Complement Activation and Polymorphonuclear Neutrophil Leukocyte Elastase in Sepsis Correlation With Marco
Gardinali, MD; Pietro Padalino, MD; Sergio Vesconi, MD; Anna Calcagno, MD; Salvatore Ciappellano; Luisa Conciato, MD; Osvaldo Chiara, MD; Angelo Agostoni, MD; Angelo Nespoli, MD
Complement activation is necessary for an adequate imand inflammatory response to infections. Activation releases anaphylatoxins that cause vasodilation, increase vascular permeability, and trigger release of polymorphonuclear neutrophil leukocyte (PMN) lysosomal enzyme and oxygen radicals. Under normal circumstances, an orderly progression of such events has a beneficial antimicrobial \s=b\
mune
effect. The same mechanism, however, when uncontrolled, may damage host tissues. To provide information about the clinical importance of such events in sepsis, different complement parameters (C3, C4, and the desarginated forms of C3a [C3ades-Arg] and C5a [C5ades-Arg]), PMN elastase, and malondialdehyde (a by-product of membrane peroxidation by oxygen radicals) were measured daily in 26 septic patients and correlated with two objectively assessed and previously validated severity scores (acute physiology and chronic health evaluation [APACHE II] and Sepsis Severity Score [SSS]). Nonsurvivors (n=12) had significantly greater and longer lasting complement activation than that in survivors, as reflected by higher levels of catabolic peptides (C3ades-Arg) and lower levels of native proteins (C3 and C4). C3ades-Arg, C3, C4, and the C3ades-Arg-C3 ratio were correlated with Sepsis Severity Scores. Polymorphonuclear neutrophil leukocyte elastase levels were higher in nonsurvivors and were correlated with C3ades-Arg and the C3ades-Arg-C3 ratio. Malondialdehyde levels were significantly higher in all patients than in controls, without, however, any relationship to severity of disease or clinical outcome. Since the higher and more persistent the complement activation and polymorphonuclear neutrophil leukocyte stimulation, the worse the patient's prognosis, we conclude that these mechanisms may be important in the clinical development of sepsis.
(Arch Surg. 1992;127:1219-1224) is
serious
that still bears a high for septic patients
clinical condition Sepsi s rate.1 mortality prognosis has a
Severity of Disease
Since the
been only partially improved by common therapeutic medical and surgical procedures,12 better comprehension of bacterial infection physiopathology is needed. According to a variety of observations from different groups, an uncon¬ trolled inflammatory response evoked by bacteria would be
Accepted for publication
November 9, 1991. From the Institute of Emergency Surgery (Drs Padalino, Chiara, and Nespoli), the Clinical Medicine (Drs Gardinali, Calcagno, Conciato, and Agostoni), Intensive Care Unit San Paolo Hospital (Dr Vesconi), and the Department of Food Science and Microbiology (Mr Ciappellano), University of Milan (Italy). Reprint requests to Istituto de Chirurgia d'Urgenza, Universit\l=a`\di Milano, Ospedale Policlinico IRCCS, via F Sforza 35-20122 Milano, Italia (Dr Padalino).
