. Since comparable alternate-pathway activity of both cellulosic and synthetic membranes was demonstrated, but their C3a release differed, it is possible that certain surfaces that activate complement also possess the ability to absorb components of the alternate pathway.     

Leukocytes, eosinophils and complement function during hemodialysis with polysulphone and polymethylmethacrylate membranes: comparison with cuprophan and polyacrylonitrile

The biocompatibility of the two new dialysis membranes, polysulphone (PS) and polymethylmethacrylate (PMMA), was evaluated versus cuprophan (CUP) and polyacrylonitrile (PAN) by studying the in vivo effects of the four different membranes on leukocyte counts, eosinophil levels and complement function both in the presence and absence of dialysis fluid. Complement function was also examined in vitro by studying the generation of chemotactic factors, whole complement activity and C3d serum conversion. Passive absorption of complement fractions by membranes has completed in vitro studies. PS, PMMA and PAN showed a higher biocompatibility than CUP, even if slight differences can be observed: PS showed a PAN-like biocompatibility pattern with a relatively high absorption of complement factors by the membrane and without complement activation. On the other hand, PMMA showed a CUP-like pattern and caused complement activation, even though to a lower intensity than CUP. PMMA biocompatibility appears to stand in-between CUP and the other two synthetic membranes PS and PAN. Our results confirm the important role played by membrane-induced complement activation on hemodialysis leukopenia. Dialysis fluid does not have a significant influence on membrane biocompatibility, but represents the major factor in determining intradialytic eosinopenia. Eosinophils seem to represent a more important marker of dialysis than of membrane biocompatibility.     

Failure to demonstrate complement activation during bronchial challenge test

In order to demonstrate a possible complement activation during early bronchospastic reaction in asthma, we have measured plasmatic C3d (a split product of C3) and the C3d/C3 index, both of which are sensitive indices of complement activation. Twenty-nine allergenic bronchial challenge tests were accomplished, with an absence of response in six cases, an early reaction in sixteen cases and a dual reaction in seven cases. Changes in plasmatic C3d or C3d/C3 five min after an early reaction, or five min after the last dose of allergen (in the six cases without bronchial response) were insignificant. However, complement activation in the lungs during asthmatic reaction cannot be completely excluded without studies using the bronchoalveolar technique.     

Activation of the classical complement pathway in spontaneous bacterial peritonitis.

To investigate the possibility that low complement concentrations in the plasma and ascites of patients with severe liver disease could be secondary to complement consumption, complement activation was studied in 32 patients with severe liver disease, 11 of whom had spontaneous bacterial peritonitis (SBP). In patients with SBP, plasma C3 and C4 were significantly lower than in uninfected patients (mean values 0.74 v 1.13 g/l, p less than 0.01 and 0.20 v 0.28 g/l, p less than 0.05 respectively). Plasma complement activation via the classical pathway, as shown by C4d/C4, was significantly increased in patients with SBP compared with uninfected patients (37.3 v 22.2, p less than 0.01) as was C3d/C3 (14.0 v 8.11, p less than 0.01), but there was no significant difference in Ba/B between SBP and uninfected patients. Ascitic C3 concentrations were higher in patients without SBP than in infected patients (0.37 v 0.08 g/l, p less than 0.05), as were factor B values (0.11 v 0.03 g/l, p less than 0.05). There was no significant difference in ascitic C4 concentrations in patients with SBP compared with uninfected patients (0.03 v 0.07 g/l). Although consumption of C3, as shown by C3d/C3 in ascites, was increased in infected patients compared with uninfected patients (79.1 v 36.1, p less than 0.05), there was no difference in ascitic complement activation between the groups for either the classical or alternative pathways. In SBP, decreased plasma C3 and C4 are primarily caused by increased activation of the classical pathway and not impaired hepatic synthesis. Activation and consumption of C3 is one factor causing the low ascitic C3 concentrations observed in SBP.     

Acute hemorrhagic pancreatic necrosis in mice: lack of a pathogenetic role for complement.

Feeding a choline-deficient diet containing 0.5% DL-ethionine induces an acute hemorrhagic pancreatitis in 100% of young female mice. Evidence for deposition of the third component of complement (C3) on acinar cell plasma membranes was sought, during the inductive stages, by a sandwich immunofluorescence technique. Such a localization could not be demonstrated even though the method is capable of detecting less than 8 x 10(-5) microgram of protein/mm2 of cell membrane. Artifactual binding of immunoglobulin reagents was encountered when goat antisera, with high levels of circulating immune complexes, formed the middle layer in the sandwich technique. This was attributed to the appearance of Fc receptors on the plasma membrane of degenerating acinar cells, and could be avoided by ultracentrifuging acinar cells, and could be avoided by ultracentrifuging the goat antisera prior to sue. In view of the fact that C3 cleavage represents an amplification loop in both the calssical and alternate pathways of complement activation, the lack of demonstrable C3 staining in tbe present experiments strongly suggests that complement plays no role in acinar cell necrosis in this model of pancreatitis.     

