- Email: [email protected]
Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients
Accepted Manuscript Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients Loïc Panh, Jean Bernard Ruidavets, Hervé Rousseau, Antoine Petermann, Vanina Bongard, Emilie Bérard, Dorota Taraszkiewicz, Olivier Lairez, Michel Galinier, Didier Carrié, Jean Ferrières PII:
S0021-9150(17)30132-6
DOI:
10.1016/j.atherosclerosis.2017.03.030
Reference:
ATH 15005
To appear in:
Atherosclerosis
Received Date: 1 December 2016 Revised Date:
15 March 2017
Accepted Date: 22 March 2017
Please cite this article as: Panh L, Ruidavets JB, Rousseau H, Petermann A, Bongard V, Bérard E, Taraszkiewicz D, Lairez O, Galinier M, Carrié D, Ferrières J, Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients, Atherosclerosis (2017), doi: 10.1016/j.atherosclerosis.2017.03.030. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients Loïc Panha, Jean Bernard Ruidavetsb, Hervé Rousseauc, Antoine Petermannc, Vanina Bongardb, Emilie Bérardb, Dorota Taraszkiewicza, Olivier Laireza, Michel Galiniera, Didier
a
RI PT
Carriéa, Jean Ferrièresa.
Department of Cardiology, Toulouse-Rangueil University Hospital (CHU), TSA 50032,
Toulouse University School of Medicine, 1 Avenue du Professeur Jean Poulhes 31059
b
SC
Toulouse Cedex 9, France.
Department of Epidemiology, Unité de Soutien Méthodologique à la Recherche (USMR),
France. c
M AN U
Toulouse University Hospital (CHU), 37 Allées Jules Guesde, 31073 Toulouse Cedex 7,
Department of Radiology, Toulouse-Rangueil University Hospital (CHU), TSA 50032,
Toulouse University School of Medicine, 1 Avenue du Professeur Jean Poulhes 31059
TE D
Toulouse Cedex 9, France.
Corresponding author: CHU Toulouse Rangueil, Department of Cardiology, 1, avenue du
EP
Pr Jean Poulhes, 31059 Toulouse Cedex 9, France.
AC C
Email: [email protected] (J. Ferrières)
1
ACCEPTED MANUSCRIPT
ABSTRACT Background and aims: A high level of serum alkaline phosphatase (ALP) is associated with an increased risk of mortality and myocardial infarction. ALP hydrolyses inorganic
RI PT
pyrophosphate, which is a strong inhibitor of calcium phosphate deposition. The aim of this study was to determine whether ALP is associated with the coronary artery calcium score (CACS).
SC
Methods: We examined the association of CACS, assessed by computed tomography scanning, and ALP, in 500 patients consecutively recruited, free of cardiovascular disease.
M AN U
The CACS were categorized into two groups: no calcification (CACS=0) (n=187) and with calcification (CACS>0) (n=313). ALP activity was divided into three tertile groups: low ALP level (<55 IU/L), intermediate (55 to 66 IU/L) and high ALP level (>66 IU/L). Results: The mean age was 60.9 ± 10.8 years, 49.6% of the patients were women. ALP ranged
TE D
from 22 to 164 IU/L (mean 62.6 IU/L, SD 19.3). In univariate analysis, traditional cardiovascular risk factors, statin use (p=0.001), and ALP (p=0.001) were significantly associated withCACS. After adjusting for cardiovascular risk factors, only age (p=0.001) and
EP
sex (p=0.001) were independently associated with CACS. Compared to the tertile group with
AC C
low levels of ALP, the intermediate tertile group [OR 2.11, 95%CI (1.12;3.96), p=0.02], as well as the high tertile group [OR 3.89, 95% CI (2.01;7.54), p=0.001)], was independently associated with CACS.
Conclusions: In patients free of cardiovascular disease, high ALP levels are positively and independently associated with coronary artery calcification. The metabolic pathway of ALP and inorganic pyrophosphate could be a target for new therapies against vascular calcification. KEYWORDS Coronary artery disease; Calcification; Alkaline phosphatase; Statin.
2
ACCEPTED MANUSCRIPT 1- INTRODUCTION Intimal atherosclerotic plaque calcification is considered as the result of an inflammatory process initiated by infiltrating oxidized lipids. Coronary artery calcification is generally assessed by non-contrast computed tomography and coronary artery calcium score (CACS)
RI PT
measurement [1]. CACS is a reflection of the atheromatous disease and a potent predictor of cardiovascular events and mortality [2]. It is due to several active and passive mechanisms of ectopic mineralization leading to hydroxyapatite crystal formation on the layers of the arterial
SC
wall. Active mechanisms of calcification involve osteoblastic differentiated cells, and lead to
M AN U
an organized calcified bone-like matrix [3]. These cellular processes are highly regulated, under the influence of inflammatory cytokines, oxidant stress, or mineral metabolism [4]. However, the majority of vascular mineralization is due to passive phenomena, subject to the neutral charge theory of calcification [5], leading to unstructured and amorphous calcified areas. Contrary to previous thinking, passive calcium phosphate deposition is also regulated.
TE D
One of the most powerful passive calcium phosphate deposition inhibitors is inorganic pyrophosphate (PPi) [6]. The biochemical effects of PPi are an inhibition of calcium and phosphate aggregation [7], hydroxyapatite crystal growth [8] and aggregation [9]. In vivo, PPi
EP
is hydrolyzed into inorganic phosphate by serum alkaline phosphatase (ALP). Thus, an
AC C
increased ALP activity can induce an imbalance between inorganic phosphate and pyrophosphate, promoting ectopic calcification. Thus, a pathological increase in ALP activity is linked to extensive vascular calcification, leading to premature atherosclerosis and cardiovascular events, such as Hutchinson-Gilford progeria syndrome or the Generalized Arterial Calcification of Infancy syndrome [10,11]. Besides these rare diseases, high ALP levels are associated with an increased risk of cardiovascular events and mortality in the general population and in secondary cardiovascular prevention [12,13] and higher coronary
3
ACCEPTED MANUSCRIPT artery calcifications in hemodialysis patients [14]. However, data on a possible role of ALP in coronary artery calcification in a primary cardiovascular prevention population are lacking. The aim of this study was to assess the potential association between ALP and CACS in
RI PT
patients free of any cardiovascular disease.
2- PATIENTS AND METHODS 2-1- Study population
SC
Our study sample included 555 patients ranging in age from 25 to 83 years, consecutively
M AN U
enrolled between September 2014 and October 2016 in the department of preventive cardiology (Toulouse University Hospital). This was a primary prevention population, selfreferred or referred by their primary-care physician or cardiologist for cardiovascular risk stratification. The inclusion criterion was all consecutive patients over 18 years old. The exclusion criteria concerned secondary prevention patients, defined as having undergone
TE D
acute coronary syndrome, stable angina, coronary revascularization (surgical or percutaneous intervention), or stroke. Patients with acute liver, gallbladder or other gastrointestinal diseases, active infectious diseases, potentially responsible for an increase of ALP activity
EP
were also excluded. Authorization to use these data was obtained in accordance with the
AC C
French law (Commission Nationale de l’Informatique et Libertés - CNIL).
2-2- Cardiovascular risk factors and laboratory data Subjects were given a questionnaire to assess cardiovascular risk factors during a medical consultation. The cardiovascular risk factors included smoking status, diabetes, hypertension, dyslipidemia, age, sex, and family history of premature cardiovascular event. Data concerning personal medical history, lifestyle habits and drug intake were recorded. Blood pressure was measured after at least 5 minutes of seated rest, with a standard automated
4
ACCEPTED MANUSCRIPT sphygmomanometer. Blood pressure was measured three times on both arms at 5 minute intervals. The mean of the measurements was used for statistical analysis. A standardized procedure was used to measure height and weight. Body mass index was calculated as the weight divided by the height squared (kg/m²).
RI PT
A blood sample was taken after over 12 hours of overnight fasting at the central laboratory of Toulouse University Hospital. Serum total cholesterol and triglycerides were measured by enzymatic assays (Cobas 8000, Roche Diagnostics, Germany). High density lipoprotein
SC
cholesterol (HDL-C) was measured after sodium phosphotungstate-magnesium chloride precipitation of apolipoprotein B-containing proteins. Low density lipoprotein cholesterol
M AN U
(LDL-C) was determined using the Friedewald formula when triglycerides were < 4.6 mmol/l. Glucose concentration were measured using a conventional enzymatic method based on hexokinase-glucose-6-phosphate dehydrogenase.