ultimately responsible for a patient's death.3,4 Bacteria and lipopolysaccharide (LPS) or other constituents of gramnegative bacteria walls activate the complement system in both antibody-dependent and independent ways.5,6 Complement activation is fundamental for bacterial opsonization since C3 fragments bind to bacterial walls and interact with specific complement receptors on phagocytic cells.7 Bacterial lysis is then completed by insertion of C5b-9 complexes into the cell membrane.8 Moreover, C3a
and C5a are vasodilators, and C5a is also a potent chemoattractant that recruits inflammatory cells to the site of infection. C5a elicits release of polymorphonuclear neu¬ trophil leukocyte (PMN) lysosomal enzymes and oxygen radicals, which contribute to bacterial lysis.9,10 However, in particular conditions, host cell membranes may become the target of these microbicidal mechanisms, leading to failure of different organs. Moreover, anaphylatoxins, particularly in the presence of LPS, cause production of such cytokines as interleukin 1 and tumor necrosis factor, which can be additionally detrimental to the host.11,12 Im¬ pairment of PMN functions may also follow anaphylatoxin binding to cell receptors.13,14 Recent experiments support the idea that complement activation is responsible for hemodynamic derangements and contributes to mortality in animal models of sepsis.1517 Likewise, various clinical studies have already demonstrated that complement activation occurs in septic
patients.1825
The aim of this study was to evaluate whether the extent of complement activation reflects the severity of sepsis. For this purpose, the levels of different complement parame¬ ters were measured. Both native protein C3 and C4 and catabolic peptide C3ades.Arg and C5ades.Arg levels were eval¬ uated. If complement activation occurs, both consumption of the native proteins and generation of catabolic peptides would be expected. The severity of the clinical conditions was assessed by calculation of two previously validated scores that grade the severity of the illness and predict
mortality.1,26,27 Polymorphonuclear neutrophil leukocyte lysosomal enzyme release was evaluated by measuring the complexes between elastase and arproteinase inhibitor,
which binds up to 90% of plasma elastase. The levels of these complexes (PMN elastase-ai-proteinase inhibitor) are considered to be suitable markers of PMN degranulation.28,29 Polymorphonuclear neutrophil leukocyte elastase has a proteolytic effect on connective tissue structures, as well as on plasma proteins. The levels of malondialdehyde (MDA), which is generated by lipid peroxidation, were measured as a marker of cell membrane damage by PMN oxygen radicals.30
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/31/2015
Table 1.—Case List* Infectious
Patient/Age, y/Sex
Outcome
Disease
From
Agents Isolated
Biological
Fluids
1/3 6/M
S
Subphrenic abscess
Escherichia coli-Klebsiella
2/43/M
D
IPN
E coli-Pseudomonas
3/3 8/M 4/43/F
D
IPN
E coli-P aeruginosa
S
Peritonitis
E
5/4 7/M
S
IPN
6/54/M 7/2 6/M
D
Peritonitis
Enterobacter cloacae-P aeruginosa
S
IPN
E
8/69/F
D
IPN
E coli-Proteus mirabilis
9/19/M
D
Liver trauma
E coli-P mirabilis
10/53/F
D
Intestinal infarction
11/52/F
S
Peritonitis
12/45/M 13/59/M
D
Intestinal infarction
mirabilis-S aureus~P E coli
S
Peritonitis
S aureus-P
14/65/M
S
Peritonitis
S aureus-P
15/55/F 16/52/M
s
Peritonitis
E coli-P
D
Endocarditis
1 7/38/M
S
IPN
18/54/M
S
Unknown
19/62/M
D
S
pneumoniae aeruginosa
coli-Streptococcus faecalis Staphylococcus epidermidis coli-Staphylococcus aureus aeruginosa-P mirabilis-Candida albicans
liquefaciens-C albicans
aeruginosa-Achromobacter xyloxidans
aeruginosa aeruginosa-K pneumoniae S fecalis-S aureus-P aeruginosa S fecalis-S aureus-P aeruginosa E coli-S
aureus
20/61/M
S
Pulmonary Pulmonary abscess
21/70/M
D
Peritonitis
E coli
22/70/M
D
Intestinal infarction
C albicans
23/51/F 24/50/M
S
Peritonitis
E cloacae-S aureus-E coli-S faecalis
D
Pulmonary abscess
S
25/68/M
S
Peritonitis
Serraba marcescens-P aeruginosa-E cloacae
S
Peritonitis
E cloacae-S marcescens-S
26/58/M *S indicates
aeruginosa-S aureus-C albicans
abscess
aureus
aureus
aureus
survived; D, dead; and IPN, infected pancreatic necrosis. PATIENTS
