Serum Complement C3/C4 Ratio, a Novel Marker for Recurrent Cardiovascular Events Anil Palikhe, MDa,d, Juha Sinisalo, MD, PhDa, Mikko Seppänen, MDb, Heikki Haario, PhDe, Seppo Meri, MD, PhDc, Ville Valtonen, MD, PhDb, Markku S. Nieminen, MD, PhDa, and Marja-Liisa Lokki, PhDd,* Acute coronary syndrome is an inflammatory disease, during which the complement cascade is activated. We assessed the complement C3 and C4 concentration ratio (C3/C4 ratio) in serum as a potential measurement to predict cardiovascular attacks. Patients with acute coronary syndrome (n ⴝ 148) were followed after an initial attack for subsequent ischemic cardiovascular events (composite end point of death, myocardial infarction, recurrent unstable angina, or stroke). During the follow-up period (average 555 days), 44 patients met an end point. Blood samples were taken at hospitalization, 1 week, 3 months, and 1 year after hospital admission. Serum complement C3 and C4 concentrations and the C3/C4 ratio were analyzed. Patients with an end point had, throughout the follow-up period, a higher C3/C4 ratio than patients without these end points (repeated measures analysis of variance, p ⴝ 0.007). When all traditional cardiovascular risk factors and other potential confounding factors were included in a Cox multivariate logistic regression survival analysis, the C3/C4 ratio emerged as the novel risk factor for any new cardiovascular event (odds ratio 1.33, 95% confidence interval 1.08 to 1.63, p ⴝ 0.007). When the C3/C4 ratio was divided into 4 quartiles, 24% in quartiles 1 and 2 (lowest) and 48% in quartile 4 (highest) had end points during follow-up (odds ratio 3.04, 95% confidence interval 1.27 to 7.29, p ⴝ 0.01). In conclusion, increased serum C3/C4 ratio is a readily available and novel marker for recurrent cardiovascular events in acute coronary syndrome. The relative increase in serum C3 protein and decrease in C4 protein could explain changes in the C3/C4 ratio. © 2007 Elsevier Inc. All rights reserved. (Am J Cardiol 2007; 99:890 – 895)
Acute coronary syndrome is a state of persistent inflammation,1 during which the complement cascade is activated.2,3 The complement cascade activation modifies inflammation and can cause additional tissue injury. The complement system consists of 3 different activation pathways: classic, lectin, and alternative. All 3 pathways merge at the level of complement component C3 activation. The complement component C4 plays a central role in the activation of the classic and lectin pathways. Complement system functions at the interface between innate and adaptive immunities. It plays an essential role in host defense against infections and in the elimination of circulating immune complexes.4 Per-
sistently increased levels of circulating immune complexes accelerate atherosclerosis.5–7 In myocardial infarction, complement activation occurs in damaged tissues.3,8 Necrotic and apoptotic cells increase complement activation,9 resulting in increased consumption or depletion of C3 and C4. Increased serum C310 –12 and the presence of C4 null alleles13–15 are risk factors for coronary heart disease. Low production, increased consumption of C4, and increased expression of C3 regulated by different inflammatory cytokines modify the C3/C4 ratio. Therefore, we tested whether the C3/C4 ratio predicts ischemic cardiovascular morbidity and its relation to other known risk factors. Methods
