Increased circulating C-reactive protein and macrophage-colony stimulating factor are complementary predictors of long-term outcome in patients ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
European Heart Journal (2005) 26, 1618–1624
doi:10.1093/eurheartj/ehi192
Clinical research
Increased circulating C-reactive protein and
macrophage-colony stimulating factor are
complementary predictors of long-term outcome
in patients with chronic coronary artery disease
Ignatios Ikonomidis1*, John Lekakis1, Ioanna Revela1, Felicita Andreotti2, and
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
Petros Nihoyannopoulos2
1
Department of Clinical Therapeutics, University of Athens, Alexandra Hospital, Vas. Sofias 80, Athens 11528, Greece; and
2
Imperial College School of Medicine, National Heart & Lung Institute, Cardiology Department, Hammersmith Hospital,
London, UK
Received 27 September 2004; revised 30 December 2005; accepted 3 February 2005; online publish-ahead-of-print 30 March 2005
KEYWORDS Aims We investigated, in a 6 year follow-up study, whether circulating levels of C-reactive protein (CRP)
Inflammation; and macrophage colony stimulating factor (MCSF) have an independent or complementary prognostic
Coronary artery disease; value in patients with chronic coronary artery disease (CAD).
Long-term prognosis Methods and results MCSF and CRP were measured in 100 patients with chronic CAD. Of 95 (33%)
patients, 31 who completed the 6 year follow-up presented adverse events (death, myocardial infarc-
tion, and unstable angina). In multivariable analysis (including traditional risk factors and medications),
the upper tertiles of MCSF (814 pg/mL) and CRP (2.5 mg/L) levels were independently associated
with a 13- and 6-fold increase in risk of events, respectively (P , 0.01). Patients with combined high
CRP and MCSF had a higher absolute risk of events than patients with elevated MCSF or CRP alone (75
vs. 59 vs. 32%, respectively, P , 0.01). The mean event-free time was 39, 64, and 52 months in patients
with elevated MCSF, elevated CRP, and their combination, respectively.
Conclusion In patients with chronic CAD, the prognostic value of MCSF is independent and complemen-
tary to that of CRP. MCSF is a particularly useful prognostic marker when CRP levels are low, but also
provides additional information concerning risk and time-course of events in patients with elevated CRP.
Introduction contribute to the development of acute coronary events in
patients with atherosclerosis.
Macrophage colony stimulating factor (MCSF) released by an Circulating CRP and MCSF levels have been associated
injured endothelium1 causes monocyte/macrophage acti- with death and recurrent ischaemic events in patients
vation,2 increases the macrophages’ cholesterol uptake,3 with acute coronary syndromes.17–21 In addition, CRP pre-
mediates monocyte-induced apoptosis of vascular smooth dicts outcome in patients with chronic coronary artery
muscle cells,4 and favours foam cell formation.5 MCSF- disease (CAD).17–19 However, it remains unclear whether
activated macrophages migrate to the fibrous cap of an CRP and MCSF have an independent or additive prognostic
atherosclerotic plaque,6 produce metalloproteinases7 and, value in patients with chronic CAD during a long-term
thus, may cause plaque destabilization.4,7 MCSF also pro- follow-up period. We, therefore, prospectively compared
motes platelet activation,8,9 tissue factor expression,10 and the prognostic significance of CRP to that of MCSF or their
the release of the procoagulant cytokine, interleukin-6 combination in patients with chronic CAD during a 6 year
(IL-6),11 which leads to C-reactive protein (CRP) pro- follow-up period.
duction.12 CRP has a direct proinflammatory effect on the
endothelium,13 mediates LDL uptake by macrophages,14
and may initiate the process of coagulation by inducing vas- Methods
cular smooth muscle cell apoptosis15 and monocyte Patient selection
expression of tissue factor.16 Through these actions, both
MCSF and CRP may promote thrombosis and, thus, Between 1994 and 1997, we prospectively examined 150
consecutive outpatients with: effort angina of 1 year duration,
documented exercise-induced ischaemia, 50% luminal diameter
* Corresponding author. Tel: þ30 210 338 1497; fax: þ30 210 777 0473. stenosis of one or more epicardial artery at angiography
E-mail address: ignoik@otenet.gr (performed within 5 + 2 months of screening), and written
& The European Society of Cardiology 2005. All rights reserved. For Permissions, please e-mail: journals.permissions@oupjournals.orgCRP and MCSF as complementary predictors in patients with chronic CAD 1619
informed consent. All patients underwent a treadmill exercise
test according to the Bruce protocol. The number of metabolic Table 1 Clinical characteristics of the study population
equivalents (Mets) achieved at ST-segment depression .0.1 mV, Age (years) 54 + 5 (32–68)
60 ms after the J point, was used to quantify the ischaemic Gender (M/F) 84/16 (84%)
threshold. Of these, 50 patients were excluded for the following Cholesterol (mmol/L) 5.9 + 0.99 (4.8–8.1)
reasons: 25 for an acute coronary event or coronary revasculariza- Triglycerides (mmol/L) 1.8 + 1.2 (0.8–5.7)
tion within the previous 6 months, seven for cerebral or periph- Body mass index (kg/m2) 28 + 4 (22–36)
eral vascular disease, 10 for diabetes mellitus, and eight for Smoking 60 (60%)
malignant or known inflammatory diseases. Patients with diabetes Hypertension 34 (34%)
were excluded becasue hyperglycaemia is a powerful activator of Parental CAD 50 (50%)
MCSF production.22 Thus, the final study cohort consisted of 100 CCS anginal class
patients (84 men, 16 women, mean age 54 + 5 years, range 1 40 (40%)
32–68) (Table 1 ). The study was approved by the local Research 2 47 (47%)
Ethics Committee. 3 13 (13%)
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
Diseased coronary arteries
Blood sampling 1 29 (30%)
2 36 (35%)
A fasting morning blood sample was taken from each patient for the
3 35 (35%)
measurement of plasma MCSF and CRP at inclusion in the study.