Twenty-six patients (age range, 19 to 70 years; 20 men and six women) who were admitted to the intensive care units (ICUs) of the Institute of Emergency Surgery and San Paolo Hospital of Milan (Italy) University entered the study. Four of six of the following sepsis criteria were needed for in¬ clusion in the study: (1) temperature of 39°C or greater; (2) chills; (3) tachycardia (heart rate, >100 beats per minute); (4) tachypnea (respirations, >30/min); (5) mental impairment; and (6) leukocytosis (white blood cell count, >10xl07L). Bacterial infection was documented in all of the patients by bac¬ teriological studies. Patients with neoplasms, chronic liver or kid¬ ney diseases, or diabetes were excluded, because these diseases may have affected the complement system.m No patient was treated with corticosteroids, since these drugs may have affected both complement activation and PMN stimulation by anaphylatoxins.12 Patients with infected pancreatic necrosis were admitted to the study only when they fulfilled the above-mentioned sepsis criteria (usually 1 to 2 weeks after onset of pancreatitis). In all of them, laparotomy proved the presence of infective foci. Table 1 re¬ ports age, sex, site of infection, and outcome of the patients. Pa¬ tients were followed up from admission to the ICU until their dis¬ charge (corresponding to the end of hemodynamic monitoring by removal of a Swan-Ganz catheter) or death. Patients were monitored for a mean±SD of 8.5±9.8 days. The pe¬ riod of observation was slightly longer for survivors than for non¬ survivors (9.7±11.0 days vs 6.9±7.8 days, P=not significant). The Sepsis Severity Score (SSS) was calculated according to the method of Stevens.2* Briefly, dysfunction in seven organs or systems (lung,
Wdney, cardiovascular, coagulation, liver, gastrointestinal, and cen¬ tral nervous) was graded from 1 to 5 according to its severity. The score was calculated by adding the values (squared) assigned to each of three organs or systems with the most severe dysfunction. Thus, the maximum possible rating was 75 (52X3). The acute physiology and chronic health evaluation (APACHE) II score was calculated according to the method of Knaus et al.27 Eleven physiological variables were graded 1 to 4, and the assigned values were added. Additional points were given according to the patient's Glasgow coma scale, age, and previous health status, to a theoretical maximum rating of 71. Both scores were calculated daily. MATERIALS AND METHODS Blood samples were drawn daily, early in the morning, from an antecubital vein. Blood was then centrifuged (10 minutes at 2500# at 4°C). Plasma (or serum) samples were stored at -70°C until they were tested.
Complement System C3 and C4 levels were measured (in serum) by radial immunodiffusion with the use of plates (Nor Partigen, Behring Institute, Scoppito, Italy). The normal values for our laboratory (mean ± SD) were as follows: C3, 0.87±0.13 g/L; C4, 0.36±0.11 g/L. C3ades.Arg and C5adcs.ArR values were measured by radioimmunoassay according to the method of Hugh and Chenoweth11 with the use of radioimmunoassay kits (Amersham International, Amersham, United Kingdom). Values were calculated from standard curves in which the correlation coefficient was greater than .98. Samples for anaphylatoxin measurement were collected in tubes contain-
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/31/2015
ing Na2-edetic acid (EDTA) (7.5 µg/mL of blood). Normal C3aJc.sAr8 values (mean±SD) were equal to 219±79 ng/mL. C5ades.Alg levels in healthy subjects were constantly less than 10 ng/mL (ie, the lower limit of detection of the assay). The
C3adM_Arg-C3
ratio
was
calculated to evaluate the extent
Table
Distribution of Scores Clinical Outcome*
2.—Frequency
by
(in
percent) of native protein conversion according to the following equation: (C3adcs.Arg/C3)x (Molecular Weight of C3/Molecular Weight of C3ados.Arg). Molecular weights of 9 kd (C3a) and 185 kd (C3) were used. Normal values (mean±SD) were equal to 0.7%±0.3%. PMN
Elastase-a,-Proteinase
Inhibitor
Complexes
R inhibitor
Serum PMN elastase-a,-proteinase complexes were measured by an immunoenzymatic method, according to the method of Neumann et al28 based on the experiments of Hoshino and colleagues,34 with the use of a PMN elastase immunoassay (Merck, Darmstadt, Federal Republic of Germany). Normal val¬ ues
(mean±SD)
were
equal to
105±22
µg/L.
Malondialdehyde
vivors).