Divisions of aCardiology and bInfectious Diseases, Department of Medicine, Helsinki University Central Hospital; and cDepartment of Bacteriology and Immunology and dTransplantation Laboratory, Haartman Institute, University of Helsinki, Helsinki; and eLaboratory of Applied Mathematics, Lappeenranta University of Technology, Lappeenranta, Finland. Manuscript received August 29, 2006; revised manuscript received and accepted November 7, 2006. This study was funded by the University of Helsinki Foundation, Helsinki; the Finnish-Norwegian Medical Foundation, Helsinki; the Aarne Koskelo Foundation, Helsinki; the Finnish Foundation for Cardiovascular Research, Helsinki; the Helsinki University Central Hospital Research Funds (EVO), Helsinki; and the Finnish Red Cross Blood Transfusion Service Research Fund, Helsinki, Finland. *Corresponding author: Tel: 358-9-191-26614; fax: 358-9-241-1227. E-mail address:
[email protected] (M.-L. Lokki). 0002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2006.11.034
Patients with acute coronary syndrome (n ⫽ 148) were collected from September 1998 to December 2000 from 9 central hospitals in Finland, as previously published.16 Patients with prolonged chest pain with ST-wave changes indicating unstable angina (n ⫽ 43) or non–Q-wave myocardial infarction (n ⫽ 105) were enrolled. ST-elevation myocardial infarction was an exclusion criterion. For inclusion, patients had to have clear symptoms of angina with electrocardiographic evidence of myocardial ischemia. Patients who met the anginal pain inclusion criteria but none of the electrocardiographic criteria were eligible to enter the trial if their cardiac enzymes were consistent with the occurrence of myocardial infarction. Patient treatment, followwww.AJConline.org
Coronary Artery Disease/Complement C3/C4 Ratio and ACS Table 1 Patient characteristics Characteristic
Baseline (n ⫽ 148)
Without End Points (n ⫽ 104)
With End Points (n ⫽ 44)
Age (yrs) Women Body mass index (kg/m2) Smoking Never Ex-smoker Current smoker Hypertension Diabetes Hypercholesterolemia High-sensitivity C-reactive protein (g/ml) Medications Acetylsalicylic acid -adrenergic blockers ACE/AT inhibitor Statins
63.7 ⫾ 0.6 48 (32.4%) 27.1 ⫾ 0.3
63.2 ⫾ 1.0 34 (32.7%) 26.8 ⫾ 0.4
64.5 ⫾ 1.4 14 (31.8%) 27.7 ⫾ 0.6
63 (42.6%) 45 (30.4%) 40 (27.0%) 62 (41.9%) 28 (18.9%) 113 (76.4%) 16.9 ⫾ 2.6
43 (41.3%) 34 (32.7%) 27 (26.0%) 40 (38.5%) 14 (13.5%)* 78 (75.0%) 17.8 ⫾ 3.4
20 (45.5%) 11 (25.0%) 13 (29.5%) 22 (50.0%) 14 (31.8%)* 35 (79.5%) 14.9 ⫾ 3.8
141 (95.3%) 126 (85.1%) 29 (19.6%) 61 (41.2%)
99 (95.2%) 87 (83.7%) 17 (16.3%) 43 (41.3%)
42 (95.5%) 39 (88.6%) 12 (27.3%) 18 (40.9%)
Data are means ⫾ SEMs or numbers of patients (percentages). * Significant difference was found between presence and absence of end points (OR 3.0, 95% CI 1.29 to 7.01, p ⫽ 0.009). The remaing characteristics were not significant. ACE ⫽ angiotensin-converting enzyme; AT ⫽ angiotensin.