Previous MI 52 (52%)
Aliquots of plasma were stored at 2708C and analysis were per-
Drugsa
formed within a year of sampling.
b-Blockers 52 (52%)
Ca2þ-Blockers 51 (51%)
Laboratory assays Long-acting nitrates 26 (26%)
Plasma MCSF was measured by enzyme-linked immunoassay (‘human Lipid lowering 41 (41%)
MCSF Quantikinine’, sensitivity 20 pg/mL, R&D system, Minneapolis, Antiplatelets 100 (100%)
MN, USA). CRP was measured using particle-enhanced immuno- Diuretics 15 (15%)
nephelometry (N Latex CRP mono, Behring Diagnostics). The assay ACE-inhibitors 25 (25%)
detects a range of values between 0.175 and 1100 mg/L. The Monotherapyb 48 (48%)
intra-assay coefficients of variation were ,5% for both tests. GTN onlyc 16 (16%)
CCS, Canadian Cardiovascular Society; SA, chronic stable angina; ACE,
Follow-up angiotensin-converting enzyme; GTN, glyceryl nitrate. Lipid lowering:
40% statins and 1% fibrates.
Of the 100 patients, 95 (95%) completed a mean follow-up of a
Medical treatment at inclusion. The patients’ medication was similar
6 years (72+3 months). The remaining five were lost to follow at the beginning and at the end of the follow-up period (P . 0.05).
up and were considered as censored cases. Patients were followed b
Patients on only one of the three classes of antianginal drugs.
c
as outpatients at 6–12 month intervals, starting from the day of Patients on only sublingual GTN at inclusion.
blood sampling. The clinical endpoints considered in our analysis
were the occurrence of cardiac death, acute myocardial infarction
(MI), and hospital admission for unstable angina. Elective percuta-
neous coronary intervention (PCI) or coronary artery bypass graft- overall sample of 80 patients was required to detect a 10%
ing (CABG) was not considered, as they were not thought to increase of the event rate in the group of patients with high con-
represent an acute coronary event. Of the 100 patients (23%), 23 centrations of inflammatory markers using a two-tailed test with a
had a revascularization procedure (6 PCI and 17 CABG) during significance level of 5%, a power level of 90%, a drop-out rate of
the 6 year follow-up. Deaths of cardiac origin were confirmed by 5%, and a total follow-up period of 6 years. We prospectively stra-
death certificates and were verified by medical record review or tified patients into three groups, on the basis of the tertiles of the
primary care physician interview. Non-cardiac death was not measured biochemical indices (MCSF: ,449 pg/mL, n ¼ 34;
included in the analysis. MI was confirmed by ST-elevation or 450–813 pg/mL, n ¼ 34; 814 pg/mL, n ¼ 32; CRP: ,1.09 mg/L,
non-ST elevation changes on ECG and diagnostic increases in n ¼ 33; 1.1–2.4 mg/L, n ¼ 33; 2.5 mg/L, n ¼ 34),18 and cardiac
serum cardiac enzymes. Unstable angina was diagnosed by the event-free survival curves were constructed by Kaplan–Meier
occurrence of rest angina with non-ST segment elevation ischaemic analysis.18 Differences among curves were assessed using the log
ECG changes or with a positive troponin T test without concomi- rank test. Because the outcomes for the first and second tertile
tant increase in serum creatine kinase levels. of MCSF and CRP plasma levels were not statistically different,
they were combined for the final Kaplan–Meier and Cox pro-
portional hazard analyses.18,24 Univariate Cox proportional hazard
Statistical analysis
analysis24 was used to assess the predictors of cardiac events for
Inflammatory indices are presented as medians and interquartiles, the following covariates: age, gender, smoking status, hyperten-
as the data were not normally distributed. Stata 8.0 (Stata sion, hyperlipidaemia, parental CAD, previous (.6 months) MI,
Corporation, College Station, TX, USA) software was used. Differ- multivessel disease (two and three vessel disease), non-use of
ences within and among groups were analysed by Wilcoxon signed b-blockers, calcium blockers, nitrates, angiotensin-converting
rank test, Mann–Whitney U test, or analysis of variance (Kruskal– enzyme inhibitor, lipid-lowering medication, MCSF 814 pg/mL,
Wallis and Friedman test). Simple relations were assessed by and CRP 2.5 mg/L. The covariates with a P-value ,0.10 at
Spearman’s rank correlation. Multiple relations were assessed by univariate analysis were entered in the multivariable model.24
linear regression analysis after logarithmic transformation of MCSF Cox multivariable analysis with a stepwise selection method was
and CRP. Categorical variables were compared by contingency x2 used to estimate the final predictors of cardiac events.