Serum
creatinine
measurement
was
performed with an enzymatic method with calibration certified by isotopic dilution mass
TE D
spectroscopy. Glomerular filtration rate (GFR) was estimated using the serum creatinine level measurement and the CKD-EPI equation calculation [15]. ALP activity was measured by a spectrophotometric method, using p-nitrophenylphosphate as
EP
a substrate, in an automated analyzer (Cobas 8000, Roche Diagnostics, Germany). Interassay variability ranged between 0.9 and 2.4%, and the normal reference ranges in our laboratory
AC C
were 35 to 104 IU/L in women and 40 to 129 IU/L in men. Hypertension was defined as mean blood pressure >140/90 mmHg at rest, or personal history of hypertension or treatment for hypertension. Smoking was defined as current smoking or no smoking. Patients who had stopped smoking at least one month before the investigation were counted as non-smokers. Family history of premature cardiovascular disease was defined as coronary disease or cerebrovascular disease before the age of 55 in men and 65 in women, for first-degree family members. Dyslipidemia was defined in this moderate cardiovascular risk
5
ACCEPTED MANUSCRIPT population as a total plasma cholesterol ≥ 6.5 mmol/l, or LDL-C ≥ 4 mmol/l, or statin use, or other lipid lowering treatment, or triglycerides ≥ 2.3 mmol/l. Diabetes was defined as plasma glucose level ≥ 7 mmol/l or history of medical treatment for diabetes. Cardiovascular disease
and femoral arteries assessed by ultrasound imaging.
RI PT
was excluded by a negative stress test and the lack of atheroma in the carotid arteries, aorta
2-3- Cardiac multislice computed tomography acquisition and interpretation
SC
All computed tomography scans were performed with a dual source CT-system (Somaton Définition; Siemens Healthcare, Forchheim, Germany). Scan parameters for the calcium score
M AN U
protocol were as follows: 3mm slice thickness, collimation 6x3, 120kV tube voltage, 68mAs, field of view 250mm during a breath-hold maneuver. Image reconstruction was performed at 70% of R-R interval. Approximately 50 sections were obtained at the level of the carina and proceeding caudally to the level of the diaphragm. A calcified lesion was defined as more
TE D
than 3 contiguous pixels with a peak of attenuation of at least 130 Hounsfield units. The total coronary artery calcium score (CACS) was determined according to the Agatston method, by summing individual lesion scores from each of 4 anatomic sites (left main, left anterior
EP
descending, circumflex, and right coronary artery). CACS were stratified in two groups, the
AC C
calcified patients (CACS>0) and non-calcified patients (CACS=0).
2-4- Statistical analysis
Data are presented as means and standard deviations for quantitative variables and percentages for categorical variables. ALP levels were stratified in three balanced tertiles groups: <55 IU/L, 55 to 66 IU/L, >66 IU/L. The Chi2 test was used to compare the distribution of qualitative variables between the two CACS groups. Mean values of quantitative variables were compared by one-way analysis of variance. Shapiro-Wilk’s and
6
ACCEPTED MANUSCRIPT Levene’s tests were used to test the normality of the distribution of residuals and the homogeneity of variances, respectively. When basic assumptions were not satisfied, a logarithmic transformation of the variables was carried out. Multivariate analyses were performed using multiple logistic regression, to test the association between ALP
RI PT
concentrations and CACS groups (CACS=0 taken as reference group) and to calculate odds ratios (OR) and 95% confidence intervals. A backward selection method of the variables was used for all the models. All selected variables were entered in a full model and then a step-by
remove non-significant contributory variables.
SC
step-elimination was performed. A likelihood ratio test was made between nested models to
M AN U
First-order interactions between ALP and all other variables remaining in the final model were tested systematically by adding an interaction term. A multiplicative product of ALP with each other variable was introduced into the model taking into account the nature of the variable studied; categorical, ordinal or continuous. Statistical significance was tested using
TE D
the likelihood ratio test between nested models with and without the term of interaction. Logistic regression analyses with polynomial terms (quadratic and cubic) were performed to examine for possible non-linear relationships between ALP, used as a continuous variable,
EP
and the binary variable CACS. The relation between ALP and CACS was considered as a linear response function because the second and third order terms were not significant.
AC C
All the statistical analyses were conducted using SAS (SAS Institute, version 9.4, Cary, NC) and Stata (Stata, release 14.0, College Station, TX). We considered p<0.05 as significant.
3- RESULTS 3-1- Characteristics of the study population and associations to CACS After excluding 55 patients according to the exclusion criteria, the study population consisted of 500 patients who had ALP measured as baseline and were eligible for statistical analysis.
7
ACCEPTED MANUSCRIPT The mean age was 60.9 ± 10.8 years, hypertension affected 28.8% of patients, 6.9% had diabetes, 70.7% had dyslipidemia, 37.2% underwent statin therapy. The mean GFR was 81.6 ± 13.9 mL/min/1.73 m2, and only 5.4% had a GFR below 60 mL/min/1.73 m2. ALP levels ranged from 22 to 164 IU/L (median 60 IU/L, mean 62.6 IU/L, SD 19.3). ALP levels outside
RI PT
of the normal range were observed in 39 patients (7.8%), with 26 patients (5.2 %) below and 13 patients (2.6%) above the reference value. The mean CACS was 163 ± 484. The proportion of CACS = 0 was 37.4% (n=187) and 62.6% (n=313) had CACS > 0. The
SC
characteristics of each ALP group and the distribution of CACS are summarized in Table 1. In univariate analysis, traditional cardiovascular risk factors such as age (p=0.001), sex
M AN U
(p=0.001), hypertension (p=0.001), diabetes (p=0.004), and GFR (p=0.003) were significantly associated with CACS. There was no significant association between CACS and dyslipidemia, HDL cholesterol or LDL cholesterol. Statin use was associated with CACS (p=0.001). Before adjustment, ALP was significantly associated withCACS (p=0.001) (Table
TE D
2).
3-2- Adjusted association with CACS and interaction between ALP and statin treatment
EP
A multiple logistic regression was performed on cardiovascular risk factors (age, sex, diabetes, hypertension, smoking, and GFR), statin treatment, and ALP activity. In the full
AC C
sample, only age (p=0.001) and sex (p= 0.001) and ALP were independently associated withCACS. Compared to the group with low levels of ALP (<55 IU/L), intermediate levels of ALP (55 IU/L to 66 IU/L) were significantly associated withCACS [OR 2.11, 95% CI (1.12;3.96), p=0.02], and high levels of ALP (>66 IU/L) were associated with CACS [OR 3.89, 95% CI (2.01;7.54), p=0.001)]. There was a linear association (p for trend 0.001) between ALP and CACS (Table 3). Otherwise, results were similar using continuous ALP levels analysis [for one ALP unit (IU/L) increase OR=1.03, 95% CI (1.02-1.05), p<0.001].
8
ACCEPTED MANUSCRIPT Age was the strongest independent predictor of CACS. However, there was no interaction between age and ALP levels (p=0.35). For one ALP unit (IU/L) increase, adjusted OR were similar in the youngest patients [age below the median; OR=1.03, 95%CI (1.01; 1.05), p=0.007] as well as the oldest patients [age above the median; OR=1.05, 95% CI (1.01; 1.08),
RI PT
p=0.009].
We found a significant interaction between ALP and statin use (p=0.004). Thus, statin treatment had an opposite effect on ALP according to the calcified or non-calcified patient’s
SC
status. In non-calcified patients (CACS=0), ALP activity was significantly higher with statin treatment [56.8 IU/L vs. 65.2IU/L (p=0.005)]. Conversely in calcified patients (CACS>0),
M AN U
ALP activity was significantly lower with statin treatment [67.0IU/L vs. 61.8IU/L (p=0.008)]. In the same multivariate analysis, according to statin treatment, the intermediate and high ALP levels were associated with CACS in patients without statin, but not in patients with statin. The other results were similar as the full sample, and only age and sex were
4- DISCUSSION
TE D
independently associated with CACS (Table 4).
EP
The major finding of this study was that, in a primary cardiovascular prevention population, a
AC C
high level of ALP activity was positively and independently associated with the presence of coronary artery calcification. Thus, ALP levels higher than 55 IU/L and particularly higher than 66 IU/L were potent predictors of coronary artery calcification. Age was the most powerful predictor of CACS, but did not directly interfere with the relationship between ALP and CACS which was significant regardless of age. The mechanism behind the relationship between ALP and artery calcification is based on previous observations on the inhibitory effect of PPi on vascular calcification and its
9
ACCEPTED MANUSCRIPT hydrolysis by ALP [16]. A physiological concentration of PPi can completely inhibit calcium and phosphate deposition under a physiological concentration of calcium and phosphate [17]. Epidemiological studies showed that high ALP levels are linked to cardiovascular events both in the general population and in secondary cardiovascular prevention patients [12,13]. Our
RI PT
results could provide a pathophysiological explanation for the increase in coronary artery calcification, which is a potent marker of cardiovascular atheroma and cardiovascular events. Interestingly, more than 90% of patients included in both these studies as well as our study
SC
had ALP levels in the range usually considered as normal.
This link between ALP and CACS is subject to several confounding factors, such as GFR,
M AN U
inflammatory status, or liver enzymes which are known to increase the cardiovascular risk. To try to reduce the influence of these potential confounding factors, we excluded active gastrointestinal, liver, or infectious diseases, and no woman was pregnant. A multiple adjustment was done on these factors, and intermediate as well as high ALP levels were still
TE D
significant predictors. However, some other potential confounding factors were not recorded, such as osteoporotic status, which is commonly associated with coronary artery calcification. We found a significant interaction with statin treatment on ALP levels according to CACS.
