Association,of,serum,complement,C1q,with,cardiovascular,outcomes,among,patients,with,acute,coronary,syndrome,undergoing,percutaneous,coronary,intervention

Qiu-Xuan LI, Xiao-Teng MA, Qiao-Yu SHAO, Zhi-Qiang YANG, Jing LIANG, Li-Xia YANG,Dong-Mei SHI, Yu-Jie ZHOU, Zhi-Jian WANG

Department of Cardiology, Beijing Anzhen Hospital, Capital Medical University, Beijing Institute of Heart Lung and Blood Vessel Disease, Beijing Key Laboratory of Precision Medicine of Coronary Atherosclerotic Disease, Clinical Center for Coronary Heart Disease, Capital Medical University, Beijing, China

ABSTRACT OBJECTIVE To determine the association of serum complement C1q levels with cardiovascular outcomes among patients with acute coronary syndrome (ACS) undergoing percutaneous coronary intervention (PCI), and evaluate the value of C1q modified by high-sensitivity C-reactive protein (hs-CRP) levels as an independent predictor. METHODS As a single-center prospective observational study, we analyzed 1701 patients who had received primary or elective PCI for ACS at Beijing Anzhen Hospital, Capital Medical University, Beijing, China between June 1, 2016 and November 30,2017. The associations of C1q modified by hs-CRP with major adverse cardiovascular events (MACE) were determined in survival analysis. RESULTS Patients with the lowest C1q tertile had the highest cumulative risk of MACE (log-rank P = 0.007). In fully adjusted Cox regression models, stratifying the total population according to hs-CRP dichotomy, C1q was significantly associated with MACE in patients with hs-CRP levels less than 2 mg/L but not in those with 2 mg/L or more (Pinteraction = 0.02). In patients with hs-CRP levels less than 2 mg/L, with the lowest C1q tertile as reference, the risk of MACE was reduced by 40.0% in the middle C1q tertile [hazard ratio (HR) = 0.600, 95% CI: 0.423-0.852, P = 0.004] and by 43.9% in the highest C1q tertile (HR = 0.561, 95% CI: 0.375-0.840, P = 0.005). CONCLUSIONS Serum complement C1q is significantly associated with cardiovascular outcomes in patients with ACS undergoing PCI, only when hs-CRP levels are less than 2 mg/L. This finding implicates the usefulness of C1q for the risk stratification in ACS patients with reduced systemic inflammation.

Acute coronary syndrome (ACS) is an important challenge due to the increased mortality risk.[1]The development of the percutaneous coronary intervention (PCI) techniques and instruments, and the launch of newer medications have significantly improved the short- and long-term prognoses of such patients. However, patients with ACS undergoing PCI are still at a high risk of future ischemic events.[2]Therefore, risk stratification is essential for optimizing subsequent medical treatment and improving cardiovascular outcomes.

The formation of atherosclerotic plaque is a chronic inflammatory process in which lipids and inflammatory cells gradually accumulate. The inflammatory response to the complement system plays an important role in the development of atherosclerosis. As the promoter of the classical complement system, C1q forms the C1 complex (C1qC1r2C1s2),which is a major contributor to atherosclerosis progression, activating the classical complement cascade (CCC) with the participation of other complement components.[3]The combination of cholesterol crystals deposited in the vessel wall with C1q has been proven to activate CCC and trigger plaque inflammation in the early stage of atherosclerosis.[4]It has been reported that anaphylatoxins C5a and C3a as downstream signals of CCC attracting inflammatory cells are excellent predictors of cardiovascular morbidity and mortality.[5,6]In the later stage of atherosclerosis, a large number of inflammatory signals are activated. The necrotic cells in the plaques increase, forming larger lipid necrotic core and thinner fiber cap. Up-regulation of inflammatory signals promotes the development and rupture of unstable plaques, leading to ACS.[7]C-reactive protein(CRP) is abundant in the necrotic core of the plaque,[8]and increased CRP levels can independently predict all-cause and cardiovascular mortality.[9]It has been proved that CRP combines with the C1q,[10]promoting C1q-mediated complement activation,and enhancing the recruitment of leukocytes in inflammatory tissues.[11]