up, and end points were as referenced previously.16 Patients’ baseline characteristics are listed in Table 1. All together, 44 patients developed ischemic cardiovascular events (composite end points of death, myocardial infarction, recurrent unstable angina, or stroke) during the follow-up (555.41 ⫾ 21.17 days). Subjects were from a placebo-controlled clarithromycin study on acute coronary syndrome.16 Blood samples were taken within 48 hours of hospitalization (visit 1), at 1 week (visit 2), at 3 months (visit 3), and at 1 year (visit 4) after hospital admission. At the first visit, 36 patients (24.32%) had a documented previous myocardial infarction. We reanalyzed the C3 and C4 concentrations measured initially, and the mean level from all 4 visits was calculated to indicate the “mean of all visits.” All patients gave written informed consent. The study protocol was approved by the ethics committee of the Department of Medicine, Hospital District of Helsinki and Uusimaa (Helsinki, Finland). C3 and C4 concentrations in serum samples were measured by nephelometry using polyclonal antibodies and reference samples from Boehringer Mannheim (Mannheim, Germany). Complement C4 allotyping (protein variants and deficiencies for C4A and C4B) was performed as referenced elsewhere.17 Cholesterol, glucose, blood cell count, probrain natriuretic peptide (pro-BNP), high-sensitivity C-reactive protein, cardiac enzymes, glycohemoglobin, liver enzymes, and thromboplastin time were measured according to laboratory standards of the Helsinki University Central Hospital (Helsinki, Finland). Ratios of C3 and C4 concentrations in serum samples were calculated. To represent the entire study period, analyses were made using the mean of all visits, if not otherwise stated. The C3/C4 ratio was considered first as a continuous
891
and then as a categorical parameter, as described in the following. The continuous variables were expressed as mean ⫾ SEM. In the entire study group, C3/C4 ratio values followed a normal distribution. Thus, parametric tests were employed. Student’s t test and repeated measures analysis of variance (ANOVA) were employed to assess whether the C3/C4 ratios differed between patients with and those without end points and C4 null alleles (C4A*Q0 or C4B*Q0 allele, indicated as C4*Q0). Multivariate logistic regression analysis was used to study major risk factors (age, gender, blood pressure, diabetes mellitus, body mass index, and smoking status), lipids (total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglyceride levels in serum), high-sensitivity C-reactive protein, Canadian Cardiac Society classes, pro-BNP, troponin T, medications (acetylsalicylic acid, -adrenergic blockers, angiotensin-converting enzyme or angiotensin inhibitor, and statins), complement factors (C3/C4 ratio), and other clinical and laboratory parameters (pulse rate, glucose, leukocyte count, thrombocyte count, hemoglobin, glycohemoglobin, creatinine, liver enzymes, and thromboplastin percentage). By Cox multivariate logistic regression survival analysis, the relation between baseline covariates and end points was evaluated with odds ratios, their 95% confidence intervals (CIs), and p values. To provide a categorical parameter, C3/C4 ratio values were divided into low (Q1 ⫹ Q2), medium (Q3), and highest (Q4) quartiles. To confirm the quartile as a result, a cut-off value was determined (Figure 1). Categorical variables (quartiles, having a C3/C4 ratio above or below the cut-off value, end points, and C4 null alleles) were analyzed with chi-square test or Fisher’s exact test to assess differences between study groups. Relative risks were expressed as odds ratios (ORs). Ischemic cardiovascular events as end points during follow-up were analyzed with the Kaplan-Meier test. Follow-up period as time, end points as events, and categorical parameters studied (quartiles and C4A*Q0 allele) were compared with log-rank test. For all statistical tests, a p value ⬍0.05 indicated a statistically significant difference. Data were analyzed with SPSS 12.0.1 (SPSS, Inc., Chicago, Illinois). Results Patients’ baseline characteristics are listed in Table 1. The used medications (Table 1) showed no significant differences between patients with and those without the end points. C3 and C4 concentrations remained stable throughout the entire trial (Figure 2). In the entire series of recruited patients, concentrations showed a strong positive correlation (Pearson correlation r ⫽ 0.77, p ⫽ ⬍0.00001, C4 ⫽ ⫺0.14 ⫹ 31 ⫻ C3; Figure 2, regression line not plotted). Of 148 patients recruited, 44 patients met ⱖ1 end point (unstable angina pectoris, n ⫽ 14; myocardial infarction, n ⫽ 19; cerebral infarction, n ⫽ 4; or death, n ⫽ 7). Elective bypass operation (coronary artery bypass grafting, n ⫽ 36) or balloon angioplasty (percutaneous transluminal coronary angioplasty, n ⫽ 33) were not considered end points. No significant (p ⫽ 0.38) difference was found between unstable angina pectoris (n ⫽ 15, 34.9%) and non–Q-wave myo-
892
The American Journal of Cardiology (www.AJConline.org)
Figure 1. Schematic diagram and procedures of determining the (A) C3/C4 ratio cut-off value and (B) cut-off value analysis at receiver operating characteristic curve analysis. Averages of C3/C4 ratios on visits 1 and 2 were aligned. The mean of all C3/C4 values (mean) was higher in patients with end points than in those without end points. The 2 groups defined an overlapping region (dark area). Within the overlapping region, the mean value (mean of overlapping region) was higher in patients with end points than in those without end points. The cut-off value (4.53) was determined from this region and further examined with receiver-operating characteristic curve statistics.