The cytolytic C5b-9 complement complex: feedback inhibition of complement activation

We describe a regulatory function of the terminal cytolytic C5b-9 complex [C5b-9(m)] of human complement. Purified C5b-9(m) complexes isolated from target membranes, whether in solution or bound to liposomes, inhibited lysis of sensitized sheep erythrocytes by whole human serum in a dose-dependent manner. C9 was not required for the inhibitory function since C5b-7 and C5b-8 complexes isolated from membranes were also effective. No effect was found with the cytolytically inactive, fluid-phase SC5b-9 complex. However, tryptic modification of SC5b-9 conferred an inhibitory capacity to the complex, due probably to partial removal of the S protein. Experiments using purified components demonstrated that C5b-9(m) exerts a regulatory effect on the formation of the classical- and alternative-pathway C3 convertases and on the utilization of C5 by cell-bound C5 convertases. C5b-9(m) complexes were unable to inhibit the lysis of cells bearing C5b-7(m) by C8 and C9. Addition of C5b-9(m) to whole human serum abolished its bactericidal effect on the serum-sensitive Escherichia coli K-12 strain W 3110 and suppressed its hemolytic function on antibody-sensitized, autologous erythrocytes. Feedback inhibition by C5b-9(m) represents a biologically relevant mechanism through which complement may autoregulate its effector functions.     

Cardiovascular function and arterial blood gases during sham dialysis in healthy man

Cardiovascular function and alveolar gas exchange were studied in healthy subjects undergoing sham dialysis (SHD)--i.e. the circulation of blood through a cuprophane dialyzer with the dialysate compartment closed to avoid diffusion and convective transport of fluid and solutes. The blood-membrane contact induced complement activation (rise in C3d) and transient leukopenia, as described during clinical hemodialysis. PaO2, PaCO2 and calculated oxygen uptake remained unchanged. Heart rate, cardiac index (thermodilution), systemic vascular resistance index and brachial and pulmonary arterial blood pressures did not change significantly during 150 min of SHD (n = 8). In 12 subjects, in whom more frequent measurements were made during the first 30 min of SHD, pulmonary arterial systolic and diastolic blood pressures decreased significantly while the dialyzer and the tubing set filled with blood, and pulmonary arterial mean blood pressure did not change significantly. Pulmonary capillary wedge pressure fell during the filling phase, but did not change significantly during SHD; pulmonary vascular resistance index remained unchanged. We conclude that in nonuremic subjects sham dialysis with a cuprophane dialyzer does not result in hypoxemia, pulmonary vascular constriction and pulmonary hypertension, in spite of complement activation and marked leukopenia.     

Accentuated complement activation in patient plasma during the adult respiratory distress syndrome: a potential mechanism for pulmonary inflammation

The adult respiratory distress syndrome (ARDS) represents an acute inflammatory lung disorder, characterized by both refractory hypoxemia and a mortality rate approaching 95%. Researchers have proposed that activation of the complement (C) system may play a role in the development of the pulmonary inflammation associated with ARDS. The purpose of this investigation was to determine whether complement activation occurs to a greater extent in patients with ARDS than in patients without ARDS, thereby providing a potential mechanism for the acute inflammation seen in ARDS. In this study, we assessed plasma complement activation by measuring complement activation by-products: C1rC1s-C1 inhibitor complex and C3b-P complex, generated subsequent to activation of the classical and alternative pathways, respectively, and also the terminal complement complex, formed after activation of either the classical or alternative complement pathway. Multivariate discriminant analysis revealed that these three complement activation complexes could distinguish patients with ARDS from those without ARDS (p less than 0.04). Furthermore, these three complexes provided a more sensitive discriminator of ARDS than did plasma levels of C3a desarginine (p greater than 0.30), C5a desarginine (p greater than 0.41) and total hemolytic complement activity (CH50) (p greater than 0.72). We conclude that a temporal association exists between the complement activation and the development of ARDS. Therefore, we suggest that complement activation by-products be included in the armamentarium for ARDS.