Analyses
Mean values of different parameters for each patient were cal¬ culated and used for statistical analysis. Student's t test (or Mann-Whitney U test) was used to deter¬ mine the significance of differences between groups (ie, survivors vs nonsurvivors). Student's t test for paired data (or Wilcoxon Matched-Pairs Signed-Rank test) analyzed differences between parameters in the first and last samples that were collected from each patient. Correlations between two biochemical parameters or between biochemical parameters and scores were determined by the Spearman Rank Correlation Coefficient test. P
Gardinali, MD; Pietro Padalino, MD; Sergio Vesconi, MD; Anna Calcagno, MD; Salvatore Ciappellano; Luisa Conciato, MD; Osvaldo Chiara, MD; Angelo Agostoni, MD; Angelo Nespoli, MD
Complement activation is necessary for an adequate imand inflammatory response to infections. Activation releases anaphylatoxins that cause vasodilation, increase vascular permeability, and trigger release of polymorphonuclear neutrophil leukocyte (PMN) lysosomal enzyme and oxygen radicals. Under normal circumstances, an orderly progression of such events has a beneficial antimicrobial \s=b\
mune
effect. The same mechanism, however, when uncontrolled, may damage host tissues. To provide information about the clinical importance of such events in sepsis, different complement parameters (C3, C4, and the desarginated forms of C3a [C3ades-Arg] and C5a [C5ades-Arg]), PMN elastase, and malondialdehyde (a by-product of membrane peroxidation by oxygen radicals) were measured daily in 26 septic patients and correlated with two objectively assessed and previously validated severity scores (acute physiology and chronic health evaluation [APACHE II] and Sepsis Severity Score [SSS]). Nonsurvivors (n=12) had significantly greater and longer lasting complement activation than that in survivors, as reflected by higher levels of catabolic peptides (C3ades-Arg) and lower levels of native proteins (C3 and C4). C3ades-Arg, C3, C4, and the C3ades-Arg-C3 ratio were correlated with Sepsis Severity Scores. Polymorphonuclear neutrophil leukocyte elastase levels were higher in nonsurvivors and were correlated with C3ades-Arg and the C3ades-Arg-C3 ratio. Malondialdehyde levels were significantly higher in all patients than in controls, without, however, any relationship to severity of disease or clinical outcome. Since the higher and more persistent the complement activation and polymorphonuclear neutrophil leukocyte stimulation, the worse the patient's prognosis, we conclude that these mechanisms may be important in the clinical development of sepsis.
(Arch Surg. 1992;127:1219-1224) is
serious
that still bears a high for septic patients
clinical condition Sepsi s rate.1 mortality prognosis has a
Severity of Disease
Since the
been only partially improved by common therapeutic medical and surgical procedures,12 better comprehension of bacterial infection physiopathology is needed. According to a variety of observations from different groups, an uncon¬ trolled inflammatory response evoked by bacteria would be
Accepted for publication
November 9, 1991. From the Institute of Emergency Surgery (Drs Padalino, Chiara, and Nespoli), the Clinical Medicine (Drs Gardinali, Calcagno, Conciato, and Agostoni), Intensive Care Unit San Paolo Hospital (Dr Vesconi), and the Department of Food Science and Microbiology (Mr Ciappellano), University of Milan (Italy). Reprint requests to Istituto de Chirurgia d'Urgenza, Universit\l=a`\di Milano, Ospedale Policlinico IRCCS, via F Sforza 35-20122 Milano, Italia (Dr Padalino).