cardial infarction (n ⫽ 29, 27.6%) when examined with end points. C4 values were lower in patients with end points than in patients without end points (repeated measures ANOVA, p ⫽ 0.06). Patients with end points showed statistically nonsignificant but increasing trends of C3 levels at all 4 visits (data not shown); however, the most prominent difference was at visit 2 (Student’s t test, p ⫽ 0.057). C3/C4 ratio values in patients who received the original study medication, clarithromycin (4.59 ⫾ 0.12), showed no difference compared with values in patients receiving placebo (4.69 ⫾ 0.15, repeated measures ANOVA, p ⫽ 0.3). C3/C4 ratios in all patients (4.64 ⫾ 1.00) were divided into quartiles or as being below or above the cut-off value. When regression lines of quartiles were plotted on serum C3 and C4 concentrations, the highest quartile showed the lowest slope and the low quartiles the highest slope. A 1-U change in C4 concentration generated a greater change in C3 concentration in the highest quartile than in the medium and low quartiles (Figure 2). Similar effects were found using the cut-off value (data not shown). Patients’ medical
history showed that 36 had a previous acute myocardial infarction. Patients with previous acute myocardial infarction had marginally higher C3/C4 ratio values than patients without previous acute myocardial infarction (4.94 ⫾ 0.23 vs 4.55 ⫾ 0.10, Student’s t test, p ⫽ 0.087). Patients with unstable angina had significantly higher C3/C4 ratio values than those with non–Q-wave myocardial infarction (4.99 ⫾ 0.19 vs 4.50 ⫾ 0.11, Student’s t test, p ⫽ 0.019). Patients with end points showed significantly higher C3/C4 ratios than patients without end points (repeated measures ANOVA, p ⫽ 0.007), except at visit 2, when only a borderline difference was reached with Student’s t test. The C3/C4 ratio was lowest at visit 2 (7 days after an acute coronary syndrome attack) and increased thereafter (Figure 2). When we performed Cox multivariate logistic regression survival analysis of patients with end points between baseline covariates, C3/C4 ratio (OR 1.33, 95% CI 1.08 to 1.63, p ⫽ 0.007) and diabetes mellitus (OR 2.12, 95% CI 1.06 to 4.24, p ⫽ 0.03) appeared as risk factors. Patients in the highest quartile had more end points and the worst outcome
Coronary Artery Disease/Complement C3/C4 Ratio and ACS
893
Figure 2. (A) Means of serum C3 and C4 concentrations. (B) Regression lines of C3/C4 ratio quartile on serum C3 and C4 correlation. The formula slope ⫽ ⌬Y/⌬X (where ⌬Y represents a 1-U change in serum C4 and ⌬X represents a corresponding change in serum C3) showed that a 1-U change in C4 concentration generated a 1.70-fold change in C3 concentration in the highest quartile and a 1.24-fold change in the medium quartile compared with the change in the low quartiles. (C) Difference in C3/C4 ratio for presence versus absence of end points. (D) Effect of C4*Q0 alleles on C4 concentrations. aThe p values by Student’s t test when visits were tested separately.
