Abstract

Aims

Emerging data have indicated unexpected complexity in the regulation of vascular and bone calcification. In particular, several recent studies have challenged the concept of a universally positive relationship between body morphology [weight, height, body mass index (BMI), body surface area (BSA)] and the extent of vascular calcification. We sought to clarify these discrepancies and investigated the relationship between index lesion coronary artery calcification (CAC) and body morphology in patients undergoing percutaneous coronary intervention (PCI) using three-dimensional intravascular ultrasound (IVUS).

Methods and results

We analysed CAC in patients who underwent PCI with pre-intervention IVUS imaging. The main outcome measure was the calcium index (CalcIndex); a three-dimensional IVUS-derived measure of total calcification per obstructive coronary lesion. A total of 346 patients (65.3 ± 10.6 years; 29.5% females) underwent PCI with IVUS-based CAC assessment. CalcIndex was categorized as zero–low (0–0.1399; n = 152) or intermediate–high (0.1400–1.2541; n = 194). All measures of body morphology were lower in patients with intermediate–high CalcIndex (height, P = 0.024; weight, P = 0.008; BMI, P = 0.064; BSA, P = 0.005). In adjusted multivariable models, weight and BSA were independent inverse predictors of intermediate–high CalcIndex [weight: odds ratio (OR) 0.986, P = 0.017; BSA: OR 0.323, P = 0.012] while CalcIndex also trended towards an inverse association with both height (P = 0.068) and BMI (P = 0.064). These independent inverse associations were consistent across multiple clinical subgroups, including stratification by age, race, gender, diabetes, and renal impairment.

Conclusion

Using three-dimensional IVUS to assess vascular calcification, these data confirm an independent, inverse relationship between body size and index lesion CAC in patients with obstructive coronary artery disease.

Introduction

Disturbances in mineral metabolism are common in older adults1 and are associated with adverse cardiovascular outcomes.1,2 As an example, low levels of serum 25-hydroxyvitamin D are associated with decreased bone mineral density1 while at the same time being associated with incident myocardial infarction and increased all-cause mortality.2,3 Of importance, low bone mineral density is an independent risk factor for cardiovascular events.4

In older persons, it is known that an inverse relationship exists between bone mineral density and vascular calcification, such that lower bone mineral density is associated with increased vascular calcification.5–8 This recently described relationship has been dubbed the ‘vascular calcification paradox’. Furthermore, after middle age, a robust positive association exists between bone mineral density and various measures of body morphology such as body weight, body mass index (BMI), and body surface area (BSA).9–11 These relationships indicate profound complexity and interplay between the physiological pathways of vascular and skeletal ossification, and indirectly suggest that there may be an age-dependent component in their regulation. Furthermore, while numerous studies have suggested that a positive relationship exists between the extent of vascular calcification and body morphology, a number of very recent studies have suggested that this may be an overly simplistic outlook and that in certain populations this relationship may not hold true.12–17 In particular, isolated studies have now suggested that in older women13 or those with clinically significant coronary disease,17 there may be an inverse relationship between body size and vascular calcification.

Given these uncertainties, we sought to rigorously explore the relationship between index lesion coronary artery calcification (CAC) and body morphology in patients with coronary artery disease undergoing percutaneous coronary intervention (PCI), using an independent data set and performing quantitative CAC assessment using three-dimensional intravascular ultrasound (IVUS) imaging.

Methods

We analysed 346 consecutive patients who underwent PCI with pre-intervention IVUS imaging. Detailed characteristics of these patients and the PCI procedure have been published previously.18 Only a single culprit target lesion and associated target vessel per patient were included in this study. IVUS of other vessels was not clinically indicated and was not performed. Key enrolment criteria were guideline- appropriate requirement for PCI, typically based on severe disease, positive stress test results, or presentation with unstable coronary syndromes; age 18 years; and signed informed consent. Exclusion criteria were presentation with ST-segment elevation myocardial infarction, serum creatinine >2.0 mg/dL, baseline anaemia (haemoglobin <10.5 g/dL), thrombocytopaenia (platelet count <125 000/mL), prior heart transplantation, active autoimmune disease, illicit drug use, human immunodeficiency virus infection, prior malignancy with mediastinal irradiation, bone marrow transplantation, high-dose chemotherapy, or adult congenital heart disease. PCI procedures were performed according to current standard guidelines, and the type of stent implanted and the use of pharmacological agents were at the discretion of the operator.18 The decision to perform IVUS was also at the discretion of the operator and was made entirely independently of this study, with all IVUS images acquired before this study was conceived.