ultimately responsible for a patient's death.3,4 Bacteria and lipopolysaccharide (LPS) or other constituents of gramnegative bacteria walls activate the complement system in both antibody-dependent and independent ways.5,6 Complement activation is fundamental for bacterial opsonization since C3 fragments bind to bacterial walls and interact with specific complement receptors on phagocytic cells.7 Bacterial lysis is then completed by insertion of C5b-9 complexes into the cell membrane.8 Moreover, C3a
and C5a are vasodilators, and C5a is also a potent chemoattractant that recruits inflammatory cells to the site of infection. C5a elicits release of polymorphonuclear neu¬ trophil leukocyte (PMN) lysosomal enzymes and oxygen radicals, which contribute to bacterial lysis.9,10 However, in particular conditions, host cell membranes may become the target of these microbicidal mechanisms, leading to failure of different organs. Moreover, anaphylatoxins, particularly in the presence of LPS, cause production of such cytokines as interleukin 1 and tumor necrosis factor, which can be additionally detrimental to the host.11,12 Im¬ pairment of PMN functions may also follow anaphylatoxin binding to cell receptors.13,14 Recent experiments support the idea that complement activation is responsible for hemodynamic derangements and contributes to mortality in animal models of sepsis.1517 Likewise, various clinical studies have already demonstrated that complement activation occurs in septic
patients.1825
The aim of this study was to evaluate whether the extent of complement activation reflects the severity of sepsis. For this purpose, the levels of different complement parame¬ ters were measured. Both native protein C3 and C4 and catabolic peptide C3ades.Arg and C5ades.Arg levels were eval¬ uated. If complement activation occurs, both consumption of the native proteins and generation of catabolic peptides would be expected. The severity of the clinical conditions was assessed by calculation of two previously validated scores that grade the severity of the illness and predict
mortality.1,26,27 Polymorphonuclear neutrophil leukocyte lysosomal enzyme release was evaluated by measuring the complexes between elastase and arproteinase inhibitor,
which binds up to 90% of plasma elastase. The levels of these complexes (PMN elastase-ai-proteinase inhibitor) are considered to be suitable markers of PMN degranulation.28,29 Polymorphonuclear neutrophil leukocyte elastase has a proteolytic effect on connective tissue structures, as well as on plasma proteins. The levels of malondialdehyde (MDA), which is generated by lipid peroxidation, were measured as a marker of cell membrane damage by PMN oxygen radicals.30
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/31/2015
Table 1.—Case List* Infectious
Patient/Age, y/Sex
Outcome
Disease
From
Agents Isolated
Biological
Fluids
1/3 6/M
S
Subphrenic abscess
Escherichia coli-Klebsiella
2/43/M
D
IPN
E coli-Pseudomonas
3/3 8/M 4/43/F
D
IPN
E coli-P aeruginosa
S
Peritonitis
E
5/4 7/M
S
IPN
6/54/M 7/2 6/M
D
Peritonitis
Enterobacter cloacae-P aeruginosa
S
IPN
E
8/69/F
D
IPN
E coli-Proteus mirabilis
9/19/M
D
Liver trauma
E coli-P mirabilis
10/53/F
D
Intestinal infarction
11/52/F
S
Peritonitis
12/45/M 13/59/M
D
Intestinal infarction
mirabilis-S aureus~P E coli
S
Peritonitis
S aureus-P
14/65/M
S
Peritonitis
S aureus-P
15/55/F 16/52/M
s
Peritonitis
E coli-P
D
Endocarditis
1 7/38/M
S
IPN
18/54/M
S
Unknown
19/62/M
D
S
pneumoniae aeruginosa
coli-Streptococcus faecalis Staphylococcus epidermidis coli-Staphylococcus aureus aeruginosa-P mirabilis-Candida albicans
liquefaciens-C albicans
aeruginosa-Achromobacter xyloxidans
aeruginosa aeruginosa-K pneumoniae S fecalis-S aureus-P aeruginosa S fecalis-S aureus-P aeruginosa E coli-S
aureus
20/61/M
S
Pulmonary Pulmonary abscess
21/70/M
D
Peritonitis
E coli
22/70/M
D
Intestinal infarction
C albicans
23/51/F 24/50/M
S
Peritonitis
E cloacae-S aureus-E coli-S faecalis
D
Pulmonary abscess
S
25/68/M
S
Peritonitis
Serraba marcescens-P aeruginosa-E cloacae
S
Peritonitis
E cloacae-S marcescens-S
26/58/M *S indicates
aeruginosa-S aureus-C albicans
abscess
aureus
aureus
aureus
survived; D, dead; and IPN, infected pancreatic necrosis. PATIENTS