(Figure 3). Patients in the low quartiles had the best outcome. Quartiles started to separate from the visit 1, and the difference between the highest and low quartiles increased for up to 300 days and remained approximately the same thereafter. Corresponding results were observed when using the cut-off value (4.53, chi-square in those with a value higher vs lower than the cut-off value, OR 2.25, 95% CI 1.09 to 4.65, p ⫽ 0.03). The sensitivity and specificity of the cut-off value were 0.62 and 0.63, respectively (Figure 1). Because C4 levels are influenced by the number of C4 genes, the number of patients with C4 null alleles (C4A*Q0 or C4B*Q0) was greater in the highest quartile than in low quartiles (Table 2). Patients carrying C4 null alleles had a significantly lower serum C4 concentration than patients without C4 null alleles (Figure 2) at all except visit 1. Serum C4 concentration was highest at visit 2 (7 days after an acute coronary syndrome attack) and decreased thereafter. Presence of C4A*Q0 alleles was associated with a larger num-
ber of end points and worse outcome (chi-square for presence vs absence of C4A*Q0, OR 2.94, 95% CI 1.17 to 7.40, p ⫽ 0.02). Pro-BNP level in serum was measured at the first visit. Patients without end points showed lower pro-BNP values (1,131 ⫾ 175 ng/L) than those with end points (1,504 ⫾ 441 ng/L), but the difference was not statistically significant (p ⫽ 0.1). Discussion In this study, we have shown that patients with acute coronary syndrome and a high C3/C4 ratio are prone to recurrent cardiovascular attacks. A high C3/C4 ratio was a strong risk factor and a better predictor for an end point than serum C3 and C4 concentrations alone. The association between an increased C3/C4 ratio and reaching an end point in patients with acute coronary syndrome was analyzed by a
894
The American Journal of Cardiology (www.AJConline.org)
.
.
.
.
.
.
Figure 3. Kaplan-Meier plot of cumulative survival during follow-up in relation to low (n ⫽ 17 of 72), medium (n ⫽ 11 of 40), and highest (16 of 33) quartiles of C3/C4 ratio. Log-rank test for low versus highest quartiles (p ⫽ 0.005) and medium versus highest quartiles (p ⫽ 0.08). Chi-square test for low versus highest quartiles (OR 3.04, 95% CI 1.27 to 7.29, p ⫽ 0.01). Data are numbers of patients with ischemic cardiovascular events/ total number of patients (missing values, n ⫽ 3). Table 2 Association of C4 null alleles* with C3/C4 ratio quartiles C3/C4 Ratio Quartiles Low quartiles (1 ⫹2) Medium quartile (3) High quartile (4) p value‡ OR (95% CI)‡
C4A*Q0
C4B*Q0
C4*Q0†
9/49 (18.4%)
18/49 (36.7%)
27/49 (55.1%)
7/25 (28.0%)
8/25 (32.0%)
15/25 (60.0%)
10/24 (41.7%)
14/24 (58.3%)
21/24 (87.5%)§
0.03 0.08 0.006 3.17 (1.07–9.41) 2.41 (0.89–6.54) 5.70 (1.50–21.66)
Data are numbers of patients/total patients (percentages). * Of 148 patients 98 patients randomly selected were studied. † Presence of C4A (C4A*Q0) or C4B (C4B*Q0) null allele. ‡ The p and OR values were calculated with chi-square or Fisher’s exact test for highest versus low quartiles. § Patients with C4A*Q0 and C4B*Q0 alleles (n ⫽ 3).