EP
Results indicate that statins seem to have a paradoxical effect on ALP activity. Non-calcified patients under statins had greater ALP levels, and conversely, ALP levels were significantly
AC C
lower in calcified patients. This epidemiological result has to take into account the fact that hypercholesterolemia could be another confounding factor. This effect could be linked to an exposition to a severe hypercholesterolemia or a higher cardiovascular risk requiring statin use. However, the biochemical effect of ALP activity reduction by statins in an osteogenic environment has already been shown in vitro. Osman et al. showed, on interstitial cells treated with osteogenic media, that atorvastatin induced a major 14-fold reduction of alkaline phosphatase activity [18]. Furthermore, this effect is reversed with mevalonate, which
10
ACCEPTED MANUSCRIPT indicates that the reduction of alkaline phosphatase activity is due to HMG-CoA reductase inhibition [19]. Clinical studies on the impact of statins on coronary artery calcification using computed tomography quantitative assessment are inconsistent [20-23]. Intravascular ultrasonography showed that statins promotes plaque calcification, which could explain their
RI PT
stabilizing effects [24]. In fact, statins could have a differential effect according to the atherosclerotic stage, with a pro calcifying effect in moderately involved lesions but not in advanced lesions [25,26]. These different observations seem to show a paradoxical effect of
SC
statins on coronary calcification, depending on the calcified or non-calcified status, and calls for further investigations.
M AN U
Evidence shows that coronary artery calcification follows the atherosclerotic plaque evolution, from microcalcification in unstable plaques to large calcified area in stabilized atherosclerotic lesions. Microcalcifications, occurring at an early stage of the calcification process, may promote shear stress and plaque rupture [27]. For the moment, no medical or
TE D
invasive therapy is effective to stop the progress of coronary calcification [28]. The metabolic pathway of ALP and PPi could be an effective target for new therapies against ectopic mineralization. Thus, bisphosphonates, non-hydrolysable agonists of pyrophosphate might
EP
reduce vascular calcification. In patients from the MESA cohort, bisphosphonates had differing effects depending on age, reducing coronary artery calcification in patients over 65
AC C
years, but promoting calcifications in younger patients [29]. Moreover, the association of etidronate and atorvastatin in a randomized controlled trial showed a greater abdominal aortic atheroma reduction than atorvastatin alone [30]. Other drugs have been developed to induce a direct inhibition of ALP pyrophosphatase function. Narisawa et al. showed a reduction of pyrophosphate hydrolysis of more than 20% with levamisole, and up to 40% with novel ALP inhibitors [31]. However, no human studies on ALP inhibitors are currently available.
11
ACCEPTED MANUSCRIPT This study has several limitations. First, we did not obtain complete data on mineral metabolism such as parathyroid hormone or vitamin D status, serum phosphorus or calcium concentrations. These different factors could have an impact on ALP activity. Secondly, we did not measure each ALP isoforms (liver, kidney, and bone), which could change the total
RI PT
ALP activity. However, PPi hydrolysis is a non-specific activity and each ALP isoform could drive this chemical reaction. Third, we did not obtain serum PPi concentrations. Thus, we cannot affirm the pathophysiological link between ALP activity and CACS. Fourthly, there
SC
also was a potential selection bias, as indicated by the high proportion of dyslipidemic patients in our population. However, only 47% of dyslipidemic patients had statin treatment,
M AN U
according to the French recommendation in primary cardiovascular prevention, which recommends to treat medically only moderate to high risk patients. Thus, our study population had a higher risk than the general population but had no previous cardiovascular events.
TE D
To conclude, we demonstrated, for the first time to our knowledge, an independent association between high ALP levels and coronary artery calcification in a sample of primary cardiovascular prevention patients. These results provide a pathophysiological explanation for
EP
previous observations associating high ALP levels to coronary artery events and cardiovascular mortality. In addition, the metabolic pathway of ALP and PPi could open a
AC C
new field for therapies against coronary artery calcification. However, these preliminary results must be carefully interpreted, particularly because of different confounding factors, and also support the need for further larger prospective studies to confirm the role of ALP and to improve our understanding of the impact of PPi on vascular calcification in humans.
12
ACCEPTED MANUSCRIPT CONFLICT OF INTEREST The authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.
FINANCIAL SUPPORT
RI PT
Loïc PANH was supported by the grant 2015-A00853-46 from the French Federation of Cardiology.
AUTHOR CONTRIBUTIONS
SC
Prof Ferrières and Dr Ruidavets had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Panh, Ruidavets, Ferrières.
M AN U
Acquisition, analysis, or interpretation of data: All Authors Drafting of the manuscript: Panh, Ruidavets, Ferrières.
Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Ruidavets, Ferrières. Obtained funding: Panh.
TE D
Study supervision: Galinier, Carrié, Ferrières.
ACKNOWLEDGEMENTS
AC C
EP
We thank Dr Anne Thuéry, PhD, for her help in revising the manuscript.
13
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
REFERENCES [1] A.S. Agatston, W.R. Janowitz, F.J. Hildner, N.R. Zusmer, M. Viamonte et al., Quantification of coronary artery calcium using ultrafast computed tomography, J Am Coll Cardiol. 15 (1990) 827-832. [2] M.J. Budoff, L.J. Shaw, S.T. Liu, S.R. Weinstein, T.P. Mosleret al., Long-term prognosis associated with coronary calcification: observations from a registry of 25,253 patients, J Am Coll Cardiol. 49 (2007) 1860-1870. [3] L.L. Demer, Y. Tintut, Inflammatory, metabolic, and genetic mechanisms of vascular calcification, Arterioscler Thromb Vasc Biol. 34 (2014) 715-723. [4] L.L. Demer, Y. Tintut, Vascular calcification: pathobiology of a multifaceted disease, Circulation. 117 (2008) 2938-2948. [5] D.W. Urry, Neutral sites for calcium ion binding to elastin and collagen: a charge neutralization theory for calcification and its relationship to atherosclerosis, Proc Natl Acad Sci U S A. 68 (1971) 810-814. [6] K.A. Lomashvili, S. Cobbs, R.A. Hennigar, K.I. Hardcastle, W.C. O'Neill, Phosphateinduced vascular calcification: role of pyrophosphate and osteopontin, J Am Soc Nephrol. 15 (2004) 1392-1401. [7] H.A. Fleisch, R.G. Russell, S. Bisaz, R.C. Mühlbauer, D.A. Williams, The inhibitory effect of phosphonates on the formation of calcium phosphate crystals in vitro and on aortic and kidney calcification in vivo, Eur J Clin Invest. 1 (1970) 12-8. [8] M.D. Francis, The inhibition of calcium hydroxypatite crystal growth by polyphosphonates and polyphosphates, Calcif Tissue Res. 3 (1969) 151-162. [9] N.M. Hansen, R. Felix, S. Bisaz, H. Fleisch, Aggregation of hydroxyapatite crystals, Biochim Biophys Acta. 451 (1976) 549-559. [10] R. Villa-Bellosta, J. Rivera-Torres, F.G. Osorio, R. Acín-Pérez, J.A. Enriquez et al., Defective extracellular pyrophosphate metabolism promotes vascular calcification in a mouse model of Hutchinson-Gilford progeria syndrome that is ameliorated on pyrophosphate treatment. Circulation 127 (2013) 2442-2451. [11] C. St Hilaire, S.G. Ziegler, T.C. Markello, A. Brusco, C. Groden et al., NT5E mutations and arterial calcifications, N Engl J Med. 364 (2011) 432-442. [12] M. Tonelli, G. Curhan, M. Pfeffer, F. Sacks, R. Thadhani, et al., Relation between alkaline phosphatase, serum phosphate, and all-cause or cardiovascular mortality, Circulation. 120 (2009) 1784-1792. [13] J.B. Park, D. Kang, H.M. Yang, H.J. Cho, K.W. Park et al., Serum alkaline phosphatase is a predictor of mortality, myocardial infarction, or stent thrombosis after implantation of coronary drug-eluting stent, Eur Heart J. 34 (2012) 920–931. [14] R. Shantouf, C.P. Kovesdy, Y. Kim, N. Ahmadi, A. Luna et al., Association of serum alkaline phosphatase with coronary artery calcification in maintenance hemodialysis patients, Clin J Am Soc Nephrol. 4 (2009) 1106-1014. [15] A.S. Levey, L.A. Stevens, C.H. Schmid, Y. Zhang, A.F. Castro, et al., A New Equation to Estimate Glomerular Filtration Rate, Ann Intern Med. 150 (2009) 604–612. [16] K.A. Lomashvili, P. Garg, S. Narisawa, J.L. Millan, W.C. O'Neill, Upregulation of alkaline phosphatase and pyrophosphate hydrolysis: Potential mechanism for uremic vascular calcification, Kidney Int. 73 (2008) 1024–1030. [17] R. Villa-Bellosta, V. Sorribas, Calcium Phosphate Deposition With Normal Phosphate Concentration-Role of Pyrophosphate, Circ J. 75 (2011) 2705-2710. [18] L. Osman, M.H. Yacoub, N. Latif, M. Amrani, A.H. Chester, Role of human valve interstitial cells in valve calcification and their response to atorvastatin, Circulation. 114 (2006) 547-552.