test. Significance between models was calculated by the likelihood
Only the first event of the combined primary outcome variable ratio test. In multivariable analysis, a P , 0.05 was considered
(cardiac death, non-fatal MI, and unstable angina) was counted statistically significant.
as an endpoint. On the basis of an annual event rate of 2% for Interactions between MCSF, CRP, smoking, and multivessel
the primary end-point in patients with stable angina,23 an disease were also assessed in the multivariable analysis. To1620 I. Ikonomidis et al.
adjust the final model for other atherosclerotic risk factors and for patients with MCSF and CRP levels lower than the corre-
medication, a forced entry approach was used.24 The results of the sponding upper tertiles (Figure 1 ), suggesting a complemen-
Cox regression analysis are expressed as hazard ratios (HR) and tary role of MCSF and CRP.
corresponding 95% confidence intervals (CI). The appropriateness Kaplan–Meier analysis showed that patients with
of the risk assumption was examined by preparing the log (2log)
MCSF 814 pg/mL had a worse prognosis than patients
plots of the survival function. The final multivariable model was
validated by plotting the residuals against the fitted values (gener-
with MCSF ,814 pg/mL (log rank: 28.7, P , 0.0001,
alized Cox–Snell residuals) and testing the goodness of fit. Figure 2A ), and that patients with CRP 2.5 had a greater
Receiver operating characteristic curve analysis was also used to risk of adverse events than patients with CRP ,2.5 mg/L
compare the predictive value of MCSF vs. CRP levels. The curves (log rank: 12.64, P , 0.001, Figure 2B ). Patients with com-
were constructed by plotting sensitivity against (one-specificity). bined elevated CRP (2.5 mg/L) and MCSF (814 pg/mL)
The area under the curve (AUC) for MCSF or CRP was also calcu- had a higher risk of adverse events than patients with elev-
lated, in addition to the standard risk factors (age . 60 years, ated CRP levels alone (log rank: 8.35, P ¼ 0.003, Figure 3 ).
sex, smoking, hypelipidaemia, and hypertension) using logistic Thus, the prognostic information provided by MSCF may be
regression analysis. Wilcoxon’s signed rank test for dependent
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
complementary to that provided by CRP. Finally, survival
samples was used for comparisons between the unadjusted AUCs,
tables showed that the mean event-free time was 64
and one-sample test of proportions for comparisons between the
adjusted AUCs.
months (95% CI: 53–75) in patients with elevated CRP
alone, compared with 39 months (95% CI: 23–53) in patients
with elevated MCSF alone (log rank ¼ 4.67, P ¼ 0.03), and
Results 52 months (95% CI: 40–65) in patients with combined elev-
ated CRP and MCSF (log rank ¼ 8.35, P ¼ 0.003), indicating
Clinical characteristics (Table 1) a significant difference in the time-course of events pre-
All patients had an ejection fraction 55% at left ventricu- dicted by high levels of MCSF vs. those of CRP.
lography. None presented with signs of heart failure at
inclusion. Twenty-five patients (25%) presented myocardial
Univariate and multivariable predictors of cardiac
ischaemia at a workload of ,7 Mets during treadmill exer-
events
cise test. The proportion of patients receiving antianginal,
antiplatelet, or lipid-lowering drugs was similar at the Univariate Cox proportional hazard analysis showed that
beginning and at the end of the follow-up period (P . 0.05). male gender, smoking status, multivessel disease, and the
upper tertiles of MCSF or CRP levels were significant predic-
Association of MCSF and CRP with atherosclerotic tors of cardiac events (P , 0.05, Table 2 ) among traditional
risk factors risk factors and medications.