quartile approach and by using a cut-off value for the C3/C4 ratio. These results corroborated with each other. An increased ratio was caused by a relative increase in C3 level and a relative decrease in C4 level. In the highest quartile, as many as every second patient had a cardiovascular event during the 1.5-year follow-up. In the low quartiles, 3 of 4 patients had no cardiovascular attacks. The complement system is composed of ⬎30 serum proteins and has immunoprotective and proinflammatory properties.2,3 Changes in serum complement levels occur in the acute-phase response18 and in several acquired diseases
in which complement consumption and synthesis are altered.19,20 Inherited complement deficiencies can also lead to low levels of the respective complement proteins.21 Complement factors C3 and C4 are mostly synthesized in the liver but also in various other tissues22,23 and in the infarcted heart.8 C3 and C4 protein levels in blood are increased during the acute-phase reaction18,24 and consumed during complement activation, e.g., by damaged tissue, C-reactive protein complexes, or immune complexes. During classic pathway activation, C4 may become consumed. In the presence of C4 null alleles, the level of C4 synthesis may not match the need for C4. In acute coronary syndrome, persistent inflammation is a main underlying factor.1 An increased serum C3 concentration and the presence of C4 null alleles are risk factors for cardiovascular disease.10 –15 With this in mind, we prospectively measured the ratio of C3 and C4 serum concentrations and found it to be a novel risk marker of forthcoming cardiovascular attacks in patients who had an acute coronary syndrome. An increased C3/C4 ratio associated positively with end points regardless of the time point of sample collection or disease phase. To our knowledge, this is the first study to use the C3/C4 ratio as a predictive marker in human disease. In repeated measure ANOVA, end points were found to correlate with high C3/C4 ratio values. Multivariate analysis confirmed the results; after including all the confounding factors, the increased value of C3/C4 ratio emerged as the novel risk factor. These results were analyzed by dividing the values into quartiles and by calculating a cut-off value. A C3/C4 ratio in the highest quartile and a C3/C4 ratio higher than the cut-off value were associated with the composite of the end point. We compared the low quartiles (Q1 ⫹ Q2) with the highest quartile (Q4) and found that the low quartiles had fewer end points than the highest quartile. However, when the lowest quartile (Q1) was compared with the highest quartile (Q4), we found only a trend-like difference, possibly caused by the small sample. We further attempted to confirm the quartile results by defining 4.53 as the cut-off value for the C3/C4 ratio. This had good sensitivity and specificity for the end points. The results using the cut-off value corroborated the results from the quartile method. Due to a lack of a control population, the cut-off value obtained from patients with acute coronary syndrome was compared with ratios from a healthy control group of ⬎55,000 Caucasians from the northern New England area.25 Ratios of the control group were calculated by using the median value obtained from the reference ranges. For these control subjects, the C3/C4 ratio (4.18) was lower than in our patients. Acknowledgment: We thank Abhishek Tripathi, MSc, for his advice on statistics. 1. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med 2005;352:1685–1695. 2. Walport MJ. Complement. First of two parts. N Engl J Med 2001;344: 1058 –1066. 3. Walport MJ. Complement. Second of two parts. N Engl J Med 2001; 344:1140 –1144. 4. Schifferli JA, Ng YC, Peters DK. The role of complement and its receptor in the elimination of immune complexes. N Engl J Med 1986;315:488 – 495.
