Site and Size of Vascular Calcifications Are Different in Dialysis Patients with Various Underlying Diseases

Computed tomography (CT) constitutes the gold standard for quantification of vascular calcification (VC) and, being the most effective and widely available with reproducible measurements, is also useful for monitoring progression as well as assessing the effect of therapeutic strategies to modify progression [1] [2]. VC has a significant effect in cardio‐ vascular diseases on dialysis patients. Tanne et al. [3] focused on calcification of the thoracic aorta and found that it associated with coronary and valvular calcification in hy‐ pertensive patients. In the Calcification Outcome in Renal Disease (CORD) study, ab‐ dominal aortic calcification was found to have the predictive value for the occurrence of cardiovascular events and mortality in dialysis patients [4]. Coll et al. [5] reported that VC in large, conduit arteries was prevalent in patients on dialysis patients, and that age, dialysis vintage, past medical history of cardiovascular disease, atherosclerosis and in‐ flammation were variable significantly influencing VC. From these studies, it is suggest‐ ed that VC occurs in vessels of various diameters. However, no definitive studies have determined the significance of VC in different vessels in patients receiving dialysis thera‐ py until the present time. Moreover, there have been few studies examining a relation between semi-quantitative measures of VC and their contributing factors. The aim of this work presented here is to examine a relation between semi quantitatively measured calci‐ fication of three major vessels, the thoracic aorta, the abdominal aorta and the iliac arter‐ ies and several known contributing factors to VC such as underlying diseases, age, gender, vintage of dialysis, values of serum calcium and phosphate, use of calciumbased phosphate binders and so on.


Introduction
Computed tomography (CT) constitutes the gold standard for quantification of vascular calcification (VC) and, being the most effective and widely available with reproducible measurements, is also useful for monitoring progression as well as assessing the effect of therapeutic strategies to modify progression [1] [2]. VC has a significant effect in cardiovascular diseases on dialysis patients. Tanne et al. [3] focused on calcification of the thoracic aorta and found that it associated with coronary and valvular calcification in hypertensive patients. In the Calcification Outcome in Renal Disease (CORD) study, abdominal aortic calcification was found to have the predictive value for the occurrence of cardiovascular events and mortality in dialysis patients [4]. Coll et al. [5] reported that VC in large, conduit arteries was prevalent in patients on dialysis patients, and that age, dialysis vintage, past medical history of cardiovascular disease, atherosclerosis and inflammation were variable significantly influencing VC. From these studies, it is suggested that VC occurs in vessels of various diameters. However, no definitive studies have determined the significance of VC in different vessels in patients receiving dialysis therapy until the present time. Moreover, there have been few studies examining a relation between semi-quantitative measures of VC and their contributing factors. The aim of this work presented here is to examine a relation between semi quantitatively measured calcification of three major vessels, the thoracic aorta, the abdominal aorta and the iliac arteries and several known contributing factors to VC such as underlying diseases, age, gender, vintage of dialysis, values of serum calcium and phosphate, use of calciumbased phosphate binders and so on.

Methods
All HD patients received three dialysis sessions of at least 4 h duration per week. HD was performed using low flux polysalphose dialyhsers (1.5-2.0 m2 APS Asahi Medical R Tokyo, Japan). All HD patients were dialyzed using bicarbonate-bound 1.25 mmol/L, calcium and 134 mmol/sodium containing dialysate. Patients were all dialysed at the Dialysis Unit of Irumadai Hospital.
This was an observational and cross-sectional study that included 79 hemodialysis patients at the Dialysis Unit of Irumadai Hospital, who gave their informed consent to enroll in this study. The inclusion criteria were patient providing informed consent, age ≥ 40 years and duration of dialysis ≥ 1 year. Exclusion criteria were significant fetal diseases that were estimated to reduce life expectancy to < 6 months and patients in whom it was impossible to measure CT scan.
The recorded cardiovascular history and smoking status were obtained. The following baseline biochemical data were obtained; serum calcium, phosphorus, intact parathyroid hormone, albumin, total cholesterol, low-density lipoprotein cholesterol. Data on weight, height, body mass index and duration of dialysis and use of medications: phosphate binders, vitamin D, statins, erythropoietin and antihypertensive agents. Clinical characteristics and laboratory variables including dual-energy x-ray absorptiometry and pulse wave velocity. This study complies with the Declaration of Helsinki and was in agreement with the guidelines approved by the ethics committee at the institution.