Patients with high MCSF (814 pg/mL) or CRP (2.5 mg/L)
By multivariable regression analysis, MCSF and CRP levels had a 5.9-fold (95% CI: 2.74–12.81) (P , 0.001) and 2.6-fold
were related to smoking (r ¼ 0.37 and r ¼ 0.28 respectively, (95% CI: 1.29–5.42) (P ¼ 0.008) higher risk of cardiac events,
P , 0.01) and MCSF additionally to multivessel disease compared with patients with low levels of MCSF or CRP,
(r ¼ 0.21, P ¼ 0.035), among age, gender, body mass respectively.
index, hypertension, hyperlipidaemia, family history of The multivariable Cox regression model included gender,
CAD and medications. MCSF levels were related to those of smoking, multivessel disease, and elevated MCSF and CRP,
CRP (r ¼ 0.47, P , 0.05). as only these variables were significant at P , 0.1 at univari-
ate analysis among traditional risk factors and medication.
Cardiac events Using the backwards stepwise selection method, smoking
Thirty-one patients presented cardiac events within the status was removed from the model. Then, using the
follow-up period. Of these, 6 died, 9 suffered a non-ST forward stepwise selection method, the non-significant cov-
segment elevation MI, and 16 had an episode of unstable ariates previously examined in univariate analysis were
angina. The median time from the start of follow-up to added. No new significant covariates were added in the mul-
the occurrence of cardiac events was 24 (12–68) months. tivariable model at the end of this analysis. At this point,
Of the 31 (23%), patients with cardiac events, 7 had a interactions between inflammatory indices and smoking or
revascularization procedure (two PCI and five CABG) at multivessel disease were examined in the multivariable
a median time of 13 (12–21) months of follow-up or at a model. A significant interaction emerged between the sub-
median time of 16 (9–51) months before the adverse groups of MCSF and CRP (P ¼ 0.03 for interaction),
event. Moreover, 16 of the 69 patients without events suggesting that the adjusted HR for cardiac events was
(23%) had a revascularization procedure (4 PCI and different between each subgroup of MCSF and CRP. The
12 CABG) at a median time of 14 (12–46) months of interaction term between MCSF and CRP remained signi-
follow-up. There was no difference in revascularization ficant after using the backward selection procedure
rate between patients with or without events (P ¼ 0.178). (P ¼ 0.02). However, multivessel disease became non-
significant (P . 0.05) and, thus, was removed from the
final model. The final multivariable model was therefore
Predictors of outcome constructed by male gender, MCSF levels 814 pg/mL, CRP
Kaplan–Meier life-time analysis for MCSF and CRP levels 2.5 mg/L, and their interaction term, as
Survival tables showed that the absolute risk (%) of cardiac summarized in Table 2.
events was 75% in patients with combined high CRP Cox proportional hazard analysis showed that patients
(.2.5 mg/L) and MCSF (.814 pg/mL), 59% in patients with with high MCSF (814 pg/mL) alone had a 13.17-fold (95%
high MCSF alone, 32% in patients high CRP alone, and 8% in CI: 4.14–42) higher risk of cardiac events compared withCRP and MCSF as complementary predictors in patients with chronic CAD 1621
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
Figure 1 Absolute risk (%) of cardiac events in patients with high or low
MCSF ( or ,814 pg/mL) and CRP ( or , 2.5 mg/L) levels derived by survi-
val tables (P , 0.05). N indicates the number of patients in each subgroup.
Figure 3 Kaplan–Meier survival curves for patients with combinations of
high or low MCSF ( or , 814 pg/mL) and CRP ( or , 2.5 mg/L) levels
(Log rank: 45.09, P ¼ 0.00001). The event-free rate in patients with com-
bined high CRP and MCSF is lower than that in patients with high CRP alone
(Log rank: 8.35, P ¼ 0.003). The survival curves for patients with combined
high MCSF and CRP and for patients with high MCSF alone crossed over
during follow-up indicating an interaction between the subgroups of CRP
and MCSF.
and CRP did not change: 14 (95% CI: 3.6–60) for MCSF and
6.29 (95% CI: 1.46–27) for CRP (P ¼ 0.001).
Comparison of the usefulness of MCSF and CRP
as predictors of cardiac events
Receiver operating characteristic curve analysis showed
a greater AUC for plasma MCSF levels than for CRP levels
[80% (95% CI: 70 to 91) vs. 61% (95% CI: 55 to 74),
z ¼ 8.419, P ¼ 0.00001 for the unadjusted AUC and
z ¼ 2.605; P ¼ 0.009 for the adjusted AUC], suggesting a
superior prognostic value of MCSF than of CRP for
predicting cardiac events (Figure 4 ).