14
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
[19] B. Wu, S. Elmariah, F.S. Kaplan, G. Cheng, E.R. Mohler, Paradoxical effects of statins on aortic valve myofibroblasts and osteoblasts implications for end-stage valvular heart disease, Arterioscler Thromb Vasc Biol. 25 (2005) 592-597. [20] T.Q Callister, P. Raggi, B. Cooil, N.J. Lippolis, D.J. Russo, Effect of HMG-CoA reductase inhibitors on coronary artery disease as assessed by electron-beam computed tomography, N Engl J Med. 339 (1998) 1972-1978. [21] M.J. Budoff, K.L. Lane, H. Bakhsheshi, S. Mao, B.O. Grassmann et al., Rates of progression of coronary calcium by electron beam tomography, Am J Cardiol. 86 (2000) 811. [22] S. Achenbach, D. Ropers, K. Pohle, A. Leber, C. Thilo et al., Influence of lipid-lowering therapy on the progression of coronary artery calcification: a prospective evaluation, Circulation. 106 (2002) 1077-1082. [23] M. Henein, G. Granåsen, U. Wiklund, A. Schmermund, A. Guerci et al., High dose and long-term statin therapy accelerate coronary artery calcification. Int J Cardiol. 184 (2015) 581-586. [24] R. Puri, S.J. Nicholls, M. Shao, Y. Kataoka, K. Uno et al., Impact of statins on serial coronary calcification during atheroma progression and regression, J Am Coll Cardiol. 65 (2015) 1273-1282. [25] J.K. Williams, G.K. Sukhova, D.M. Herrington, P. Libby, Pravastatin has cholesterollowering independent effects on the artery wall of atherosclerotic monkeys, J Am Coll Cardiol. 31 (1998) 684-691. [26] I. Dykun, N. Lehmann, H. Kälsch, S. Möhlenkamp, S. Moebus et al., Statin Medication Enhances Progression of Coronary Artery Calcification: The Heinz Nixdorf Recall Study, J Am Coll Cardiol. 68 (2016) 2123-2125. [27] K. Imoto, T. Hiro, T. Fujii, A. Murashige, Y. Fukumoto, et al. Longitudinal structural determinants of atherosclerotic plaque vulnerability: a computational analysis of stress distribution using vessel models and three-dimensional intravascular ultrasound imaging, J Am Coll Cardiol. 46 (2005) 1507-1515. [28] M.V. Madhavan, M. Tarigopula, G.S. Mintz, A. Maehara, G.W. Stone, et al. Coronary artery calcification: pathogenesis and prognostic implications, J Am Coll Cardiol. 63 (2014) 1703-1714. [29] S. Elmariah, J.A. Delaney, K.D. O'Brien, M.J. Budoff, J. Vogel-Claussen, et al., Bisphosphonate Use and Prevalence of Valvular and Vascular Calcification in Women MESA (The Multi-Ethnic Study of Atherosclerosis), J Am Coll Cardiol. 56 (2010) 1752-1759. [30] T. Kawahara, M. Nishikawa, C. Kawahara, T. Inazu, K. Sakai, et al., Atorvastatin, Etidronate, or Both in Patients at High Risk for Atherosclerotic Aortic Plaques: A Randomized Controlled Trial, Circulation. 127 (2013) 2327-2335. [31] S. Narisawa, D. Harmey, M.C. Yadav, W.C. O'Neill, M.F. Hoylaerts et al., Novel inhibitors of alkaline phosphatase suppress vascular smooth muscle cell calcification, J Bone Miner Res. 22 (2007) 1700-1710.
15
ACCEPTED MANUSCRIPT TABLES Table 1. Characteristics of each tertile of ALP and CACS distribution. Mean
SD
ALP <55 IU/L
171 (34.2)
45.7
6.6
ALP 55-66 IU/L
167 (33.4)
60.2
3.5
ALP >66 IU/L
162 (32.4)
83.1
19.0
CACS = 0
187 (37.4)
-
CACS > 0
313 (62.6)
257
CACS 1-100
169 (33.8)
29
CACS 101-400
92 (18.4)
209
CACS ≥401
52 (10.4)
1058
RI PT
n (%)
-
588 30
89
SC
Variable
1100
AC C
EP
TE D
M AN U
CACS, coronary artery calcium score; ALP, serum alkaline phosphatase; SD, standard deviation.
ACCEPTED MANUSCRIPT Table 2. Baseline characteristics according to CACS. Variables
All patients
CACS=0
CACS>0
p
(n=500)
(n=187)
(n=313)
60.9 (10.8)
55.3 (11.1)
64.3 (9.2)
0.001
Woman (%)
49.6
62.6
41.1
0.001
Smokers (%)
13.0
14.0
12.7
0.68
Hypertension (%)
28.8
18.7
34.6
0.001
AC C
EP
TE D
M AN U
SC
Age, mean (SD), years
RI PT
Clinical
6.9
2.7
9.5
0.004
Dyslipidemia (%)
70.7
66.8
73.1
0.14
BMI, mean (SD), kg/m²
25.5 (4.7)
25.5 (5.4)
25.5 (4.3)
0.96
SBP, mean (SD), mmHg
136 (17)
131 (15)
138 (17)
0.001
DBP, mean (SD), mmHg
82 (12)
81 (11)
83 (12)
0.14
37.2
26.2
43.7
0.001
81.6 (13.9)
84.1 (14.3)
80.2 (13.5)
0.003
Total cholesterol, mean (SD), mmol/L
2.22 (0.47)
2.26 (0.44)
2.20 (0.49)
0.19
Triglycerides, mean (SD), mmol/L
1.27 (1.03)
HDL cholesterol, mean (SD), mmol/L
0.59 (0.18)
LDL cholesterol, mean (SD), mmol/L
1.38 (0.44)
Statin (%) Biological Glomerular filtration rate, mean (SD), ml/mn/1.73 m2
ALP, mean (SD), IU/L a
1.29 (1.27)
1.27 (0.86)
0.44 a
0.61 (0.19)
0.58 (0.18)
0.06
1.40 (0.42)
1.38 (0.45)
0.56
2.1 (2.9)
2.0 (2.8)
2.2 (3.0)
0.25a
62.6 (19.3)
59.0 (17)
64.7 (20)
0.001a
M AN U
CRP, mean (SD), mg/L
RI PT
Diabetes (%)
SC
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Tesr on log transformed data. CACS,coronary artery calcium score; SBP,systolic blood pressure; DBP,diastolic blood pressure; CRP,C reactive protein; ALP,serum alkaline phosphatase. Data are presented as % or mean (standard deviation).
ACCEPTED MANUSCRIPT Table 3. Logistic regression for prediction of CACS>0 in 500 primary prevention patients. 95% CIa
pa
Age, years
1.12
1.08 ; 1.15
0.001
Men
3.84
2.39 ; 6.15
0.001
Diabetes
2.93
0.93 ; 9.24
0.07
Hypertension
1.06
0.62 ; 1.79
0.84
Current vs. no smokers
1.61
0.84 ; 3.06
0.15
Glomerular filtration rate, ml/mn/1.73 m2
1.02
0.99 ; 1.03
0.10
Statin
0.49
0.23 ; 1.05
0.07
ALP 55-66 vs. <55 IU/L
2.11
1.12 ; 3.96
0.02
ALP >66 vs. <55IU/L
3.89
2.01 ; 7.54
0.001
M AN U
SC
RI PT
ORa
Variables
a
AC C
EP
TE D
With interaction between statin and ALP. For ALP p for trend=0.001 in full sample and in subsample without statin analyses. ALP,serum alkaline phosphatase; OR,odds ratio; CI,confidence interval.
Table 4. Logistic regression for prediction of CACS>0 in patients with or without statin treatment. ACCEPTED MANUSCRIPT Patients without statin (n=314) Variables
Patients with statin (n=186)
95% CI
p
OR
95% CI
p
Age, years
1.13
1.09 ; 1.17
0.001
1.10
1.04 ; 1.15
0.001
Men
3.68
2.06 ; 6.59
0.001
4.47
1.94 ; 10.3
0.001
Diabetes
3.25
0.64 ; 16.3
0.16
2.74
0.54 ; 13.8
0.22
Hypertension
1.38
0.70 ; 2.72
0.35
0.77
0.32 ; 1.87
0.56
Current vs. no smokers
1.31
0.60 ; 2.87
0.51
2.52
0.74 ; 8.65
0.14
1.02
0.99 ; 1.04
0.11
1.01
0.98 ; 1.04
0.60
ALP 55-66 vs. <55 IU/L
2.19
1.14 ; 4.19
0.018
1.56
0.59 ; 4.10
0.37
ALP >66 vs. <55IU/L
3.87
1.96 ; 7.61
0.58
0.24 ; 1.45
0.25
ml/mn/1.73 m2
SC
Glomerular filtration rate,
RI PT
OR
0.001
AC C
EP
TE D
M AN U
ALP,serum alkaline phosphatase; OR,odds ratio; CI,confidence interval.
ACCEPTED MANUSCRIPT HIGHLIGHTS •
Alkaline phosphatase is an independent predictor of coronary artery calcification.
•
Statins have an impact on alkaline phosphatase activity.