Computed tomography
CT scan of the aorta and arteries was performed with a 16-detector CT scan {Prime Purpose MDCT (GE Healthcare, Milwaukee, WI USA)}. Scanning time was 0.5 s for two contiguous 1.25 mm sections and 20±5 seconds for the entire zone of interest. Examination was performed during a single, unforced, withheld inspiration. During scanning with the tube rotating at 2 rotation/second and the table moving at 55 mm/s with a 1:1.375 scanning pitch, images were obtained with an effective section thickness of 10 mm. Scanning was performed with 120 kVp and 350 mAs, standard resolution, and a 28-36 cm field of view. The total duration of the procedure was 5min. The range of CT scan was illustrated in Fig. 1.

Evaluation of thoracic and abdominal aorta and iliac artery
Volume acquisitions were analyzed using Volume Viewer software (GE Healthcare). The thoracic and abdominal aorta were segmented manually. In order to reduce errors due to noise, a cut-off of 130 Housefield Unit (HU) was applied. The total calcification volume was calculated as the sum of all voxels in the remaining volume.

Biochemistry
Blood samples were collected at monthly intervals. The results presented here were timeaveraged results from the preceding 6 months prior to the CT scan.

Blood pressure
Three blood pressure (BP) recording were taken suing automated device.

Statistics
Data are expressed as means ± SD. Using variables found to be significant in the univariate analysis and potential confounders, we applied forward stepwise logistic regressions, in order to determine which of these variables were most significantly associated with calcification of the thoracic and abdominal aorta and the arteries of the lower limbs. F-to-Remove was set at 2.9. P<0.05 was considered as significance.
Baseline demographics and laboratory and hemodynamic values of the study population are shown in Tables 1, 2 and 3 and current medications used are listed in Table 4.

Calcification of vessels
In Table 5, the average of calcification scores is shown.
All three lesions correlated significantly with each other. Stepwise regression was applied in which the independent variables were identified from the univariate analyses. Significant associations were seen for the following: the prevalence of calcification; the thoracic aorta with period of dialysis, elevations of both systolic and diastolic blood pressure and levels of serum albumin (Table 6); in the abdominal aorta with age, presence of diabetes, and calcium supplement ( Table 7); arteries of the lower limbs with presence of diabetes mellitus, use of sevelamer and cinacalcet and serum levels of intact parathyroid hormone and albumin (Table 8). Table 5. Calcification scores of thoracic aorta, abdominal aorta and iliac artery SBP: systolic blood pressuer, DBP: diastolic blood pressure