Discussion
In this prospective study of patients with chronic CAD, we
demonstrate for the first time that CRP and MCSF plasma
levels are complementary predictors of adverse outcome,
among traditional atherosclerotic risk factors such as age,
male gender, smoking, hypertension, hyperlipidaemia,
Figure 2 Kaplan–Meier survival curves for patients with chronic CAD. (A ) family history of CAD, and multivessel disease, during a 6
The event-free rate in patients with MCSF 814 pg/mL (upper tertile of year follow-up period. The combination of increased CRP
MCSF levels) is significantly lower than that in patients with MCSF ,814 pg/
mL (Log rank: 28.7, P ¼ 0.00001). (B ) The event-free rate in patients with
and MCSF levels was associated with a greater absolute
CRP levels 2.5 mg/L (upper tertile of CRP levels) is significantly lower and relative risk of future events and a shorter event-free
than that in patients with CRP levels ,2.5 mg/L (Log rank: 12.64, period compared with elevated CRP levels alone in patients
P ¼ 0.0001). with chronic CAD.
patients with low MCSF and low CRP (P ¼ 0.0001), whereas
MCSF, CRP, and acute cardiac events
patients with high CRP levels (2.5 mg/L) alone had a In the present study, male gender, multivessel disease,
6.24-fold (95% CI: 1.74–22.42) higher risk of cardiac events smoking status, and the upper tertiles of MCSF and CRP
compared with patients with low CRP and low MCSF levels levels were significant predictors of prognosis by univariate
(P ¼ 0.005, Table 2 ). analysis. However, in multivariable analysis only male
After adjustment for other potential confounders such as gender and the upper tertiles of MCSF and CRP levels
age, hypertension, hyperlipidaemia, parental CAD, previous retained their prognostic value among age, smoking, hyper-
MI, and medications, the HR associated with elevated MCSF tension, hyperlipidaemia, family history of CAD, and1622 I. Ikonomidis et al.
Table 2 Cox proportional hazard analysis for the predictors of cardiac events
Variables Univariate analysis Multivariable analysis
P-values HR (95% CI) P-values HR (95% CI)
a
Inflammatory indices
MCSF 814 pg/mL 0.0001 5.9 (2.74–12.81) 0.0001 13.17 (4.1–42)c
CRP 2.5 mg/L 0.008 2.6 (1.29–5.42) 0.005 6.24 (1.74–22.42)
MCSF CRPb 0.02
Coronary risk factors
Age 0.80 1.005 (0.97–1.05)
Male gender 0.06 6.6 (0.9–49) 0.03 8.34 (1.13–61.4)
Smoking 0.03 2.9 (1.09–7.5)
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
Hyperlipidemia 0.49 1.32 (0.6–2.9)
Hypertension 0.76 1.11 (0.5–1.2)
Parental CAD 0.17 1.7 (0.8–3.4)
Previous MI 0.29 2.02 (0.5–7.5)
Multivessel disease 0.05 2.5 (0.98–6.2)
Medicationa
No use of nitrates 0.17 1.6 (0.8–3.4)
No use of b-blockers 0.22 1.56 (0.7–3.1)
No use of Ca-blockers 0.60 0.83 (0.4–1.6)
No use of lipid-lowering drugs 0.38 1.4 (0.7–3.8)
No use of ACE-Inhibitors 0.23 1.54 (0.76–3.1)
All covariates except age are categorical. HR, hazard ratio; CI, confidence interval.
a
The cut-off values for MCSF and CRP correspond to the upper tertiles of their levels that are compared with the combination of the first and second tertile
of their levels by multivariable analysis.18
b
MCSF CRP indicates the interaction term between CRP 2.5 mg/L and MCSF 814 pg/mL entered in the multivariable model.
c
HR for patients with high MCSF or CRP alone vs. patients with combined low MCSF and CRP; multivessel: two and three vessel; ACE, angiotensin converting
enzyme; NS, not significant (P . 0.05).
cardiac events in patients with stable and unstable angina
during a 2 year follow-up period. Saito et al. 21 showed the
prognostic significance of increased MCSF levels for future
cardiac events in a mixed population of patients with
stable and unstable angina during 14 months of follow-up.
However, in their study, increased MCSF levels on admission
and incidence of future events were associated with the
diagnosis of unstable angina at inclusion, indicating that
the results of their study were mainly driven by the subgroup
of patients with unstable angina. Conversely, in the present
study, we showed that high MCSF levels predicted an adverse
outcome in patients with chronic CAD.
Moreover, in this study, patients with CRP levels ,2.5 mg/L,
who are considered at low risk of cardiac events,17 had a
13-fold risk of future events when their MCSF levels were
.814 pg/mL compared with when they were ,814 pg/mL.
Figure 4 Receiver operating curves for the prediction of cardiac events
Thus, in the presence of low CRP levels, MCSF becomes par-
showed a greater area under the curve for MCSF levels than for CRP levels ticularly useful in distinguishing patients at a substantially
[80% (95% CI: 70–91) vs. 61% (95% CI: 50–74), P ¼ 0.00001]. lower or higher risk of cardiac events.