•
The pyrophosphate pathway could be a target for therapies against vascular
AC C
EP
TE D
M AN U
SC
RI PT
calcification.
S0021-9150(17)30132-6
DOI:
10.1016/j.atherosclerosis.2017.03.030
Reference:
ATH 15005
To appear in:
Atherosclerosis
Received Date: 1 December 2016 Revised Date:
15 March 2017
Accepted Date: 22 March 2017
Please cite this article as: Panh L, Ruidavets JB, Rousseau H, Petermann A, Bongard V, Bérard E, Taraszkiewicz D, Lairez O, Galinier M, Carrié D, Ferrières J, Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients, Atherosclerosis (2017), doi: 10.1016/j.atherosclerosis.2017.03.030. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Association between serum alkaline phosphatase and coronary artery calcification in a sample of primary cardiovascular prevention patients Loïc Panha, Jean Bernard Ruidavetsb, Hervé Rousseauc, Antoine Petermannc, Vanina Bongardb, Emilie Bérardb, Dorota Taraszkiewicza, Olivier Laireza, Michel Galiniera, Didier
a
RI PT
Carriéa, Jean Ferrièresa.
Department of Cardiology, Toulouse-Rangueil University Hospital (CHU), TSA 50032,
Toulouse University School of Medicine, 1 Avenue du Professeur Jean Poulhes 31059
b
SC
Toulouse Cedex 9, France.
Department of Epidemiology, Unité de Soutien Méthodologique à la Recherche (USMR),
France. c
M AN U
Toulouse University Hospital (CHU), 37 Allées Jules Guesde, 31073 Toulouse Cedex 7,
Department of Radiology, Toulouse-Rangueil University Hospital (CHU), TSA 50032,
Toulouse University School of Medicine, 1 Avenue du Professeur Jean Poulhes 31059
TE D
Toulouse Cedex 9, France.
Corresponding author: CHU Toulouse Rangueil, Department of Cardiology, 1, avenue du
EP
Pr Jean Poulhes, 31059 Toulouse Cedex 9, France.
AC C
Email: [email protected] (J. Ferrières)
1
ACCEPTED MANUSCRIPT
ABSTRACT Background and aims: A high level of serum alkaline phosphatase (ALP) is associated with an increased risk of mortality and myocardial infarction. ALP hydrolyses inorganic
RI PT
pyrophosphate, which is a strong inhibitor of calcium phosphate deposition. The aim of this study was to determine whether ALP is associated with the coronary artery calcium score (CACS).
SC
Methods: We examined the association of CACS, assessed by computed tomography scanning, and ALP, in 500 patients consecutively recruited, free of cardiovascular disease.
M AN U
The CACS were categorized into two groups: no calcification (CACS=0) (n=187) and with calcification (CACS>0) (n=313). ALP activity was divided into three tertile groups: low ALP level (<55 IU/L), intermediate (55 to 66 IU/L) and high ALP level (>66 IU/L). Results: The mean age was 60.9 ± 10.8 years, 49.6% of the patients were women. ALP ranged
TE D
from 22 to 164 IU/L (mean 62.6 IU/L, SD 19.3). In univariate analysis, traditional cardiovascular risk factors, statin use (p=0.001), and ALP (p=0.001) were significantly associated withCACS. After adjusting for cardiovascular risk factors, only age (p=0.001) and
EP
sex (p=0.001) were independently associated with CACS. Compared to the tertile group with
AC C
low levels of ALP, the intermediate tertile group [OR 2.11, 95%CI (1.12;3.96), p=0.02], as well as the high tertile group [OR 3.89, 95% CI (2.01;7.54), p=0.001)], was independently associated with CACS.
Conclusions: In patients free of cardiovascular disease, high ALP levels are positively and independently associated with coronary artery calcification. The metabolic pathway of ALP and inorganic pyrophosphate could be a target for new therapies against vascular calcification. KEYWORDS Coronary artery disease; Calcification; Alkaline phosphatase; Statin.
2
ACCEPTED MANUSCRIPT 1- INTRODUCTION Intimal atherosclerotic plaque calcification is considered as the result of an inflammatory process initiated by infiltrating oxidized lipids. Coronary artery calcification is generally assessed by non-contrast computed tomography and coronary artery calcium score (CACS)
RI PT
measurement [1]. CACS is a reflection of the atheromatous disease and a potent predictor of cardiovascular events and mortality [2]. It is due to several active and passive mechanisms of ectopic mineralization leading to hydroxyapatite crystal formation on the layers of the arterial
SC
wall. Active mechanisms of calcification involve osteoblastic differentiated cells, and lead to
M AN U
an organized calcified bone-like matrix [3]. These cellular processes are highly regulated, under the influence of inflammatory cytokines, oxidant stress, or mineral metabolism [4]. However, the majority of vascular mineralization is due to passive phenomena, subject to the neutral charge theory of calcification [5], leading to unstructured and amorphous calcified areas. Contrary to previous thinking, passive calcium phosphate deposition is also regulated.
TE D
One of the most powerful passive calcium phosphate deposition inhibitors is inorganic pyrophosphate (PPi) [6]. The biochemical effects of PPi are an inhibition of calcium and phosphate aggregation [7], hydroxyapatite crystal growth [8] and aggregation [9]. In vivo, PPi
EP
is hydrolyzed into inorganic phosphate by serum alkaline phosphatase (ALP). Thus, an
AC C
increased ALP activity can induce an imbalance between inorganic phosphate and pyrophosphate, promoting ectopic calcification. Thus, a pathological increase in ALP activity is linked to extensive vascular calcification, leading to premature atherosclerosis and cardiovascular events, such as Hutchinson-Gilford progeria syndrome or the Generalized Arterial Calcification of Infancy syndrome [10,11]. Besides these rare diseases, high ALP levels are associated with an increased risk of cardiovascular events and mortality in the general population and in secondary cardiovascular prevention [12,13] and higher coronary
3
ACCEPTED MANUSCRIPT artery calcifications in hemodialysis patients [14]. However, data on a possible role of ALP in coronary artery calcification in a primary cardiovascular prevention population are lacking. The aim of this study was to assess the potential association between ALP and CACS in
RI PT
patients free of any cardiovascular disease.
2- PATIENTS AND METHODS 2-1- Study population
SC
Our study sample included 555 patients ranging in age from 25 to 83 years, consecutively
M AN U
enrolled between September 2014 and October 2016 in the department of preventive cardiology (Toulouse University Hospital). This was a primary prevention population, selfreferred or referred by their primary-care physician or cardiologist for cardiovascular risk stratification. The inclusion criterion was all consecutive patients over 18 years old. The exclusion criteria concerned secondary prevention patients, defined as having undergone
TE D
acute coronary syndrome, stable angina, coronary revascularization (surgical or percutaneous intervention), or stroke. Patients with acute liver, gallbladder or other gastrointestinal diseases, active infectious diseases, potentially responsible for an increase of ALP activity
EP
were also excluded. Authorization to use these data was obtained in accordance with the
AC C
French law (Commission Nationale de l’Informatique et Libertés - CNIL).
2-2- Cardiovascular risk factors and laboratory data Subjects were given a questionnaire to assess cardiovascular risk factors during a medical consultation. The cardiovascular risk factors included smoking status, diabetes, hypertension, dyslipidemia, age, sex, and family history of premature cardiovascular event. Data concerning personal medical history, lifestyle habits and drug intake were recorded. Blood pressure was measured after at least 5 minutes of seated rest, with a standard automated
4
ACCEPTED MANUSCRIPT sphygmomanometer. Blood pressure was measured three times on both arms at 5 minute intervals. The mean of the measurements was used for statistical analysis. A standardized procedure was used to measure height and weight. Body mass index was calculated as the weight divided by the height squared (kg/m²).
RI PT
A blood sample was taken after over 12 hours of overnight fasting at the central laboratory of Toulouse University Hospital. Serum total cholesterol and triglycerides were measured by enzymatic assays (Cobas 8000, Roche Diagnostics, Germany). High density lipoprotein
SC
cholesterol (HDL-C) was measured after sodium phosphotungstate-magnesium chloride precipitation of apolipoprotein B-containing proteins. Low density lipoprotein cholesterol
M AN U
(LDL-C) was determined using the Friedewald formula when triglycerides were < 4.6 mmol/l. Glucose concentration were measured using a conventional enzymatic method based on hexokinase-glucose-6-phosphate dehydrogenase.
Serum
creatinine
measurement
was
performed with an enzymatic method with calibration certified by isotopic dilution mass
TE D
spectroscopy. Glomerular filtration rate (GFR) was estimated using the serum creatinine level measurement and the CKD-EPI equation calculation [15]. ALP activity was measured by a spectrophotometric method, using p-nitrophenylphosphate as
EP
a substrate, in an automated analyzer (Cobas 8000, Roche Diagnostics, Germany). Interassay variability ranged between 0.9 and 2.4%, and the normal reference ranges in our laboratory
AC C
were 35 to 104 IU/L in women and 40 to 129 IU/L in men. Hypertension was defined as mean blood pressure >140/90 mmHg at rest, or personal history of hypertension or treatment for hypertension. Smoking was defined as current smoking or no smoking. Patients who had stopped smoking at least one month before the investigation were counted as non-smokers. Family history of premature cardiovascular disease was defined as coronary disease or cerebrovascular disease before the age of 55 in men and 65 in women, for first-degree family members. Dyslipidemia was defined in this moderate cardiovascular risk
5
ACCEPTED MANUSCRIPT population as a total plasma cholesterol ≥ 6.5 mmol/l, or LDL-C ≥ 4 mmol/l, or statin use, or other lipid lowering treatment, or triglycerides ≥ 2.3 mmol/l. Diabetes was defined as plasma glucose level ≥ 7 mmol/l or history of medical treatment for diabetes. Cardiovascular disease
and femoral arteries assessed by ultrasound imaging.