However, several experiments have emphasized that C1q may have a protective effect on the development of atherosclerosis. C1q in combination with apoptotic cells through complement activation eliminates apoptotic cells and cell debris from plaques.[12,13]C1q also enhances clearance of atherogenic lipoproteins by macrophages as a non-complement role,[14]during which C1q promotes macrophage anti-inflammatory M2-like polarization.[15]

From the above-mentioned aspects, C1q may play a complex role in the development of atherosclerosis. However, to date, the effect of C1q on ACS has not been fully studied.[16]In view of the complexity of the role of C1q in atherosclerosis, it is necessary to carry out relevant studies to explore the prognostic value of C1q in patients with ACS.

Study Design and Population

We admitted all patients who had received primary or elective PCI for ACS at Beijing Anzhen Hospital, Capital Medical University, Beijing, China between June 1, 2016 and November 30, 2017. ACS was diagnosed according to the American College of Cardiology/American Heart Association guidelines.[17,18]All PCI procedures were performed according to current standard guidelines, with the treatment strategies at the discretion of the operator. Weight-adjusted intraprocedural unfractionated heparin (with a goal-activated clotting time of 250-300 s)was administered during the procedure and was routinely discontinued at the end of the procedure. All patients received dual antiplatelet therapy from the day of the procedure. The type of P2Y12inhibitor (clopidogrel or ticagrelor) and the duration of the dual antiplatelet therapy for each patient were determined by the operators or the clinical physicians. We excluded patients with prior coronary artery bypass grafting, renal dysfunction with creatinine clearance ＜ 30 mL/min, Killip class ＞ II, left ventricular ejection fraction ＜ 30%, acute and/or chronic infection, autoimmune diseases, and known malignancy.

This study was performed by the Helsinki Declaration of Human Rights and was approved by the Institutional Review Board of Beijing Anzhen Hospital, Capital Medical University, Beijing, China (No.2016034X). All patients gave their written informed consent before study inclusion. The registration number of this study is ChiCTR1800017417 (http://www.chictr.org.cn/hvshowproject.aspx?id=21397).

Measurements

Demographics, lifestyle, medical history, and medication history were collected using standard questionnaires. The concentrations of complement C1q,high-sensitivity C-reactive protein (hs-CRP), triglycerides, total cholesterol, high-density lipoprotein cholesterol, fasting plasma glucose, and glycated hemoglobin in the first fasting blood samples during hospitalization, which were obtained after 12 h of fasting, were measured in the Central Laboratory of Beijing Anzhen Hospital, Capital Medical University, Beijing, China. Brain natriuretic peptide (BNP)and cardiac troponin (cTn) were based on the levels at admission. The Friedewald equation was used to calculate the low-density lipoprotein (LDL) cholesterol level. Creatinine clearance was calculated by using the Cockcroft-Gault formula. Hypertension was defined as blood pressure measured three times on different days ≥ 140/90 mmHg and/or the use of antihypertensive drugs. Diabetes mellitus was defined as diabetic symptoms with random blood glucose ≥ 200 mg/dL or fasting plasma glucose ≥ 126 mg/dL or 2-hour blood glucose concentration ≥ 200 mg/dL from 75 g oral glucose tolerance test, and/or use of anti-diabetic drugs. Dyslipidemia was defined as total cholesterol ＞ 200 mg/dL, LDL cholesterol ＞ 130 mg/dL, triglycerides ＞ 150 mg/dL, highdensity lipoprotein cholesterol ＜ 40 mg/dL, and/or use of lipid-lowering drugs. Left ventricular ejection fraction was obtained from echocardiography measuring by professional echocardiologists after admission. The records of diseased blood vessels and stents used were jointly decided by two or more senior PCI surgeons, which were based on the results of the angiography during the operation. With the basic data of patients, we calculated the GRACE risk score and SYNTAX score according to the corresponding guidelines.[19,20]