Furthermore, patients with the combined elevation of
MCSF and CRP levels had a higher absolute and relative
multivessel disease. Because atherosclerotic risk factors risk of future events when compared with patients with
promote cytokine production1–4,22,25,26 and, consequently, elevated levels of a single inflammatory factor or to patients
CRP release,12,27,28 these findings suggest that inflammatory with low levels of both inflammatory indices. Thus, high
indices may serve as a surrogate marker of the cumulative MCSF provided additional prognostic information in patients
effect of traditional risk factors on the prognosis of patients with high CRP levels and vice versa.
with chronic CAD. However, it is possible that MCSF and CRP Our results suggest that MCSF and CRP levels may be
may directly contribute to plaque rupture, thrombo- synergistic and complementary predictors of adverse
sis,4,6,7,10,15,16 and thus, to the genesis of acute cardiac outcome in patients with chronic CAD. Indeed, both CRP
events.17–21 Our findings are in agreement with those of and MCSF induce apoptosis of vascular smooth muscle
others18 who found that CRP levels independently predicted cells4,15 and tissue factor expression10,16 in atheroscleroticCRP and MCSF as complementary predictors in patients with chronic CAD 1623
lesions and, consequently, may cause a synergic increase in Thus, increased MCSF may signal patients with more aggres-
plaque thrombogenicity and instability. sive CAD and, therefore, with a higher risk for acute ischae-
Survival tables showed that the mean event-free time was mic events at an earlier follow-up time compared with
39 months in patients with elevated MCSF when compared patients with elevated CRP alone. Our results are in line with
with 64 months in patients with elevated CRP and 52 those of other investigators who found that MCSF but not
months in patients with combined elevation of CRP and CRP was an independent predictor of in-hospital and short-
MCSF. This finding suggests that increased MCSF levels term outcome in patients with unstable angina.20 For these
alone or combined with elevated CRP levels may predict reasons, MCSF levels may be considered a reliable alternative
the incidence of future coronary events occurring at an marker of outcome in patients with chronic stable CAD.
earlier time when compared with elevated CRP levels alone.
Atheroslerotic risk factors1,22,25 or infections2 may trigger Study limitations
increased MCSF production, mainly by endothelial cells,2 in
the initial process of atherosclerosis.31 Moreover, MCSF-acti- The following limitations should be acknowledged.
Inflammatory indices were measured in peripheral blood.
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
vated macrophages produce further amounts of MCSF that
can enter the systemic circulation.2 MCSF promotes the This does not allow firm conclusions on the release of
release of IL-6,11 which drives CRP production.12 Thus, these factors within the coronary circulation. Medications
MCSF production by endothelial cells and resident macro- may affect plasma levels of the measured inflammatory
phages at atherosclerotic lesions may precede the cyto- indices.19,27 However, the study subgroups did not differ in
kine-mediated release of CRP by hepatocytes.12 As a antianginal, antiplatelet, or lipid lowering treatment at
result, MCSF may promote apoptosis of smooth muscle enrolment or at the end of the follow-up period. Thus, any
cells,4 release of metalloproteinases,7 platelet acti- possible influence of medications on cytokine plasma
vation,8,9 tissue factor expression10 and, consequently, levels was evenly distributed within the study subgroups.
plaque instability at an earlier time than CRP. Although The sample size of our study population is relatively small,
CRP and MCSF have been detected in human atherosclerotic though adequately powered (.80%) for the differences
lesions,28,29 CRP is not present in the normal vessel wall,2,28 reported between the various study subgroups. The study
suggesting a role of CRP in the advanced phases of athero- population comprised a high proportion of smokers and
genesis. Recent evidence suggests that a balance of pro- male patients and thus our results should be interpreted in
thrombotic15 and anti-thrombotic effects of CRP30,31 on view of this. Finally, the number of patients using b-blockers
the vessel wall may be important in the development of and lipid lowering drugs was relatively low; thus, our study
adverse cardiac events.32 Increased MCSF production may population may not be representative of the current
disrupt this balance towards increased apoptosis and patient population, in whom the current widespread use of
plaque thrombogenicity at an early stage of atherosclerosis b-blockers and statins may influence the prognostic value
by enhancing the apoptotic effects of CRP.4,15 Alternatively, of the measured inflammatory markers.
MCSF may facilitate the structural modification of the
native, pentameric, CRP to monomeric subunits which are Conclusion
required for proinflammatory actions on endothelial
cells.33 Through this action MCSF may enhance and acceler- In this prospective study of patients with chronic CAD, high
ate the proinflammatory effects of circulating CRP on vascu- CRP, and MCSF levels were independent and complementary
lar tissues so that these are augmented and become evident predictors of adverse outcome during a 6 year follow-up
at an earlier time than they would have become in the pre- period. Furthermore, the prognostic value of elevated
sence of elevated CRP levels alone. The above pathophysio- MCSF became evident at an earlier time during follow-up
logical mechanisms may explain the greater risk of future than that of elevated CRP. Our findings are clinically rel-
events and the shorter event-free time in patients with com- evant, as MCSF levels provide additional information on
bined high CRP and MCSF levels compared with patients with the risk and time course of adverse events in patients with
elevated CRP levels alone observed in the present study. elevated CRP levels, but are also useful to further stratify
Thus, the measurement of MCSF provides additional prog- the risk of patients who have low levels of CRP.
nostic information on the risk and time-course of events
compared with the measurement of CRP in patients with Acknowledgement
chronic CAD.