Twenty-six patients (age range, 19 to 70 years; 20 men and six women) who were admitted to the intensive care units (ICUs) of the Institute of Emergency Surgery and San Paolo Hospital of Milan (Italy) University entered the study. Four of six of the following sepsis criteria were needed for in¬ clusion in the study: (1) temperature of 39°C or greater; (2) chills; (3) tachycardia (heart rate, >100 beats per minute); (4) tachypnea (respirations, >30/min); (5) mental impairment; and (6) leukocytosis (white blood cell count, >10xl07L). Bacterial infection was documented in all of the patients by bac¬ teriological studies. Patients with neoplasms, chronic liver or kid¬ ney diseases, or diabetes were excluded, because these diseases may have affected the complement system.m No patient was treated with corticosteroids, since these drugs may have affected both complement activation and PMN stimulation by anaphylatoxins.12 Patients with infected pancreatic necrosis were admitted to the study only when they fulfilled the above-mentioned sepsis criteria (usually 1 to 2 weeks after onset of pancreatitis). In all of them, laparotomy proved the presence of infective foci. Table 1 re¬ ports age, sex, site of infection, and outcome of the patients. Pa¬ tients were followed up from admission to the ICU until their dis¬ charge (corresponding to the end of hemodynamic monitoring by removal of a Swan-Ganz catheter) or death. Patients were monitored for a mean±SD of 8.5±9.8 days. The pe¬ riod of observation was slightly longer for survivors than for non¬ survivors (9.7±11.0 days vs 6.9±7.8 days, P=not significant). The Sepsis Severity Score (SSS) was calculated according to the method of Stevens.2* Briefly, dysfunction in seven organs or systems (lung,
Wdney, cardiovascular, coagulation, liver, gastrointestinal, and cen¬ tral nervous) was graded from 1 to 5 according to its severity. The score was calculated by adding the values (squared) assigned to each of three organs or systems with the most severe dysfunction. Thus, the maximum possible rating was 75 (52X3). The acute physiology and chronic health evaluation (APACHE) II score was calculated according to the method of Knaus et al.27 Eleven physiological variables were graded 1 to 4, and the assigned values were added. Additional points were given according to the patient's Glasgow coma scale, age, and previous health status, to a theoretical maximum rating of 71. Both scores were calculated daily. MATERIALS AND METHODS Blood samples were drawn daily, early in the morning, from an antecubital vein. Blood was then centrifuged (10 minutes at 2500# at 4°C). Plasma (or serum) samples were stored at -70°C until they were tested.
Complement System C3 and C4 levels were measured (in serum) by radial immunodiffusion with the use of plates (Nor Partigen, Behring Institute, Scoppito, Italy). The normal values for our laboratory (mean ± SD) were as follows: C3, 0.87±0.13 g/L; C4, 0.36±0.11 g/L. C3ades.Arg and C5adcs.ArR values were measured by radioimmunoassay according to the method of Hugh and Chenoweth11 with the use of radioimmunoassay kits (Amersham International, Amersham, United Kingdom). Values were calculated from standard curves in which the correlation coefficient was greater than .98. Samples for anaphylatoxin measurement were collected in tubes contain-
Downloaded From: http://archsurg.jamanetwork.com/ by a University of Arizona Health Sciences Library User on 05/31/2015
ing Na2-edetic acid (EDTA) (7.5 µg/mL of blood). Normal C3aJc.sAr8 values (mean±SD) were equal to 219±79 ng/mL. C5ades.Alg levels in healthy subjects were constantly less than 10 ng/mL (ie, the lower limit of detection of the assay). The
C3adM_Arg-C3
ratio
was
calculated to evaluate the extent
Table
Distribution of Scores Clinical Outcome*
2.—Frequency
by
(in
percent) of native protein conversion according to the following equation: (C3adcs.Arg/C3)x (Molecular Weight of C3/Molecular Weight of C3ados.Arg). Molecular weights of 9 kd (C3a) and 185 kd (C3) were used. Normal values (mean±SD) were equal to 0.7%±0.3%. PMN
Elastase-a,-Proteinase
Inhibitor
Complexes
R inhibitor
Serum PMN elastase-a,-proteinase complexes were measured by an immunoenzymatic method, according to the method of Neumann et al28 based on the experiments of Hoshino and colleagues,34 with the use of a PMN elastase immunoassay (Merck, Darmstadt, Federal Republic of Germany). Normal val¬ ues
(mean±SD)
were
equal to
105±22
µg/L.
Malondialdehyde
vivors).
Analyses
Mean values of different parameters for each patient were cal¬ culated and used for statistical analysis. Student's t test (or Mann-Whitney U test) was used to deter¬ mine the significance of differences between groups (ie, survivors vs nonsurvivors). Student's t test for paired data (or Wilcoxon Matched-Pairs Signed-Rank test) analyzed differences between parameters in the first and last samples that were collected from each patient. Correlations between two biochemical parameters or between biochemical parameters and scores were determined by the Spearman Rank Correlation Coefficient test. P