Follow-up and Outcomes

All patients started their first follow-up one month after discharge and every six months thereafter.Adverse events were obtained by trained personnel who never knew the baseline characteristics of the patients through telephone contact or by office visits.The primary endpoint of this study was major adverse cardiovascular events (MACE), which was identified as the composite of all-cause death, nonfatal stroke, non-fatal myocardial infarction (MI), or unplanned repeat revascularization. The definition of each component was described in detail in our previously published papers. Stroke was defined as ischemic cerebral infarction with neurological dysfunction, which had been clinically documented on brain computed tomography or magnetic resonance imaging. MI was defined as the levels of cTn or the MB fraction of creatine kinase exceeding the upper limit with either ischemic symptoms or electrocardiographic changes implicating ischemia, or the new pathological Q waves appearing in ≥ 2 contiguous leads. Within one week after the index PCI, only Qwave MI was defined as MI. Unplanned repeat revascularization was defined as any non-staged revascularization after index PCI. Staged revascularization was defined as scheduled revascularization within 90 days after index PCI unless revascularization was driven by severe ischemia. If more than one event occurred during follow-up, the most severe endpoint event was chosen for the primary endpoint analysis (all-cause death ＞ stroke ＞ MI ＞ revascularization). If more than one event of a stroke or MI or revascularization occurred, the first event of a stroke or MI or revascularization was chosen.

Statistical Analysis

Categorical variables were presented as counts(percentages) and were compared with the Pearson’s chi-squared test. Continuous variables with a normal distribution were presented as mean ± SD and were compared by the analysis of variance; however, continuous variables without a normal distribution were presented as the median (interquartile range) and were compared by the Kruskal-WallisHtests. All analyzed patients were divided into three groups according to the tertiles of C1q levels (T1:102.3-168.4 mg/L, T2: 168.4-198.7 mg/L, and T3:198.7-371.2 mg/L). The Kaplan-Meier method was used to estimate the cumulative risk of cardiovascular outcomes over time, data were compared using the log-rank test among three groups. Restricted cubic spline curves with five knots were used to show the non-linear associations between C1q and MACE.Multivariate Cox proportional hazards regression models adjusting for other baseline covariates were used to estimate the hazard ratio (HR) with the 95%confidence interval (CI) of five different outcomes(MACE, all-cause death, non-fatal stroke, non-fatal MI, and unplanned revascularization). After stratifying by the levels of hs-CRP and age, the relationships between C1q and MACE were further explored by the multivariate Cox regression analyses. The validity of the proportionality assumption was verified for all covariates by a visual examination of the log (minus log) curves and a test based on the Schoenfeld residuals. Statistical analyses were performed using the SPSS 24.0 (SPSS Inc., IBM, Chicago,IL, USA) and the R statistical software 3.6.3 (http://www.r-proje ct.org). A two-sidedP-value ＜ 0.05 was considered statistically significant.

Initially, there were 1705 eligible patients, but four patients were excluded because of missing followup data despite at least four separate attempts to contact them. Finally, 1701 patients were included in the analysis (supplemental material, Figure 1S).The average age of the 1701 patients was 60 ± 10 years,and 76.7% of the patients were men (n= 1305). At baseline, 63.6% of patients had hypertension, 46.0%of patients had diabetes mellitus, and 79.9% of patients had dyslipidemia (Table 1).

Cohort Demographics

The follow-up period lasted until November 30,2019 and the median follow-up period was 30 months (interquartile range: 30-36 months). During the follow-up period, MACE occurred in 350 patients(20.6%), of which 138 patients (24.3%) were from T1,95 patients (16.8%) were from T2, and 117 patients(20.6%) were from T3. In the 350 patients who had MACE, there were 43 patients (2.5%) had death (36 cardiovascular death and seven non-cardiovascular death), 24 patients (1.4%) had non-fatal stroke, 49 patients (2.9%) had non-fatal MI, and 286 patients (16.8%)had unplanned repeat revascularization (Table 2).