This work was supported by Hammersmith Hospital Grant no. RC/
In this study, a head-to-head comparison between the pre-
259.
dictive value of MCSF and of CRP showed that the upper
tertile of MCSF levels was associated with an approximately
two-fold higher absolute (59 vs. 32%) and relative (13.17 vs.
References
6.24) risk of events, as well as with a shorter event-free 1. Rajavashisth TB, Andalibi A, Territo MC, Navab M, Fogelman AM, Lusis AJ.
time (39 vs. 64 months) than the upper tertile of CRP Induction of endothelial cell expression of granulocyte and macrophage
colony stimulating factors by modified low-density lipoproteins. Nature
levels. In addition, by receiver operating curve analysis,
(Lond) 1990;344:254–257.
the predictive value for acute cardiac events of MCSF 2. Roth P, Stanley ER. The biology of CSF-1 and its receptor. Current Topics
levels was higher than that of CRP levels (area under the in Microbiology and Immunology. Berlin: Sonneg-Verlag, 1992, vol. 181,
curve 80 vs. 61%). MCSF is directly related to myocardial p.141–158.
ischaemia during effort and daily life activities,9,27 as well 3. Ishibashi S, Inaba T, Shimano H, Harda K, Inoue I, Mokuno H, Mori N,
Gotoda T, Takaku F, Yamada N. Monocyte colony stimulating factor
as to the angiographic extent of CAD in patients with enhances cholesterol uptake and degradation of acetylated low density
chronic stable angina.21,27 Conversely, CRP is modestly lipoproteins and cholesterol esterification in human monocyte derived
related to the anatomic extent of CAD in these patients.34 macrophages. J Biol Chem 1990;265:14109–14117.1624 I. Ikonomidis et al.
4. Seshiah PN, Kereiakes DJ, Vasudevan SS, Lopes N, Su BY, Flavahan NA, 19. Crea F, Monaco C, Lanza GA, Maggi E, Ginnetti F, Cianflone D, Niccoli G,
Goldschmidt-Clermont PJ. Activated monocytes induce smooth muscle Cook T, Bellomo G, Kjekshus J. Inflammatory predictors of mortality in
cell death. The role of macrophage colony stimulating factor and cell the Scandinavian Simvastatin Survival Study. Clin Cardiol
contact. Circulation 2002;105:174–178. 2002;25:461–466.
5. Huh HY, Pearce SF, Yesner LM, Schindler JL, Silverstein RL. Regulated 20. Rallidis LS, Zolindaki MG, Manioudaki HS, Laoutaris NP, Velissaridou AH,
expression of CD36 during monocyte to macrophage differentiation: Papasteriadis EG. Prognostic value of C-reactive protein fibrinogen, inter-
potential role of CD36 in foam cell formation. Blood 1996;87: 2020–2028. leukin 6 and macrophage colony stimulating factor in severe unstable
6. Moreno PR, Falk E, Palacios IF, Newell JB, Fuster V, Fallon JT. Macrophage angina. Clin Cardiol 2002;25:505–510.
infiltration in acute coronary syndromes: implications for plaque rupture. 21. Saitoh T, Kishiba H, Tsukada Y, Fukuma Y, Sano J, Yasutake M, Fukuma N,
Circulation 1994;90:775–778. Kusama Y. Clinical significance of increased plasma concentration of MCSF
7. Uzui H, Harpf A, Liu M, Doherty TM, Shukla A, Chai NN, Tripathi PV, in patients with angina pectoris. J Am Coll Cardiol 2000;35:655–665.
Jovinge S, Wilkin DJ, Asotra K, Shah PK, Rajavashisth TB. Increased
22. Saini A, Liu YJ, Cohen DJ, Ooi BS. Hyperglycemia augments macrophage
expression of membrane type 3-matrix metalloproteinase in human
growth response to colony stimulating factor-1. Metabolism
atherosclerotic plaque: role of activated macrophages and inflammatory
1996;45:1125–1129.
cytokines. Circulation 2002;106:3024–3030.
23. Juul-Mollers S, Edvardsson N, Jahnmatz B, Rosen A, Sorensen S, Omblus
8. Orlandi M, Bartolini G, Minghetti L, Luchetti S, Giulucci B, Chiricolo M,
R. Double-blind trial of aspirin in primary prevention of myocardial
Downloaded from https://academic.oup.com/eurheartj/article/26/16/1618/632755 by guest on 01 February 2022
Tomasi V. Prostagladin and thromboxane biosynthesis in isolated platelet
infarction in patients with stable chronic angina pectoris. The Swedish
free monocytes. III. The induction of cycloxygenase by colony stimulating
Angina Pectoris Aspirin Trial (SAPAT) Group. Lancet 1992;340:1421–1425.
factor. Prostaglandins Leukot Essent Fatty Acids 1989;36:101–106.