The following definitions were used: BMI = (weight in kg)/(height in meters)2; BSA was calculated according to the Mosteller formula = ((weight in kg)0.5 × (height)0.5)/60; estimated glomerular filtration rate (eGFR) was calculated according to the modification of diet in renal disease formula, with chronic renal impairment (CRI) defined as an eGFR <60 mL/min/1.73 m2; hypertension = use of anti-hypertensive medication(s) or a prior clinical diagnosis of hypertension documented from the primary referring physician; congestive heart failure = prior clinical diagnosis of New York Heart Association functional class III or IV heart failure, irrespective of left ventricular function; hyperlipidaemia = use of lipid-lowering therapy or a prior clinical diagnosis of dyslipidaemia documented from the primary referring physician; diabetes = use of oral or injected anti-diabetic agents.

Intravascular ultrasound imaging

Target lesion IVUS imaging studies were performed after intracoronary administration of 200 µg nitroglycerine using a commercially available IVUS system (Atlantis SR Pro, 40 MHz catheter, Boston Scientific Corp., Natick, MA, USA; Eagle Eye, 20 MHz catheter or Revolution 45 MHz catheter, Volcano Corp., Rancho Cordova, CA, USA) and prior to PCI. The IVUS catheter was advanced distal to the stenosis, and imaging was performed with retrograde pull-back to the aorto-ostial junction of the coronary artery using an automatic pullback system at a speed of 0.5 mm/s. All analyses were performed offline without knowledge of patient body morphology using planimetry software (INDEC Systems Inc., Mountain View, CA, USA). The minimum lumen cross-sectional area (CSA) site was the image slice with the smallest lumen CSA. The reference sites were the most normal-appearing cross sections within 5 mm proximal and distal to the lesion but before any side branch and were used to calculate a mean reference. For each patient, the lesion with the smallest lumen CSA was chosen for analysis. The lesion itself was defined as the segment between the proximal and distal reference sites whose length (mm) was calculated using the pullback duration and pullback speed. Quantitative analysis included the measurement of the external elastic membrane (EEM) and lumen CSA every 1 mm within the length of the lesion. Plaque and media CSA were calculated as EEM − lumen CSA. Once a complete set of CSA measurements were obtained, EEM, plaque + media, and lumen volumes were calculated using Simpson's rule. Plaque burden was calculated as plaque and media divided by EEM volume or CSA. A remodelling index was calculated as the lesion divided by the mean reference EEM CSA.18 Following two-dimensional analysis, high-resolution images of each IVUS pull-back were imported into dedicated software for three-dimensional reconstruction and analysis. Calcium was identified as an echo signal brighter than the adventitia with acoustic shadowing. The maximum arc of calcium (°) within the lesion was measured with the electronic protractor centred on the lumen. Calcium length (mm) within the lesion was measured as the length of the lesion in which there was IVUS-detectable calcium. Calcium index (CalcIndex) was calculated as total calcium length/lesion length × maximum calcium arc/360° (Figure 1).18 CalcIndex represents the total amount of calcium per coronary lesion and was therefore selected a priori as the primary dependent outcome for this study. The use of three-dimensional IVUS to quantify arterial calcification has been validated by Virmani and co-workers and exhibits a high degree of correlation with ex vivo histopathological calcification assessment.19 The in vivo human application of this methodology has previously been reported by ourselves18 and others.20

Figure 1

Derivation of Calcium Index (CalcIndex) by IVUS. (A) In this longitudinal IVUS reconstruction, the lesion measures 27.5 mm in length from the proximal to the distal reference site and contains two lengths of diffuse calcification (note acoustic shadows) that are 3 and 18 mm in length (white horizontal arrows). The vertical white arrow indicates the point of maximum calcium arc. (B) Cross-sectional image taken from the same IVUS scan showing the point of maximum calcium arc, measuring 250°. The calculation of CalcIndex is shown and was 0.53 in this case.18

Figure 1

Derivation of Calcium Index (CalcIndex) by IVUS. (A) In this longitudinal IVUS reconstruction, the lesion measures 27.5 mm in length from the proximal to the distal reference site and contains two lengths of diffuse calcification (note acoustic shadows) that are 3 and 18 mm in length (white horizontal arrows). The vertical white arrow indicates the point of maximum calcium arc. (B) Cross-sectional image taken from the same IVUS scan showing the point of maximum calcium arc, measuring 250°. The calculation of CalcIndex is shown and was 0.53 in this case.18