RI PT
was excluded by a negative stress test and the lack of atheroma in the carotid arteries, aorta
2-3- Cardiac multislice computed tomography acquisition and interpretation
SC
All computed tomography scans were performed with a dual source CT-system (Somaton Définition; Siemens Healthcare, Forchheim, Germany). Scan parameters for the calcium score
M AN U
protocol were as follows: 3mm slice thickness, collimation 6x3, 120kV tube voltage, 68mAs, field of view 250mm during a breath-hold maneuver. Image reconstruction was performed at 70% of R-R interval. Approximately 50 sections were obtained at the level of the carina and proceeding caudally to the level of the diaphragm. A calcified lesion was defined as more
TE D
than 3 contiguous pixels with a peak of attenuation of at least 130 Hounsfield units. The total coronary artery calcium score (CACS) was determined according to the Agatston method, by summing individual lesion scores from each of 4 anatomic sites (left main, left anterior
EP
descending, circumflex, and right coronary artery). CACS were stratified in two groups, the
AC C
calcified patients (CACS>0) and non-calcified patients (CACS=0).
2-4- Statistical analysis
Data are presented as means and standard deviations for quantitative variables and percentages for categorical variables. ALP levels were stratified in three balanced tertiles groups: <55 IU/L, 55 to 66 IU/L, >66 IU/L. The Chi2 test was used to compare the distribution of qualitative variables between the two CACS groups. Mean values of quantitative variables were compared by one-way analysis of variance. Shapiro-Wilk’s and
6
ACCEPTED MANUSCRIPT Levene’s tests were used to test the normality of the distribution of residuals and the homogeneity of variances, respectively. When basic assumptions were not satisfied, a logarithmic transformation of the variables was carried out. Multivariate analyses were performed using multiple logistic regression, to test the association between ALP
RI PT
concentrations and CACS groups (CACS=0 taken as reference group) and to calculate odds ratios (OR) and 95% confidence intervals. A backward selection method of the variables was used for all the models. All selected variables were entered in a full model and then a step-by
remove non-significant contributory variables.
SC
step-elimination was performed. A likelihood ratio test was made between nested models to
M AN U
First-order interactions between ALP and all other variables remaining in the final model were tested systematically by adding an interaction term. A multiplicative product of ALP with each other variable was introduced into the model taking into account the nature of the variable studied; categorical, ordinal or continuous. Statistical significance was tested using
TE D
the likelihood ratio test between nested models with and without the term of interaction. Logistic regression analyses with polynomial terms (quadratic and cubic) were performed to examine for possible non-linear relationships between ALP, used as a continuous variable,
EP
and the binary variable CACS. The relation between ALP and CACS was considered as a linear response function because the second and third order terms were not significant.
AC C
All the statistical analyses were conducted using SAS (SAS Institute, version 9.4, Cary, NC) and Stata (Stata, release 14.0, College Station, TX). We considered p<0.05 as significant.
3- RESULTS 3-1- Characteristics of the study population and associations to CACS After excluding 55 patients according to the exclusion criteria, the study population consisted of 500 patients who had ALP measured as baseline and were eligible for statistical analysis.
7
ACCEPTED MANUSCRIPT The mean age was 60.9 ± 10.8 years, hypertension affected 28.8% of patients, 6.9% had diabetes, 70.7% had dyslipidemia, 37.2% underwent statin therapy. The mean GFR was 81.6 ± 13.9 mL/min/1.73 m2, and only 5.4% had a GFR below 60 mL/min/1.73 m2. ALP levels ranged from 22 to 164 IU/L (median 60 IU/L, mean 62.6 IU/L, SD 19.3). ALP levels outside
RI PT
of the normal range were observed in 39 patients (7.8%), with 26 patients (5.2 %) below and 13 patients (2.6%) above the reference value. The mean CACS was 163 ± 484. The proportion of CACS = 0 was 37.4% (n=187) and 62.6% (n=313) had CACS > 0. The
SC
characteristics of each ALP group and the distribution of CACS are summarized in Table 1. In univariate analysis, traditional cardiovascular risk factors such as age (p=0.001), sex
M AN U
(p=0.001), hypertension (p=0.001), diabetes (p=0.004), and GFR (p=0.003) were significantly associated with CACS. There was no significant association between CACS and dyslipidemia, HDL cholesterol or LDL cholesterol. Statin use was associated with CACS (p=0.001). Before adjustment, ALP was significantly associated withCACS (p=0.001) (Table
TE D
2).
3-2- Adjusted association with CACS and interaction between ALP and statin treatment
EP
A multiple logistic regression was performed on cardiovascular risk factors (age, sex, diabetes, hypertension, smoking, and GFR), statin treatment, and ALP activity. In the full
AC C
sample, only age (p=0.001) and sex (p= 0.001) and ALP were independently associated withCACS. Compared to the group with low levels of ALP (<55 IU/L), intermediate levels of ALP (55 IU/L to 66 IU/L) were significantly associated withCACS [OR 2.11, 95% CI (1.12;3.96), p=0.02], and high levels of ALP (>66 IU/L) were associated with CACS [OR 3.89, 95% CI (2.01;7.54), p=0.001)]. There was a linear association (p for trend 0.001) between ALP and CACS (Table 3). Otherwise, results were similar using continuous ALP levels analysis [for one ALP unit (IU/L) increase OR=1.03, 95% CI (1.02-1.05), p<0.001].
8
ACCEPTED MANUSCRIPT Age was the strongest independent predictor of CACS. However, there was no interaction between age and ALP levels (p=0.35). For one ALP unit (IU/L) increase, adjusted OR were similar in the youngest patients [age below the median; OR=1.03, 95%CI (1.01; 1.05), p=0.007] as well as the oldest patients [age above the median; OR=1.05, 95% CI (1.01; 1.08),
RI PT
p=0.009].
We found a significant interaction between ALP and statin use (p=0.004). Thus, statin treatment had an opposite effect on ALP according to the calcified or non-calcified patient’s
SC
status. In non-calcified patients (CACS=0), ALP activity was significantly higher with statin treatment [56.8 IU/L vs. 65.2IU/L (p=0.005)]. Conversely in calcified patients (CACS>0),
M AN U
ALP activity was significantly lower with statin treatment [67.0IU/L vs. 61.8IU/L (p=0.008)]. In the same multivariate analysis, according to statin treatment, the intermediate and high ALP levels were associated with CACS in patients without statin, but not in patients with statin. The other results were similar as the full sample, and only age and sex were
4- DISCUSSION
TE D
independently associated with CACS (Table 4).
EP
The major finding of this study was that, in a primary cardiovascular prevention population, a
AC C
high level of ALP activity was positively and independently associated with the presence of coronary artery calcification. Thus, ALP levels higher than 55 IU/L and particularly higher than 66 IU/L were potent predictors of coronary artery calcification. Age was the most powerful predictor of CACS, but did not directly interfere with the relationship between ALP and CACS which was significant regardless of age. The mechanism behind the relationship between ALP and artery calcification is based on previous observations on the inhibitory effect of PPi on vascular calcification and its
9
ACCEPTED MANUSCRIPT hydrolysis by ALP [16]. A physiological concentration of PPi can completely inhibit calcium and phosphate deposition under a physiological concentration of calcium and phosphate [17]. Epidemiological studies showed that high ALP levels are linked to cardiovascular events both in the general population and in secondary cardiovascular prevention patients [12,13]. Our
RI PT
results could provide a pathophysiological explanation for the increase in coronary artery calcification, which is a potent marker of cardiovascular atheroma and cardiovascular events. Interestingly, more than 90% of patients included in both these studies as well as our study
SC
had ALP levels in the range usually considered as normal.
This link between ALP and CACS is subject to several confounding factors, such as GFR,
M AN U
inflammatory status, or liver enzymes which are known to increase the cardiovascular risk. To try to reduce the influence of these potential confounding factors, we excluded active gastrointestinal, liver, or infectious diseases, and no woman was pregnant. A multiple adjustment was done on these factors, and intermediate as well as high ALP levels were still
TE D
significant predictors. However, some other potential confounding factors were not recorded, such as osteoporotic status, which is commonly associated with coronary artery calcification. We found a significant interaction with statin treatment on ALP levels according to CACS.