Table 1 Baseline characteristics of study population across tertiles of C1q levels.

Continued

The baseline characteristics of all analyzed patients stratified according to C1q tertiles were summarized (Table 1). The proportion of unstable angina (UA)and ST-segment elevation MI had a non-linear relationship with C1q tertiles. T2 had the highest proportion of UA (71.0%vs.80.7%vs.71.8%,P＜ 0.001),and had the lowest proportion of ST-segment elevation MI (14.8%vs.8.3%vs.15.3%,P＜ 0.001). The levels of BNP, cTn, hs-CRP, and GRACE risk score were also non-linearly related to C1q tertiles. T2 had the lowest levels of BNP, cTn (bothP= 0.001), hs-CRP(P＜ 0.001), and GRACE risk score (P= 0.006). Other baseline characteristics among C1q tertiles were not statistically different. There were no statistical differences among the three groups in medication before admission and after discharge (allP＞ 0.05).

Relationship between C1q Tertiles and Cardiovascular Outcomes

There were significant differences in the incidences of MACE as well as all-cause death, cardiovascular death, non-fatal MI or unplanned repeat revascularization among the three C1q tertiles. As described in Table 2 and Figure 1, the T1 had the highest cumulative risk, while the T2 had the lowest cumulative risk, except for non-fatal stroke events.

Adjusted Multivariate Cox Analysis

The cutoff point of hs-CRP was considered at 2mg/L, which has been shown to have obvious grouping significance for clinical outcomes in previous studies.[21,22]Table 3 described the fully adjusted multivariable relationships between MACE stratified according to the dichotomy of hs-CRP and C1q tertiles. In the total population, higher levels of hs-CRP(＜ 2 mg/Lvs.≥ 2 mg/L) were associated with a significantly higher risk of MACE (HR = 1.595, 95% CI:1.277-1.992,P＜ 0.001). For C1q, compared with T1,T2 was associated with a lower risk of MACE (HR =0.704, 95% CI: 0.540-0.916,P= 0.009), whereas T3 was not (HR = 0.894, 95% CI: 0.697-1.147,P= 0.379).The restricted cubic spline curve also demonstrated the non-linear relationship between continuous variable C1q and the incidence of MACE (P= 0.008) (supplemental material, Figure 2S).

Figure 1 Kaplan-Meier curves of C1q tertiles and cumulative incidence of cardiovascular outcomes at follow-up. (A): MACE; (B):all-cause death; (C): cardiovascular death; (D): non-fatal stroke; (E): non-fatal myocardial infarction; and (F): unplanned repeat revascularization. The primary endpoint was defined as a composite of all-cause death, non-fatal ischemic stroke, non-fatal myocardial infarction, and unplanned repeat revascularization. MACE: major adverse cardiovascular event.

Table 2 The number and proportion of cardiovascular outcomes in tertiles of C1q levels.

C1q was significantly associated with MACE only when hs-CRP levels were less than 2 mg/L. As is shown in Table 3 and Figure 2, with T1 as reference,the risk of MACE was reduced by 40.0% in the T2(HR = 0.600, 95% CI: 0.423-0.852,P= 0.004) and by 43.9% in T3 (HR = 0.561, 95% CI: 0.375-0.840,P=0.005). When hs-CRP levels were 2 mg/L or more,C1q was no longer related to MACE. There was a significant interaction for MACE between C1q tertiles and the hs-CRP dichotomy (Pinteraction= 0.021).The multivariate Cox regression analyses of C1q and the other four clinical outcomes were summarized (supplemental material, Table 1S).