Statistical analyses

Continuous variables are expressed as mean and SD as indicated and compared using Student's t-test or Wilcoxon rank-sum test if applicable. Discrete variables are presented as numbers and percentages and compared with the χ2 test, unless the observation in any cell was <5, in which case Fisher's exact test was used. We initially examined the distribution of CalcIndex and consistent with prior reports found it to be positively skewed,21 with 55/346 (15.9%) patients having a value of zero. This distribution pattern failed to fulfil the principal assumptions that underpin the classical linear regression model: (i) The relationship between dependent and independent variables was non-linear, (ii) The error distribution was not normal, and (iii) The errors were heteroscedastic. For this reason and consistent also with prior studies examining associations of vascular calcification,17,21 CalcIndex was categorized as zero–low (0–0.1399; n = 152) or intermediate–high (0.1400–1.2541; n = 194). The independent associations between various parameters of body morphology (height, weight, BMI, and BSA) and CalcIndex were then assessed using binary logistic regression. Separate models were generated for each parameter of body morphology as the exposure of interest with intermediate–high CalcIndex as the dependent outcome. All fully adjusted models included the following candidate covariates selected using a stepwise algorithm with entry/exit criteria of 0.1/0.1: gender, age, race, current smoking, former smoking, hypertension, hyperlipidaemia, diabetes, CRI, and prior congestive heart failure. The associations between intermediate–high CalcIndex and both BSA and weight were also assessed separately in subgroups defined by: age, gender, race, diabetes, chronic renal impairment, smoking status, hyperlipidaemia, and hypertension. Current and former smokers were combined for this analysis as ‘smoking’. These models included stratification according to age, gender, renal impairment, and race/ethnicity. Interaction testing between each subgroup and BSA or weight on the outcome of intermediate–high CalcIndex was also performed. Statistical analyses were performed using the SAS software, version 9.2. (Cary, NC, USA).

Results

A total of 346 consecutive patients are included in this study who underwent index lesion PCI and pre-PCI three-dimensional IVUS assessment with offline computation of CalcIndex (Figure 1). The mean age was 65.3 ± 10.6 years (range 32.7–87.0 years), with 102 females (29.5%).

Patient demographics are presented in Table 1 and were stratified according to those with zero lesion calcification and tertiles of those with any calcification (low, intermediate, high). There was no significant trend for gender, race, or history of renal impairment among the groups of CalcIndex. Older patients tended to have greater CalcIndex (P = 0.08). Consistent with prior observations,17 a trend was noted for an association between a history of prior myocardial infarction and lower CalcIndex (P = 0.004). Lack of baseline aspirin therapy was associated with higher CalcIndex (P = 0.03), whereas the use of β-blockers was associated with increasing CalcIndex (P = 0.05).

Table 1

Baseline demographic characteristics according to groupings of no calcification and tertiles of low, intermediate, and high CalcIndex