Coronary Artery Disease/Complement C3/C4 Ratio and ACS 5. Wissler RW, Vesselinovitch D, Ko C. The effects of circulating immune complexes on atherosclerotic lesions in experimental animals and in younger and older humans. Transplant Proc 1989;21:3707– 3708. 6. Wissler RW. Update on the pathogenesis of atherosclerosis. Am J Med 1991;91(suppl):3S–9S. 7. Qiao JH, Castellani LW, Fishbein MC, Lusis AJ. Immune-complex– mediated vasculitis increases coronary artery lipid accumulation in autoimmune-prone MRL mice. Arterioscler Thromb 1993;13:932– 943. 8. Yasojima K, Schwab C, McGeer EG, McGeer PL. Human heart generates complement proteins that are upregulated and activated after myocardial infarction. Circ Res 1998;83:860 – 869. 9. Ciurana CL, Zwart B, van Mierlo G, Hack CE. Complement activation by necrotic cells in normal plasma environment compares to that by late apoptotic cells and involves predominantly IgM. Eur J Immunol 2004;34:2609 –2619. 10. Muscari A, Bozzoli C, Puddu GM, Sangiorgi Z, Dormi A, Rovinetti C, Descovich GC, Puddu P. Association of serum C3 levels with the risk of myocardial infarction. Am J Med 1995;98:357–364. 11. Szeplaki G, Prohaszka Z, Duba J, Rugonfalvi-Kiss S, Karadi I, Kokai M, Kramer J, Fust G, Kleiber M, Romics L, Varga L. Association of high serum concentration of the third component of complement (C3) with pre-existing severe coronary artery disease and new vascular events in women. Atherosclerosis 2004;177:383–389. 12. Onat A, Uzunlar B, Hergenc G, Yazici M, Sari I, Uyarel H, Can G, Sansoy V. Cross-sectional study of complement C3 as a coronary risk factor among men and women. Clin Sci (Lond) 2005;108:129 –135. 13. Szalai C, Fust G, Duba J, Kramer J, Romics L, Prohaszka Z, Csaszar A. Association of polymorphisms and allelic combinations in the tumour necrosis factor-alpha– complement MHC region with coronary artery disease. J Med Genet 2002;39:46 –51. 14. Kramer J, Rajczy K, Hegyi L, Fulop T, Mohacsi A, Mezei Z, Keltai M, Blasko G, Ferenczy E, Anh-Tuan N. C4B*Q0 allotype as risk factor for myocardial infarction. BMJ 1994;309:313–314.
895
15. Arason GJ, Kramer J, Bodvarsson S, Sigurdarson ST, Sigurdsson G, Fust G, Thorgeirsson G. Increased frequency of C4B null alleles in Icelandic patients with coronary disease. Mol Immunol 1998;35:412. 16. Sinisalo J, Mattila K, Valtonen V, Anttonen O, Juvonen J, Melin J, Vuorinen-Markkola H, Nieminen MS, for the Clarithromycin in Acute Coronary Syndrome Patients in Finland (CLARIFY) Study Group. Effect of 3 months of antimicrobial treatment with clarithromycin in acute non– q-wave coronary syndrome. Circulation 2002;105:1555– 1560. 17. Marcus D, Alper CA. Methods for allotyping complement proteins. In: Rose NR, Friedman H, Fahey JL, eds. Manual of Clinical Laboratory Immunology, 3rd Ed. Washington, DC: American Society for Microbiology, 1986:185–196. 18. Schorlemmer HU. The role of complement in the function of the monocyte-macrophage system. In: Schmalzl F, Hahn D, Schaeffer HE, eds. Haematology and Blood Transfusion. Disorders of the Monocyte Macrophage System. New York: Springer-Verlag, 1981:59 –71. 19. Nielsen CH, Fischer EM, Leslie RG. The role of complement in the acquired immune response. Immunology 2000;100:4 –12. 20. Nielsen CH, Leslie RG. Complement’s participation in acquired immunity. J Leukoc Biol 2002;72:249 –261. 21. Crawford K, Alper CA. Genetics of the complement system. Rev Immunogenet 2000;2:323–338. 22. Cox BJ, Robins DM. Tissue-specific variation in C4 and Slp gene regulation. Nucleic Acids Res 1988;16:6857– 6870. 23. Cole FS, Matthews WJ Jr, Marino JT, Gash DJ, Colten HR. Control of complement synthesis and secretion in bronchoalveolar and peritoneal macrophages. J Immunol 1980;125:1120 –1124. 24. Gabay C, Kushner I. Acute-phase proteins and other systemic responses to inflammation. N Engl J Med 1999;340:448 – 454. 25. Ritchie RF, Palomaki GE, Neveux LM, Navolotskaia O, Ledue TB, Craig WY. Reference distributions for complement proteins C3 and C4: a practical, simple and clinically relevant approach in a large cohort. J Clin Lab Anal 2004;18:1– 8.