EP
Results indicate that statins seem to have a paradoxical effect on ALP activity. Non-calcified patients under statins had greater ALP levels, and conversely, ALP levels were significantly
AC C
lower in calcified patients. This epidemiological result has to take into account the fact that hypercholesterolemia could be another confounding factor. This effect could be linked to an exposition to a severe hypercholesterolemia or a higher cardiovascular risk requiring statin use. However, the biochemical effect of ALP activity reduction by statins in an osteogenic environment has already been shown in vitro. Osman et al. showed, on interstitial cells treated with osteogenic media, that atorvastatin induced a major 14-fold reduction of alkaline phosphatase activity [18]. Furthermore, this effect is reversed with mevalonate, which
10
ACCEPTED MANUSCRIPT indicates that the reduction of alkaline phosphatase activity is due to HMG-CoA reductase inhibition [19]. Clinical studies on the impact of statins on coronary artery calcification using computed tomography quantitative assessment are inconsistent [20-23]. Intravascular ultrasonography showed that statins promotes plaque calcification, which could explain their
RI PT
stabilizing effects [24]. In fact, statins could have a differential effect according to the atherosclerotic stage, with a pro calcifying effect in moderately involved lesions but not in advanced lesions [25,26]. These different observations seem to show a paradoxical effect of
SC
statins on coronary calcification, depending on the calcified or non-calcified status, and calls for further investigations.
M AN U
Evidence shows that coronary artery calcification follows the atherosclerotic plaque evolution, from microcalcification in unstable plaques to large calcified area in stabilized atherosclerotic lesions. Microcalcifications, occurring at an early stage of the calcification process, may promote shear stress and plaque rupture [27]. For the moment, no medical or
TE D
invasive therapy is effective to stop the progress of coronary calcification [28]. The metabolic pathway of ALP and PPi could be an effective target for new therapies against ectopic mineralization. Thus, bisphosphonates, non-hydrolysable agonists of pyrophosphate might
EP
reduce vascular calcification. In patients from the MESA cohort, bisphosphonates had differing effects depending on age, reducing coronary artery calcification in patients over 65
AC C
years, but promoting calcifications in younger patients [29]. Moreover, the association of etidronate and atorvastatin in a randomized controlled trial showed a greater abdominal aortic atheroma reduction than atorvastatin alone [30]. Other drugs have been developed to induce a direct inhibition of ALP pyrophosphatase function. Narisawa et al. showed a reduction of pyrophosphate hydrolysis of more than 20% with levamisole, and up to 40% with novel ALP inhibitors [31]. However, no human studies on ALP inhibitors are currently available.
11
ACCEPTED MANUSCRIPT This study has several limitations. First, we did not obtain complete data on mineral metabolism such as parathyroid hormone or vitamin D status, serum phosphorus or calcium concentrations. These different factors could have an impact on ALP activity. Secondly, we did not measure each ALP isoforms (liver, kidney, and bone), which could change the total
RI PT
ALP activity. However, PPi hydrolysis is a non-specific activity and each ALP isoform could drive this chemical reaction. Third, we did not obtain serum PPi concentrations. Thus, we cannot affirm the pathophysiological link between ALP activity and CACS. Fourthly, there
SC
also was a potential selection bias, as indicated by the high proportion of dyslipidemic patients in our population. However, only 47% of dyslipidemic patients had statin treatment,
M AN U
according to the French recommendation in primary cardiovascular prevention, which recommends to treat medically only moderate to high risk patients. Thus, our study population had a higher risk than the general population but had no previous cardiovascular events.
TE D
To conclude, we demonstrated, for the first time to our knowledge, an independent association between high ALP levels and coronary artery calcification in a sample of primary cardiovascular prevention patients. These results provide a pathophysiological explanation for
EP
previous observations associating high ALP levels to coronary artery events and cardiovascular mortality. In addition, the metabolic pathway of ALP and PPi could open a
AC C
new field for therapies against coronary artery calcification. However, these preliminary results must be carefully interpreted, particularly because of different confounding factors, and also support the need for further larger prospective studies to confirm the role of ALP and to improve our understanding of the impact of PPi on vascular calcification in humans.
12
ACCEPTED MANUSCRIPT CONFLICT OF INTEREST The authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.
FINANCIAL SUPPORT
RI PT
Loïc PANH was supported by the grant 2015-A00853-46 from the French Federation of Cardiology.
AUTHOR CONTRIBUTIONS
SC
Prof Ferrières and Dr Ruidavets had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Panh, Ruidavets, Ferrières.
M AN U
Acquisition, analysis, or interpretation of data: All Authors Drafting of the manuscript: Panh, Ruidavets, Ferrières.
Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Ruidavets, Ferrières. Obtained funding: Panh.
TE D
Study supervision: Galinier, Carrié, Ferrières.
ACKNOWLEDGEMENTS
AC C
EP
We thank Dr Anne Thuéry, PhD, for her help in revising the manuscript.
13
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
REFERENCES [1] A.S. Agatston, W.R. Janowitz, F.J. Hildner, N.R. Zusmer, M. Viamonte et al., Quantification of coronary artery calcium using ultrafast computed tomography, J Am Coll Cardiol. 15 (1990) 827-832. [2] M.J. Budoff, L.J. Shaw, S.T. Liu, S.R. Weinstein, T.P. Mosleret al., Long-term prognosis associated with coronary calcification: observations from a registry of 25,253 patients, J Am Coll Cardiol. 49 (2007) 1860-1870. [3] L.L. Demer, Y. Tintut, Inflammatory, metabolic, and genetic mechanisms of vascular calcification, Arterioscler Thromb Vasc Biol. 34 (2014) 715-723. [4] L.L. Demer, Y. Tintut, Vascular calcification: pathobiology of a multifaceted disease, Circulation. 117 (2008) 2938-2948. [5] D.W. Urry, Neutral sites for calcium ion binding to elastin and collagen: a charge neutralization theory for calcification and its relationship to atherosclerosis, Proc Natl Acad Sci U S A. 68 (1971) 810-814. [6] K.A. Lomashvili, S. Cobbs, R.A. Hennigar, K.I. Hardcastle, W.C. O'Neill, Phosphateinduced vascular calcification: role of pyrophosphate and osteopontin, J Am Soc Nephrol. 15 (2004) 1392-1401. [7] H.A. Fleisch, R.G. Russell, S. Bisaz, R.C. Mühlbauer, D.A. Williams, The inhibitory effect of phosphonates on the formation of calcium phosphate crystals in vitro and on aortic and kidney calcification in vivo, Eur J Clin Invest. 1 (1970) 12-8. [8] M.D. Francis, The inhibition of calcium hydroxypatite crystal growth by polyphosphonates and polyphosphates, Calcif Tissue Res. 3 (1969) 151-162. [9] N.M. Hansen, R. Felix, S. Bisaz, H. Fleisch, Aggregation of hydroxyapatite crystals, Biochim Biophys Acta. 451 (1976) 549-559. [10] R. Villa-Bellosta, J. Rivera-Torres, F.G. Osorio, R. Acín-Pérez, J.A. Enriquez et al., Defective extracellular pyrophosphate metabolism promotes vascular calcification in a mouse model of Hutchinson-Gilford progeria syndrome that is ameliorated on pyrophosphate treatment. Circulation 127 (2013) 2442-2451. [11] C. St Hilaire, S.G. Ziegler, T.C. Markello, A. Brusco, C. Groden et al., NT5E mutations and arterial calcifications, N Engl J Med. 364 (2011) 432-442. [12] M. Tonelli, G. Curhan, M. Pfeffer, F. Sacks, R. Thadhani, et al., Relation between alkaline phosphatase, serum phosphate, and all-cause or cardiovascular mortality, Circulation. 120 (2009) 1784-1792. [13] J.B. Park, D. Kang, H.M. Yang, H.J. Cho, K.W. Park et al., Serum alkaline phosphatase is a predictor of mortality, myocardial infarction, or stent thrombosis after implantation of coronary drug-eluting stent, Eur Heart J. 34 (2012) 920–931. [14] R. Shantouf, C.P. Kovesdy, Y. Kim, N. Ahmadi, A. Luna et al., Association of serum alkaline phosphatase with coronary artery calcification in maintenance hemodialysis patients, Clin J Am Soc Nephrol. 4 (2009) 1106-1014. [15] A.S. Levey, L.A. Stevens, C.H. Schmid, Y. Zhang, A.F. Castro, et al., A New Equation to Estimate Glomerular Filtration Rate, Ann Intern Med. 150 (2009) 604–612. [16] K.A. Lomashvili, P. Garg, S. Narisawa, J.L. Millan, W.C. O'Neill, Upregulation of alkaline phosphatase and pyrophosphate hydrolysis: Potential mechanism for uremic vascular calcification, Kidney Int. 73 (2008) 1024–1030. [17] R. Villa-Bellosta, V. Sorribas, Calcium Phosphate Deposition With Normal Phosphate Concentration-Role of Pyrophosphate, Circ J. 75 (2011) 2705-2710. [18] L. Osman, M.H. Yacoub, N. Latif, M. Amrani, A.H. Chester, Role of human valve interstitial cells in valve calcification and their response to atorvastatin, Circulation. 114 (2006) 547-552.