 Zero CalcIndex (no calcium), n = 55 Low CalcIndex (0.001–0.1399), n = 97 Intermediate CalcIndex (0.1400–0.3759), n = 97 High CalcIndex (0.3760–1.2541), n = 97 P-value (ANOVA) P-value (trend) 
Age/gender/race 
 Age (years)a 63.0 ± 9.8 65.2 ± 9.8 65.3 ± 11.7 66.4 ± 10.8 0.303 0.080 
 Female gender 32.7% (18) 22.7% (22) 30.9% (30) 33.0% (32) 0.375 0.470 
 Race/ethnicity       
  White 68.0% (34/50) 65.2% (60/92) 64.8% (57/88) 62.1% (54/87) 0.881 0.435 
  Black 14.0% (7/50) 9.8% (9/92) 9.1% (8/88) 11.5% (10/87) 0.818 0.744 
  Hispanic 12.0% (6/50) 9.8% (9/92) 7.9% (7/88) 10.3% (9/87) 0.889 0.736 
  Other 6.0% (3/50) 15.2% (14/92) 18.2% (16/88) 16.1% (14/87) 0.265 0.158 
Body morphology 
 Height (cm) 171.4 ± 9.0 172.6 ± 10.5 169.7 ± 9.9 169.8 ± 9.9 0.128 0.084 
 Weight (kg) 86.9 ± 17.6 88.1 ± 19.7 80.6 ± 19.1 83.8 ± 18.7 0.038 0.072 
 BMI (kg/m229.59 ± 5.46 29.49 ± 5.86 27.89 ± 5.50 28.93 ± 5.29 0.163 0.229 
 BSA (m22.03 ± 0.23 2.04 ± 0.26 1.94 ± 0.26 1.98 ± 0.26 0.026 0.049 
History and risk factors 
 Prior MI 29.09% (16/55) 27.84% (27/97) 19.59% (19/97) 12.37% (12/97) 0.027 0.004 
 Prior stroke/TIA 5.45% (3/55) 6.19% (6/97) 7.22% (7/97) 7.22% (7/97) 0.967 0.634 
 Congestive heart failure 7.27% (4/55) 9.28% (9/97) 6.19% (6/97) 7.22% (7/97) 0.874 0.736 
 Hypertension 81.82% (45/55) 77.32% (75/97) 76.29% (74/97) 79.38% (77/97) 0.860 0.818 
 Hyperlipidaemia 81.82% (45/55) 84.54% (82/97) 78.35% (76/97) 79.38% (77/97) 0.702 0.435 
 Diabetes 34.55% (19/55) 29.90% (29/97) 36.08% (35/97) 28.87% (28/97) 0.673 0.683 
 Chronic renal impairment 9.09% (5/55) 6.19% (6/97) 13.40% (13/97) 8.25% (8/97) 0.362 0.70 
 Current smoking 9.09% (5/55) 7.22% (7/97) 13.40% (13/97) 10.31% (10/97) 0.549 0.483 
 Former smoking 20.00% (11/55) 19.59% (19/97) 14.43% (14/97) 10.31% (10/97) 0.247 0.052 
Medications 
 ASA 90.91% (50/55) 92.63% (88/95) 83.51% (81/97) 82.47% (80/97) 0.103 0.033 
 Plavix 63.64% (35/55) 70.53% (67/95) 46.39% (45/97) 61.86% (60/97) 0.006 0.208 
 β-Blockers 59.62% (31/52) 64.21% (61/95) 67.71% (65/96) 74.23% (72/97) 0.269 0.050 
 Statins 67.92% (36/53) 70.53% (67/95) 71.88% (69/96) 72.16% (70/97) 0.950 0.583 
 ACE-I 35.85% (19/53) 29.47% (28/95) 36.46% (35/96) 39.18% (38/97) 0.549 0.351 
 Zero CalcIndex (no calcium), n = 55 Low CalcIndex (0.001–0.1399), n = 97 Intermediate CalcIndex (0.1400–0.3759), n = 97 High CalcIndex (0.3760–1.2541), n = 97 P-value (ANOVA) P-value (trend) 
Age/gender/race 
 Age (years)a 63.0 ± 9.8 65.2 ± 9.8 65.3 ± 11.7 66.4 ± 10.8 0.303 0.080 
 Female gender 32.7% (18) 22.7% (22) 30.9% (30) 33.0% (32) 0.375 0.470 
 Race/ethnicity       
  White 68.0% (34/50) 65.2% (60/92) 64.8% (57/88) 62.1% (54/87) 0.881 0.435 
  Black 14.0% (7/50) 9.8% (9/92) 9.1% (8/88) 11.5% (10/87) 0.818 0.744 
  Hispanic 12.0% (6/50) 9.8% (9/92) 7.9% (7/88) 10.3% (9/87) 0.889 0.736 
  Other 6.0% (3/50) 15.2% (14/92) 18.2% (16/88) 16.1% (14/87) 0.265 0.158 
Body morphology 
 Height (cm) 171.4 ± 9.0 172.6 ± 10.5 169.7 ± 9.9 169.8 ± 9.9 0.128 0.084 
 Weight (kg) 86.9 ± 17.6 88.1 ± 19.7 80.6 ± 19.1 83.8 ± 18.7 0.038 0.072 
 BMI (kg/m229.59 ± 5.46 29.49 ± 5.86 27.89 ± 5.50 28.93 ± 5.29 0.163 0.229 
 BSA (m22.03 ± 0.23 2.04 ± 0.26 1.94 ± 0.26 1.98 ± 0.26 0.026 0.049 
History and risk factors 
 Prior MI 29.09% (16/55) 27.84% (27/97) 19.59% (19/97) 12.37% (12/97) 0.027 0.004 
 Prior stroke/TIA 5.45% (3/55) 6.19% (6/97) 7.22% (7/97) 7.22% (7/97) 0.967 0.634 
 Congestive heart failure 7.27% (4/55) 9.28% (9/97) 6.19% (6/97) 7.22% (7/97) 0.874 0.736 
 Hypertension 81.82% (45/55) 77.32% (75/97) 76.29% (74/97) 79.38% (77/97) 0.860 0.818 
 Hyperlipidaemia 81.82% (45/55) 84.54% (82/97) 78.35% (76/97) 79.38% (77/97) 0.702 0.435 
 Diabetes 34.55% (19/55) 29.90% (29/97) 36.08% (35/97) 28.87% (28/97) 0.673 0.683 
 Chronic renal impairment 9.09% (5/55) 6.19% (6/97) 13.40% (13/97) 8.25% (8/97) 0.362 0.70 
 Current smoking 9.09% (5/55) 7.22% (7/97) 13.40% (13/97) 10.31% (10/97) 0.549 0.483 
 Former smoking 20.00% (11/55) 19.59% (19/97) 14.43% (14/97) 10.31% (10/97) 0.247 0.052 
Medications 
 ASA 90.91% (50/55) 92.63% (88/95) 83.51% (81/97) 82.47% (80/97) 0.103 0.033 
 Plavix 63.64% (35/55) 70.53% (67/95) 46.39% (45/97) 61.86% (60/97) 0.006 0.208 
 β-Blockers 59.62% (31/52) 64.21% (61/95) 67.71% (65/96) 74.23% (72/97) 0.269 0.050 
 Statins 67.92% (36/53) 70.53% (67/95) 71.88% (69/96) 72.16% (70/97) 0.950 0.583 
 ACE-I 35.85% (19/53) 29.47% (28/95) 36.46% (35/96) 39.18% (38/97) 0.549 0.351 

aWith respect to age, 27/346 (7.8%) patients were <50 years of age and 59/346 (17.1%) were <55 years of age.