Figure 2 Cardiovascular outcomes according to C1q tertiles and hs-CRP dichotomy in fully adjusted multivariable models. The point represents the hazard ratio and the vertical line represents 95% confidence interval. T1 (102.3-168.4 mg/L) of C1q tertiles as the reference group. The multivariable models are the same as Table 3. hs-CRP: high-sensitivity C-reactive protein; MACE: major adverse cardiovascular event.

Table 3 Associations between MACE stratified according to the dichotomy of hs-CRP and C1q tertiles.

In addition, we further explored the influence of C1q on MACE in different age stratification (supplemental material, Table 2S & Table 3S). We found that people under 65 years with middle level of C1q had 40.2% reduction in MACE compared with the low level. While the value of C1q was not apparent for older patients (age ≥ 65 years). However, C1q has a certain protective effect in patients ＜ 65 years with reduced systemic inflammatory response.

Several previous studies have investigated the relationship between C1q and cardiovascular morbidity and mortality. Ni,et al.[16]found that C1q levels of acute MI patients were lower than control patients and UA patients (P＜ 0.05), and reduced levels of C1q was an adverse factor for acute MI (OR = 0.984,95% CI: 0.972-0.997,P= 0.015) and for ACS (OR =0.984, 95% CI: 0.971-0.984,P= 0.025) in drinking patients. In addition, Cavusoglu,et al.[23]found that C1q levels were an independent predictor of 10-year mortality in men with diabetes mellitus referred for coronary angiography (HR = 0.66, 95% CI: 0.52-0.84,P= 0.0006); and compared with patients with C1q levels ≤ 178 ug/mL, patients with C1q levels ＞ 178 ug/mL had a significantly lower 10-year mortality(P= 0.0109 by log-rank test). Most previous studies analyzed C1q as dichotomies and found that patients with lower C1q had worse prognoses. In our study, patients were divided into three groups according to C1q tertiles, and we found that patients with the lowest level of C1q (102.3-168.4 mg/L) had the highest incidence of MACE, while those with the middle level of C1q (168.4-198.7 mg/L) had the best prognosis. Similar to what we found, Hertle,et al.[24]in the CODAM (Cohort on Diabetes and Atherosclerosis Maastricht) study found that compared with patients in the lowest tertile of serum complement C1q levels, patients in the middle and highest tertiles had lower incidence of cardiovascular disease, although there was no statistically significant difference between the lowest and highest tertiles.These results indicated that: (1) both our study and previous studies believed that lower level of C1q had a negative impact on the long-term prognosis after PCI in ACS patients; and (2) there was a non-linear relationship between C1q level and MACE incidence.

We further explored the effect of inflammation on C1q prediction of long-term outcomes. In fully adjusted analyses, we demonstrated a significant relationship between C1q and MACE only in patients with reduced systemic inflammation (i.e., hs-CRP levels ＜2 mg/L): with the lowest C1q tertile as a reference, patients in the middle tertile had a 40.0% reduced risk of MACE, and those in the highest tertile had a 43.9%reduced risk of MACE. However, there was no such relationship in patients with enhanced systemic inflammation (hs-CRP ≥ 2 mg/L). The above-mentioned results indicated that C1q had a good predictive value for the long-term prognosis after PCI in ACS patients with reduced systemic inflammatory response. As one of the important components in the complement system, C1q causes the complement cascade effect and mediates inflammatory response by initiating the classical activation pathway.In fact, the complement system is just one of many ways to activate inflammation. Therefore, when the level of inflammation is reduced, the role of C1q is relatively prominent. When the level of inflammation is relatively high, the effect of C1q is no longer obvious. The underlying mechanism needs further experimental studies.