14
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
[19] B. Wu, S. Elmariah, F.S. Kaplan, G. Cheng, E.R. Mohler, Paradoxical effects of statins on aortic valve myofibroblasts and osteoblasts implications for end-stage valvular heart disease, Arterioscler Thromb Vasc Biol. 25 (2005) 592-597. [20] T.Q Callister, P. Raggi, B. Cooil, N.J. Lippolis, D.J. Russo, Effect of HMG-CoA reductase inhibitors on coronary artery disease as assessed by electron-beam computed tomography, N Engl J Med. 339 (1998) 1972-1978. [21] M.J. Budoff, K.L. Lane, H. Bakhsheshi, S. Mao, B.O. Grassmann et al., Rates of progression of coronary calcium by electron beam tomography, Am J Cardiol. 86 (2000) 811. [22] S. Achenbach, D. Ropers, K. Pohle, A. Leber, C. Thilo et al., Influence of lipid-lowering therapy on the progression of coronary artery calcification: a prospective evaluation, Circulation. 106 (2002) 1077-1082. [23] M. Henein, G. Granåsen, U. Wiklund, A. Schmermund, A. Guerci et al., High dose and long-term statin therapy accelerate coronary artery calcification. Int J Cardiol. 184 (2015) 581-586. [24] R. Puri, S.J. Nicholls, M. Shao, Y. Kataoka, K. Uno et al., Impact of statins on serial coronary calcification during atheroma progression and regression, J Am Coll Cardiol. 65 (2015) 1273-1282. [25] J.K. Williams, G.K. Sukhova, D.M. Herrington, P. Libby, Pravastatin has cholesterollowering independent effects on the artery wall of atherosclerotic monkeys, J Am Coll Cardiol. 31 (1998) 684-691. [26] I. Dykun, N. Lehmann, H. Kälsch, S. Möhlenkamp, S. Moebus et al., Statin Medication Enhances Progression of Coronary Artery Calcification: The Heinz Nixdorf Recall Study, J Am Coll Cardiol. 68 (2016) 2123-2125. [27] K. Imoto, T. Hiro, T. Fujii, A. Murashige, Y. Fukumoto, et al. Longitudinal structural determinants of atherosclerotic plaque vulnerability: a computational analysis of stress distribution using vessel models and three-dimensional intravascular ultrasound imaging, J Am Coll Cardiol. 46 (2005) 1507-1515. [28] M.V. Madhavan, M. Tarigopula, G.S. Mintz, A. Maehara, G.W. Stone, et al. Coronary artery calcification: pathogenesis and prognostic implications, J Am Coll Cardiol. 63 (2014) 1703-1714. [29] S. Elmariah, J.A. Delaney, K.D. O'Brien, M.J. Budoff, J. Vogel-Claussen, et al., Bisphosphonate Use and Prevalence of Valvular and Vascular Calcification in Women MESA (The Multi-Ethnic Study of Atherosclerosis), J Am Coll Cardiol. 56 (2010) 1752-1759. [30] T. Kawahara, M. Nishikawa, C. Kawahara, T. Inazu, K. Sakai, et al., Atorvastatin, Etidronate, or Both in Patients at High Risk for Atherosclerotic Aortic Plaques: A Randomized Controlled Trial, Circulation. 127 (2013) 2327-2335. [31] S. Narisawa, D. Harmey, M.C. Yadav, W.C. O'Neill, M.F. Hoylaerts et al., Novel inhibitors of alkaline phosphatase suppress vascular smooth muscle cell calcification, J Bone Miner Res. 22 (2007) 1700-1710.
15
ACCEPTED MANUSCRIPT TABLES Table 1. Characteristics of each tertile of ALP and CACS distribution. Mean
SD
ALP <55 IU/L
171 (34.2)
45.7
6.6
ALP 55-66 IU/L
167 (33.4)
60.2
3.5
ALP >66 IU/L
162 (32.4)
83.1
19.0
CACS = 0
187 (37.4)
-
CACS > 0
313 (62.6)
257
CACS 1-100
169 (33.8)
29
CACS 101-400
92 (18.4)
209
CACS ≥401
52 (10.4)
1058
RI PT
n (%)
-
588 30
89
SC
Variable
1100
AC C
EP
TE D
M AN U
CACS, coronary artery calcium score; ALP, serum alkaline phosphatase; SD, standard deviation.
ACCEPTED MANUSCRIPT Table 2. Baseline characteristics according to CACS. Variables
All patients
CACS=0
CACS>0
p
(n=500)
(n=187)
(n=313)
60.9 (10.8)
55.3 (11.1)
64.3 (9.2)
0.001
Woman (%)
49.6
62.6
41.1
0.001
Smokers (%)
13.0
14.0
12.7
0.68
Hypertension (%)
28.8
18.7
34.6
0.001
AC C
EP
TE D
M AN U
SC
Age, mean (SD), years
RI PT
Clinical
6.9
2.7
9.5
0.004
Dyslipidemia (%)
70.7
66.8
73.1
0.14
BMI, mean (SD), kg/m²
25.5 (4.7)
25.5 (5.4)
25.5 (4.3)
0.96
SBP, mean (SD), mmHg
136 (17)
131 (15)
138 (17)
0.001
DBP, mean (SD), mmHg
82 (12)
81 (11)
83 (12)
0.14
37.2
26.2
43.7
0.001
81.6 (13.9)
84.1 (14.3)
80.2 (13.5)
0.003
Total cholesterol, mean (SD), mmol/L
2.22 (0.47)
2.26 (0.44)
2.20 (0.49)
0.19
Triglycerides, mean (SD), mmol/L
1.27 (1.03)
HDL cholesterol, mean (SD), mmol/L
0.59 (0.18)
LDL cholesterol, mean (SD), mmol/L
1.38 (0.44)
Statin (%) Biological Glomerular filtration rate, mean (SD), ml/mn/1.73 m2
ALP, mean (SD), IU/L a
1.29 (1.27)
1.27 (0.86)
0.44 a
0.61 (0.19)
0.58 (0.18)
0.06
1.40 (0.42)
1.38 (0.45)
0.56
2.1 (2.9)
2.0 (2.8)
2.2 (3.0)
0.25a
62.6 (19.3)
59.0 (17)
64.7 (20)
0.001a
M AN U
CRP, mean (SD), mg/L
RI PT
Diabetes (%)
SC
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Tesr on log transformed data. CACS,coronary artery calcium score; SBP,systolic blood pressure; DBP,diastolic blood pressure; CRP,C reactive protein; ALP,serum alkaline phosphatase. Data are presented as % or mean (standard deviation).
ACCEPTED MANUSCRIPT Table 3. Logistic regression for prediction of CACS>0 in 500 primary prevention patients. 95% CIa
pa
Age, years
1.12
1.08 ; 1.15
0.001
Men
3.84
2.39 ; 6.15
0.001
Diabetes
2.93
0.93 ; 9.24
0.07
Hypertension
1.06
0.62 ; 1.79
0.84
Current vs. no smokers
1.61
0.84 ; 3.06
0.15
Glomerular filtration rate, ml/mn/1.73 m2
1.02
0.99 ; 1.03
0.10
Statin
0.49
0.23 ; 1.05
0.07
ALP 55-66 vs. <55 IU/L
2.11
1.12 ; 3.96
0.02
ALP >66 vs. <55IU/L
3.89
2.01 ; 7.54
0.001
M AN U
SC
RI PT
ORa
Variables
a
AC C
EP
TE D
With interaction between statin and ALP. For ALP p for trend=0.001 in full sample and in subsample without statin analyses. ALP,serum alkaline phosphatase; OR,odds ratio; CI,confidence interval.
Table 4. Logistic regression for prediction of CACS>0 in patients with or without statin treatment. ACCEPTED MANUSCRIPT Patients without statin (n=314) Variables
Patients with statin (n=186)
95% CI
p
OR
95% CI
p
Age, years
1.13
1.09 ; 1.17
0.001
1.10
1.04 ; 1.15
0.001
Men
3.68
2.06 ; 6.59
0.001
4.47
1.94 ; 10.3
0.001
Diabetes
3.25
0.64 ; 16.3
0.16
2.74
0.54 ; 13.8
0.22
Hypertension
1.38
0.70 ; 2.72
0.35
0.77
0.32 ; 1.87
0.56
Current vs. no smokers
1.31
0.60 ; 2.87
0.51
2.52
0.74 ; 8.65
0.14
1.02
0.99 ; 1.04
0.11
1.01
0.98 ; 1.04
0.60
ALP 55-66 vs. <55 IU/L
2.19
1.14 ; 4.19
0.018
1.56
0.59 ; 4.10
0.37
ALP >66 vs. <55IU/L
3.87
1.96 ; 7.61
0.58
0.24 ; 1.45
0.25
ml/mn/1.73 m2
SC
Glomerular filtration rate,
RI PT
OR
0.001
AC C
EP
TE D
M AN U
ALP,serum alkaline phosphatase; OR,odds ratio; CI,confidence interval.
ACCEPTED MANUSCRIPT HIGHLIGHTS •
Alkaline phosphatase is an independent predictor of coronary artery calcification.
•
Statins have an impact on alkaline phosphatase activity.
•
The pyrophosphate pathway could be a target for therapies against vascular
AC C
EP
TE D
M AN U
SC
RI PT
calcification.