With respect to body morphology, in these crude analyses presented in Table 1, an inverse association was observed between CalcIndex and patient height, weight, and BSA (P = 0.08, 0.07, 0.049, respectively). Consistent with prior reports of arterial calcification,21 we also identified that the distribution of CalcIndex was positively skewed, with 55 patients (15.9%) having no calcification detected by IVUS.

Owing to its known skewed distribution and consistent with prior studies17,21 CalcIndex was classified as a binary variable (zero–low or intermediate–high) (see Section ‘Statistical analyses’). As shown in Figure 2, compared with patients classified as having zero–low CalcIndex, all measures of body morphology were lower in patients with intermediate–high CalcIndex in unadjusted comparisons (height, P = 0.024; weight, P = 0.008; BMI, P = 0.064; BSA, P = 0.005).

Figure 2

Unadjusted comparisons of mean CalcIndex (zero or low vs. intermediate or high) according to patient (A) height, (B) weight, (C) BMI, and (D) BSA. Values shown are mean for each group and 95% confidence interval.

Figure 2

Unadjusted comparisons of mean CalcIndex (zero or low vs. intermediate or high) according to patient (A) height, (B) weight, (C) BMI, and (D) BSA. Values shown are mean for each group and 95% confidence interval.

Lesion characteristics are presented in Table 2. Lesions located in the left anterior descending (LAD) coronary artery comprised 44% of all lesions. With the exception of a small number of lesions imaged in the diagonal branch of the LAD, there was no difference in the distribution of lesions with zero–low vs. intermediate–high CalcIndex according to anatomical location in the coronary tree (Table 2). IVUS imaging analyses revealed that reference vessel and three-dimensional vessel and lesion characteristics were concordant with CalcIndex, with greater CalcIndex being associated with various measures of increasing vessel and lesion size, including greater reference vessel EEM CSA, greater plaque burden and greater lesion length. There was no relationship between CalcIndex and cross-sectional lesion characteristics at the minimal lumen area site (Table 2).

Table 2

Baseline lesion characteristics according to groups of none or low CalcIndex vs. intermediate or high CalcIndex