9. Ikonomidis I, Andreotti F, Nihoyannopoulos P. Reduction of daily-life 24. Collet D. Modelling Survival Data in Medical Research. 1st ed. London:
ischaemia by aspirin in patients with angina:underlying link between Chapman and Hall; 1994: p53–106.
thromboxane A2 and macrophage colony stimulating factor. Heart 25. Rose RM, Kobzik L, Dushay K, Wolfthal S, Hondalus M, Metzger M,
2004;90:389–393. Stoudemire J, Brain JD, Garnick M, O’ Donnell C. Characterisation of
10. Lyberg T, Stanley ER, Prydz H. Colony stimulating factor-1 induces throm- colony stimulating factor activity in the human respiratory tract.
boplastin activity in murine macrophages and human monocytes. J Cell Comparison of healthy smokers and non-smokers. Am Rev Respir Dis
Physiol 1987;132:367–370. 1992;145:394–399.
11. Bauer J, Ganter U, Geiger T, Jacobshagen U., Gilberto G. Regulation of 26. Tappia PS, Troughton KL, Langley-Evans SC, Grimble RF. Cigarette
interleukin 6 expression in cultured human blood monocytes and mono- smoking influences cytokine production and antioxidant defences. Clin
cyte derived macrophages. Blood 1988;72:1134–1140. Sci (Lond) 1995;88:485–489.
12. Bataille R, Klein B. C-reactive protein levels as a direct indicator of inter- 27. Ikonomidis I, Andreotti F, Economou E, Stefanadis C, Toutouzas P,
leukin 6 levels in humans in vivo. Arthritis Rheum 1992;35: 982–984. Nihoyannopoulos P. Increased proinflammatory cytokines in patients
13. Pasceri V, Willerson JT, Yeh ET. Direct proinflammatory effect of with chronic stable angina and their reduction by aspirin. Circulation
C-reactive protein on human endothelial cells. Circulation 2000; 1999;100:793–798.
102:2165–2168. 28. Reynolds GD, Vance RP. C-Reactive protein immunohistochemical localis-
14. Zwaka TP, Hombach V, Torzewski J. C-reactive protein-mediated low ation in normal and atherosclerotic human aortas. Arch Pathol Lab Med
density lipoprotein uptake by macrophages implications for atherosclero- 1987;111:265–269.
sis. Circulation 2001;111:1194–1197. 29. Rosenfeld ME, Yla-Herttualla S, Lipton BA, Ord VA, Witztum JL, Steinberg
15. Blaschke F, Bruemmer D, Yin F, Takata Y, Wang W, Fishbein MC, Okura T, D. Macrophage colony stimulating factor mRNA and protein in athero-
Higaki J, Graf K, Fleck E, Hsueh WA, Law RE. C-reactive protein induces sclerotic lesions of rabbits and humans. Am J Pathol 1992;140:291–300.
apoptosis in human coronary vascular smooth muscle cells. Circulation 30. Wang CH, Li SH, Weisel RD, Fedak PW, Dumont AS, Szmitko P, Li RK,
2004;110:579–587. Mickle DA, Verma S. C-reactive protein upregulates angiotensin I recep-
16. Nakagomi A, Freedman SB, Geczy CL. Interferon gamma and lipo-
tors in vascular smooth muscle cells. Circulation 2003;107:1783–1790.
polysaccharide potentiate monocyte tissue factor induction by C-reactive
31. Li SH, Szmitko PE, Weisel RD, Wang CH, Fedak PW, Li RK, Mickle DA,
protein: relationship with age sex and hormone replacement treatment.
Verma S. C-reactive protein upregulates complement-inhibitory factors
Circulation 2000;101:1785–1791.
in endothelial cells. Circulation 2004;109:834–836.
17. Haverkate F, Thompson, SC, Pyke SD, Gallimore JR, Pepys MB. For the
32. Verma S, Szmitko PE, Yeh ETH. C-Reactive protein: structure affects func-
European Concerted action on Thrombosis and disabilities Angina
tion. Circulation 2004;109:1914–1917.
Pectoris Study group. Production of C-reactive protein and risk for coron-
ary events in stable and unstable angina. Lancet 1997;349:462–466. 33. Khreiss T, József L, Potempa LA, Filep JG. Conformational rearrangement
18. Zebrack JS, Anderson JL, Maycock CA, Horne BD, Bair TL, Muhlestein JB. in C-reactive protein is required for proinflammatory actions on human
Usefulness of high sensitivity C-reactive protein in predicting long term endothelial cells. Circulation 2004;109:2016–2022.
risk of death or acute myocardial infarction in patients with unstable 34. Zebrack JS, Muhlstein JB, Horne BD, Anderson JL. C-Reactive protein and
or stable angina pectoris or acute myocardial infarction. Am J Cardiol angiographic coronary artery disease: independent and additive predic-
2002;89:145–149. tors of risk in patients with angina. J Am Coll Cardiol 2002;39:632–637.You can also read