Discussion
In the present study, we found that the contributing factors to VC were different in the different vessels. The development and progression of VC is a multifactorial process. Potentially differing factors may exert their maximum influence at either the predisposition, initiation and continuation phases of the process. The multivariate analysis performed on these data attempt to elucidate which factors might be most significant to the development of VC. In the present study, age, duration of HD, systolic and diastolic BP, presence of DM serum levels of Ca, intact PTH, calcium modulating drugs and albumin contributed differently in the different vessels. Albumin was negatively correlated with the severity of VC. This suggests that a characteristic state of low albumin as seen in malnutrition, inflammation or atherosclerosis complex is most important, as suggested by Wang et al. [6]. Factors shown to predict VC in the current study included older age, longer dialysis vintage, diabetes, higher concentrations of serum phosphorus and calcium are associated with more extensive VC among patients on HD and result partially consistent with those reported previously [7] [8] [9] [10].
Adler et al [11] demonstrated a strong association of coronary calcification and calcification of the thoracic aorta on spiral CT. The aortic calcification signifies a higher probability of coronary atherosclerosis and ischemic stroke (Cerebrovascular disease). Also, Tanne et al. [3] found that severe calcification in descending aorta is a predictor of ischemic cerebrovascular events. Calcification of the thoracic aorta is not a direct causative factor for embolic stroke, but rather a marker of increased burden of vascular (atherosclerotic disease) disease [12]. However, Honkanen et al. [4] reported that although the duration of HD correlates with calcification in coronary [1], carotid and peripheral arteries [7], the association is less clear in the thoracic arteries [8].
In the present study, calcification of the thoracic aorta had a strong association with dialysis vintage, systolic and diastolic BP and albumin, which are a major factors contributing to cardiovascular diseases. From these data, it is possible that severe calcification of the thoracic aorta is produced by hemodynamic, malnutrition and uremia in combination.
Abdominal aorta calcification has been well studied, has been associated with an increase risk of cardiovascular morbidity and mortality in patients with HD [13]. Hanada et al. [14] proposed that the section of the aorta chosen for measuring the semiquantitative calcification score is suitable for evaluation of the severity of VC because the site is associated with turbulent flow and is susceptible to development of atheroma. The chosen site is also simple to investigate radiologically since it is in a significant part of the aorta and is vertical to the transverse section. In the present study, VC of the abdominal aorta was correlated with the presence of diabetes, which is a well-known atherosclerotic risk factor. In addition, the factors relating with calcium-phsophate modulation, such as concentrations of calcium, PTH and so on are frequently evoked as the principal causes associated with vascular remodeling and/or arterial calcifications [15] [16]. Guerin et al. [17] reported that in HD patients, there is an association between the presence of aortic calcification and increased Ca x P products. In contrast, Arad et al. did not find the serum concentrations of calcium, 1,25-Vit D, and PTH to be associated with the presence of arterial calcifications [18]. Besides, the amount of CaO 3 prescribed as a phosphate binder was independently associated with the score of vascular calcifications. One of the adverse effects of calcium based phosphate binders is hypercalcemia, which may in turn result in arterial calcification. It is therefore likely that development of VC of the abdominal aorta is associated with calcium and phosphorus regulation in HD patients. Moreover, mineral bone disease-related factors such as serum calcium, phosphorus and PTH are thought to be strongly associated with the severity of VC in dialysis patients [19] [20].
Sigrist et al. [21] described a simple, sensitive low radiation dose technique as an alternative to coronary artery and aortic measurements to quantify a calcification score for the superficial femoral artery (SFA). The sector of artery chosen for this study is ideal as it avoids major bifurcations and arterial branching, and therefore, obvious site for turbulent flow and the development of atheroma. In the present study, factors contributing to VC of the iliac arteries are similar with those of the abdominal aorta.
In the Calcification Outcome in Renal Disease (CORD) study, 19% of patients had no visible calcification in their abdominal aorta [4]. These findings are partially in line with certain previous observations and it has been suggested that these individuals rarely develop calcification at follow-up [22] [8] [23]. In the present study, we did not find these individuals.
Recently, further reports from CORD study provided a new evidence that no coronary [24] or thoracic aortic calcification at baseline, but their calcification developed during 2 years of observation and was most prevalent in those receiving calcium-containing binders. Besides, retrospective and cross-sectional data have given contradicting results with some publication showing a contribution of Vit D to VC [15], whereas others do not support this contention [25]. It is therefore unlikely that HD patients receiving calcium-containing binders and Vitamin D analogues have no VC of the vessels.
Recently Allison et al. [26] demonstrated that in terms of extent of calcification, the iliac arteries showed the strongest association for all mortality and end points, consistent with the well-known association between the severity of peripheral artery disease and both CVD and total mortality [27].
In addition, they concluded [26] that higher levels of calcium in different vascular beds are associated not only with CVD mortality but also with non-CVD and total mortality and that location of the arterial calcification appears to be relevant to the strength of the association with mortality, and the CVD risk factors appear to mediate some of this association.

Study limitations
First, the imaging methods used in this study did not distinguish the two types of VC (pathy calcification of the intima and calcification of the media). As is known, mineral metabolism disturbances link specifically with medial rather than intimal V and intima calcification associates with atherosclerosis. Second, our studies was cross-sectional, it does not directly show how detection of VC in various vessels predict incident cardiovascular events in the dialysis patients. Third, VC represents the result of long-standing atherosclerotic and calcification processes. It is unclear whether the steady-state of serum chemistry such as calcium, phosphate, intact PTH concentrations measured in this study accurately represents pathological process that occurred when VC was developing.

Conclusion
Presence and extension of VC in thoracic and abdominal aortas and lower limbs might be regulated in complex manner and caution should be needed to use these variables as a marker of the burden of vascular disease. The associations between calcified atherosclerosis and mortality differ by vascular bed, suggesting that the location and severity of calcification in different vascular beds provide unique information for mortality.