 Zero or low CalcIndex (0–0.1399), n = 152 Intermediate or high CalcIndex (0.1400–1.2541), n = 194 P-value 
Lesion location 
 Left main 3.3% (5) 1.5% (3) 0.303 
 LAD 42.7% (65) 45.9% (89) 0.587 
 LAD branch (diagonal) 6.6% (10) 1.5% (3) 0.020 
 LCx 13.8% (21) 14.9% (29) 0.878 
 LCx branch (OM, LPL, ramus intermedius, LPDA) 7.9% (12) 7.2% (14) 0.839 
 RCA 22.4% (33) 26.8% (52) 0.315 
 RCA branch (RPL, RPDA, AV continuation) 4.0% (6) 2.1% (4) 0.345 
Intravascular ultrasound 
 Reference sitea 
  EEM CSA (mm211.5 ± 4.3 13.4 ± 4.9 0.0003 
  Luminal CSA (mm26.6 ± 2.2 6.8 ± 2.3 0.3075 
  Plaque burden 0.42 ± 0.10 0.48 ± 0.10 <0.0001 
Minimum luminal area site 
 EEM CSA (mm211.0 ± 4.2 11.7 ± 4.3 0.1403 
 Luminal CSA (mm23.1 ± 1.2 3.2 ± 1.3 0.7081 
 Plaque and media CSA (mm27.8 ± 3.7 8.5 ± 3.7 0.1137 
 Plaque burden 0.70 ± 0.09 0.71 ± 0.09 0.1460 
Three-dimensional IVUS data 
 EEM volume (mm3101.7 ± 73.5 148.3 ± 121.7 <0.0001 
 Luminal volume (mm339.1 ± 25.8 51.7 ± 32.3 0.0001 
 Plaque and media volume (mm362.6 ± 50.3 96.7 ± 91.7 <0.0001 
 Plaque burden 0.59 ± 0.08 0.62 ± 0.08 0.0004 
 Remodelling Index 0.96 ± 0.18 0.88 ± 0.17 0.0001 
 Lesion length (mm)b 8.9 ± 5.9 11.8 ± 9.2 0.0009 
 Maximal calcium arc (°)b 46.4 ± 46.7 199.6 ± 93.5 <0.0001 
 Calcium length (mm)b 2.3 ± 2.9 9.0 ± 6.4 <0.0001 
 Zero or low CalcIndex (0–0.1399), n = 152 Intermediate or high CalcIndex (0.1400–1.2541), n = 194 P-value 
Lesion location 
 Left main 3.3% (5) 1.5% (3) 0.303 
 LAD 42.7% (65) 45.9% (89) 0.587 
 LAD branch (diagonal) 6.6% (10) 1.5% (3) 0.020 
 LCx 13.8% (21) 14.9% (29) 0.878 
 LCx branch (OM, LPL, ramus intermedius, LPDA) 7.9% (12) 7.2% (14) 0.839 
 RCA 22.4% (33) 26.8% (52) 0.315 
 RCA branch (RPL, RPDA, AV continuation) 4.0% (6) 2.1% (4) 0.345 
Intravascular ultrasound 
 Reference sitea 
  EEM CSA (mm211.5 ± 4.3 13.4 ± 4.9 0.0003 
  Luminal CSA (mm26.6 ± 2.2 6.8 ± 2.3 0.3075 
  Plaque burden 0.42 ± 0.10 0.48 ± 0.10 <0.0001 
Minimum luminal area site 
 EEM CSA (mm211.0 ± 4.2 11.7 ± 4.3 0.1403 
 Luminal CSA (mm23.1 ± 1.2 3.2 ± 1.3 0.7081 
 Plaque and media CSA (mm27.8 ± 3.7 8.5 ± 3.7 0.1137 
 Plaque burden 0.70 ± 0.09 0.71 ± 0.09 0.1460 
Three-dimensional IVUS data 
 EEM volume (mm3101.7 ± 73.5 148.3 ± 121.7 <0.0001 
 Luminal volume (mm339.1 ± 25.8 51.7 ± 32.3 0.0001 
 Plaque and media volume (mm362.6 ± 50.3 96.7 ± 91.7 <0.0001 
 Plaque burden 0.59 ± 0.08 0.62 ± 0.08 0.0004 
 Remodelling Index 0.96 ± 0.18 0.88 ± 0.17 0.0001 
 Lesion length (mm)b 8.9 ± 5.9 11.8 ± 9.2 0.0009 
 Maximal calcium arc (°)b 46.4 ± 46.7 199.6 ± 93.5 <0.0001 
 Calcium length (mm)b 2.3 ± 2.9 9.0 ± 6.4 <0.0001 

Values are means ± SD or % (n).

aCalculated as the average of the proximal and distal reference sites.

bThese variables are used to derive the CalcIndex ( = calcium length/lesion length × maximal calcium arc/360°).

Logistic regression was then performed to evaluate the relationships between body morphometric parameters (height, weight, BMI, BSA) and the presence of intermediate–high (vs. zero–low) CalcIndex (Table 3). In both unadjusted and adjusted multivariable regression models, all measures of body morphology displayed an inverse association with CalcIndex. In the adjusted multivariable models, both weight and BSA were significant inverse predictors of intermediate–high CalcIndex (weight: OR 0.986, P = 0.017; BSA: OR 0.323, P = 0.012). While not reaching statistical significance, there was a similar trend towards an inverse association with CalcIndex and both height (P = 0.068) and BMI (P = 0.064) in these multivariable models.

Table 3

Unadjusted, partial, and adjusted multivariable regression models evaluating predictors of intermediate–high CalcIndex with measures of body morphology (height, weight, BMI, BSA) as independent covariates

 Unadjusted
 
Mutually adjusted for gender and age
 
Multivariable regression
 
 Odds ratio P-value Odds ratio P-value Odds ratio P-value 
Height (cm) 0.985 0.025 0.977 0.051 0.979 0.068 
Weight (kg) 0.985 0.009 0.986 0.022 0.986 0.017 
BMI (kg/m20.964 0.066 0.967 0.093 0.963 0.064 
BSA (m20.299 0.005 0.330 0.014 0.323 0.012 
 Unadjusted
 
Mutually adjusted for gender and age
 
Multivariable regression
 
 Odds ratio P-value Odds ratio P-value Odds ratio P-value 
Height (cm) 0.985 0.025 0.977 0.051 0.979 0.068 
Weight (kg) 0.985 0.009 0.986 0.022 0.986 0.017 
BMI (kg/m20.964 0.066 0.967 0.093 0.963 0.064 
BSA (m20.299 0.005 0.330 0.014 0.323 0.012 

Other independent covariates included in the adjusted multivariable models were gender, age, race, current smoking, former smoking, hypertension, hyperlipidaemia, diabetes, chronic renal impairment, and prior congestive heart failure.