Several experiments have investigated how C1q affects the development of atherosclerosis. In the early stage of atherosclerosis, C1q can be produced by dendritic cells, macrophages, and foam cells without other complement components.[25]Both C1q and mannan-binding lectin promote the clearance of modified forms of LDL (including oxidized LDL and acetylated LDL), and what’s more, they reduce the accumulation of free cholesterol in macrophages significantly and at the same time enhance the efflux of high-density lipoprotein-specific cholesterol.[14]Attenuated cholesterol efflux is closely related to the increased risk of atherosclerotic cardiovascular disease.[26-28]Notably, C1q was demonstrated to have a protective role in early atherosclerosis in LDL receptor-knockout mice.[13]C1q also modulates the activation of macrophages directly.[29]C1q suppresses the activation of Janus kinase and signal transducer and activator of transcription pathway, but increases the transcriptional activation of the peroxisome proliferator activating receptor signaling pathway, inhibiting the production of interleukin-6 and reducing the inflammatory signals significantly.[30]The ruptured plaque mainly contains M1-associated inflammatory cytokines, which has been indicated advanced atherosclerosis.[31]Spivia,et al.[15]found that C1q exerted a non-complement function to the polarization of macrophages toward the anti-inflammatory (M2-like) phenotype. Over-expression of antiinflammatory M2-associated cytokine interleukin-10 has been shown to inhibit atherosclerosis in hyperlipidemia mice, which is important in suppressing inflammation in the early stage of atherosclerosis.[15]Apoptosis is one of the characteristics of atherosclerosis and is closely related to vulnerable plaque. In early lesions, even if macrophages form foam cells with lipid droplets accumulated, apoptotic cells can still be effectively removed through the virtuous cycle of macrophage autophagy.[13]The combination of C1q and oxidized LDL can also downregulate specific genes in macrophages, thereby enhancing the uptake and removal of apoptotic cells.[3]On the other hand, C1q can eliminate apoptotic cells and cell debris in atherosclerotic plaques through the classical complement pathway to play a protective role.[12]

As the promoter of the classical complement system, C1q recognizes antigens and activates downstream anaphylatoxins C3 and C5, causing inflammation.[5]As one of the most important inflammatory indicators, CRP levels have been shown to be related to long-term cardiovascular events.[31]It has been shown that as the strongest mediator of C1q binding, CRP can activate the classical complement system, recruiting leukocytes and promoting the expression of downstream inflammatory cytokines.[10,11]These suggest that the interaction between C1q and CRP complicates the respective effects of C1q and CRP on the development of atherosclerosis. Therefore, in our study, we specifically introduced CRP to determine the effect of CRP on the relationship between C1q and cardiovascular outcomes. Intriguingly, our study demonstrated that in ACS patients with reduced systemic inflammation (hs-CRP ＜ 2 mg/L), C1q was a valuable predictor of cardiovascular outcomes. However, in patients with enhanced systemic inflammation (hs-CRP ≥ 2 mg/L), the predictive value of C1q disappeared. The complement pathway initiated by C1q is just one of the many pathways activating inflammation. When the inflammation is evident, the effect of complement may be obscured.

There are several limitations that must be noted.Firstly, this study was an observational study, which cannot exclude the effects of unmeasured and undetected confounding variables. Secondly, C1q is part of the classical complement pathway, but our study did not consider the influence of other complement components on the association of C1q with cardiovascular outcomes. Thirdly, we failed to analyze the effect of dynamic changes of C1q and hs-CRP levels on cardiovascular outcomes. Fourthly, the follow-up information was obtained via telephone, even though we confirmed the authenticity of adverse events by reviewing the corresponding medical records. Last but not least, all patients in the study were Chinese, and whether the conclusions from the present study can be extrapolated to other ethnic groups requires further research.

Serum complement C1q levels are significantly associated with cardiovascular outcomes in patients with ACS undergoing PCI, only when hs-CRP levels are less than 2 mg/L. This finding implicates the use of complement C1q for the risk stratification in ACS patients with reduced systemic inflammation,but requires larger-scale studies to confirm its clinical value.

This study was supported by the National Key Research and Development Program of China (2017 YFC0908800), the China Postdoctoral Science Foundation (2021M692253), and the Beijing Postdoctoral Research Foundation (2021-ZZ-023). All authors had no conflicts of interest to disclose.