We then assessed if the inverse association between BSA and CalcIndex was consistent across differing clinical subgroups. Across all examined subgroups, there were no significant differences identified, with the inverse association between BSA and CalcIndex persisting across groups differing according to age, gender, race, diabetes, renal impairment, smoking status, hyperlipidaemia, and hypertension (P interaction >0.1 for all subgroups) (Figure 3A). Similarly, for the inverse relationship between weight and CalcIndex, there were no significant differences among all examined subgroups (Figure 3B).

Figure 3

Forest plots demonstrating associations between (A) BSA and CalcIndex, and (B) weight and CalcIndex, in clinically relevant subgroups. BSA and weight were treated as continuous variables. Current and former smokers were combined in this analysis as the ‘smoking’ category. Boxes represent age-, gender-, and race/ethnicity-adjusted ORs for CalcIndex; lines represent 95% confidence intervals. Abbreviations not previously defined: DM, diabetes mellitus.

Figure 3

Forest plots demonstrating associations between (A) BSA and CalcIndex, and (B) weight and CalcIndex, in clinically relevant subgroups. BSA and weight were treated as continuous variables. Current and former smokers were combined in this analysis as the ‘smoking’ category. Boxes represent age-, gender-, and race/ethnicity-adjusted ORs for CalcIndex; lines represent 95% confidence intervals. Abbreviations not previously defined: DM, diabetes mellitus.

Discussion

The complex biology of vascular calcification is only beginning to be understood. Important recent insights include the role of circulating calcifying cells,22 the possibility that osteoclast-like cells may increase vascular osseous tissue,23 and that a close association exists between bone metabolism and the vasculature.5,6,8,24–27 Of relevance, CAC is a positive predictor of future cardiovascular events and mortality.28

In this study, we investigated the relationship between body morphology and CAC among patients undergoing PCI using three-dimensional IVUS. Although computed tomographic (CT) scanning is often used for this purpose, two-dimensional IVUS is known to offer superior sensitivity,29 while three-dimensional IVUS has been validated ex vivo and is highly accurate for the assessment and quantification of arterial calcification.19 The superior sensitivity of IVUS and CalcIndex for CAC detection is highlighted by the fact that in the CT-based MESA study, >40% of subjects had no detectable CAC,30 while only 15.9% (55/346) of our subjects had no index lesion calcification. Moreover, the current study represents one of the largest three-dimensional IVUS data sets ever analysed. The principal findings were (i) index lesion CalcIndex was inversely associated with body size, including weight, height, and BSA; (ii) the inverse relationship between CalcIndex and either weight or BSA persisted in multivariable regression models after adjustment for relevant clinical covariates; (iii) these relationships were consistent across multiple clinical subgroups.

To date, the overwhelming majority of studies that examined the relationship between CAC and body size was performed in homogeneous populations of middle aged Caucasians without known coronary artery disease.31–36 These studies, almost always performed using CT scanning, generally reported a positive association between CAC and body size.31–36 Because the current study used a more sensitive imaging modality in an older population of patients with significant coronary artery disease, our findings are not in conflict with these earlier reports.31–36 Rather, our data potentially indicate a more complex relationship between body morphology and CAC. Including the current study, a series of reports have suggested that a ‘U-shaped’ relationship exists between CAC and body size.12,14–17 That is, an initial positive association between body size and CAC may later become inverse due to older age or other factors. Interestingly, this hypothesis is consistent with the widely held belief that vascular calcification is more prevalent in ‘little old’ patients. Although our findings were similar between patients older vs. younger than 65 years of age (Figure 3), this may be due to the fact that there were few ‘young’ patients in this study, with only 59/346 (17.1%) being <55 years of age. By comparison, in many of the CT-based studies discussed above, the mean age was 55 years or less.31,33–35 Provocatively, in the CT-based Rotterdam study of older patients (mean age 72 years), obesity tended to be inversely related to the degree of arterial calcification in women.13

There are multiple biological pathways that might account for our findings. We originally undertook these investigations because bone mineralization is inversely related to CAC,5,6,8 but positively associated with measures of body size including weight, BMI, and BSA.9–11 In addition, a reasonable temporal concordance exists between the approximate age of the onset of accelerated bone loss37 and increasing vascular calcification.32 Therefore, a possible explanation for our findings may be that reduced body size with older age, in combination with other factors like menopause, leads to decreased bone mineral density which in turn leads to increased arterial calcification. In support of this, Shen et al.38 recently reported that a higher rate of bone loss is independently associated with an increased incidence of cardiovascular disease. Nevertheless, other explanations for our observations are also plausible, including that obese subjects with severe CAC do not survive to older age.

At the molecular level, numerous signalling pathways have been identified that regulate vascular calcification and bone mineralization. Notably, knockout of the klotho gene in mice leads to premature ageing, medial vascular calcification, and reduced bone mineral density.39 In humans, a relatively common functional klotho variant is an independent risk factor for coronary artery disease.40 Similarly, genetic knockout of osteoprotegerin leads to medial vascular calcification and decreased bone mineral density.41 Additional factors and signalling pathways that may be involved include fibroblast growth factor-23, fetuin-A,22,42 phosphate metabolism,43 and inflammation.26,27 Furthermore, circulating bone marrow-derived calcifying cells are also implicated in the bone-vascular axis.22 Future biomarker studies geared towards investigating these pathways will be critical for understanding the interactions between body size, bone mineralization, and vascular calcification.

Our results are in agreement with several other cardiovascular ‘obesity paradoxes’. These include (i) a survival advantage for obese patients following coronary revascularization44; (ii) reduced B-type natriuretic peptide levels and mortality in obese patients with heart failure45,46; (iii) reduced survival in lean patients undergoing valvular heart surgery47; (iv) a higher prevalence of hypertension, left ventricular hypertrophy, and increased mortality in lean haemodialysis patients48; and (v) improved survival and a lower risk of major vascular events after ischaemic stroke in overweight and obese patients.49,50 An important next step to delineate these ‘paradoxical’ relationships, including the current findings, is to define the relative importance of the morphological components that contribute to obesity, including lean mass vs. fat mass, and visceral vs. subcutaneous adipose tissue. Additional study is also required to determine whether vascular calcification plays a mechanistic role in some of these ‘obesity paradoxes’, like the survival advantage seen in obese patients after coronary revascularization. In other words, our findings provide support to the hypothesis that among an older population of patients with clinically significant coronary artery disease, one of the reasons why lean patients with a low body weight and/or BSA may ‘paradoxically’ suffer from more cardiovascular events than obese patients may be due their increased burden of calcific vascular disease.

Study limitations

We did not measure total CAC in the entire coronary tree, or other measures of body size or adiposity such as the waist/hip ratio. Body habitus at the time of PCI was unlikely to reflect lifetime BMI, BSA or weight. Grey scale IVUS is limited when compared with virtual histology (VH) IVUS to assess other plaque components such as a necrotic core, and it would be of relevance to compare VH-IVUS plaque components with body morphology parameters. Although this data set represents one of the largest three-dimensional IVUS data sets ever compiled, it remains relatively underpowered to examine complex biological associations.

Conclusions

Converging lines of evidence indicate a complex, age-dependent interplay between CAC and body morphology. In this large three-dimensional IVUS data set, we identified an independent, inverse relationship between body size and index lesion CAC in patients with significant coronary artery disease. These findings alter our framework for understanding vascular calcification and open the door to further studies investigating the interactions between body morphology, bone, and vascular disease. Moreover, our data indicate that reduced body weight or lower BSA are independent clinical risk factors for advanced calcific coronary artery disease.

Funding

No specific funding was used to perform this study.

Conflict of interest: The following authors have no conflicts of interest to declare: S.M.E., J.R.L., A.P.C., A.N.-K., N.G., A.H., U.B., S.S., M.F., and V.F. J.K. is supported by National Institutes of Health Grant K08HL111330 and has received research support from AstraZeneca. A.M. has received a research grant and acted as consultant for Boston Scientific. G.S.M. is a consultant and has received grant support from Boston Scientific and Volcano. G.D's. spouse has received consultant honoraria from Boston Scientific.

Acknowledgement

We thank Qi Zheng for her assistance with data preparation and formatting.

References

1
Kuchuk
NO
Pluijm
SM
van Schoor
NM
Looman
CW
Smit
JH
Lips
P
Relationships of serum 25-hydroxyvitamin D to bone mineral density and serum parathyroid hormone and markers of bone turnover in older persons
J Clin Endocrinol Metab
 , 
2009
, vol. 
94
 (pg. 
1244
-
50
)
2
Kestenbaum
B
Katz
R
de Boer
I
Hoofnagle
A
Sarnak
MJ
Shlipak
MG
, et al.  . 
Vitamin D, parathyroid hormone, and cardiovascular events among older adults
J Am Coll Cardiol
 , 
2011
, vol. 
58
 (pg. 
1433
-
41
)
3
Giovannucci
E
Liu
Y
Hollis
BW
Rimm
EB
25-hydroxyvitamin D and risk of myocardial infarction in men: a prospective study
Arch Intern Med
 , 
2008
, vol. 
168
 (pg. 
1174
-
80
)
4
Tanko
LB
Christiansen
C
Cox
DA
Geiger
MJ
McNabb
MA
Cummings
SR
Relationship between osteoporosis and cardiovascular disease in postmenopausal women
J Bone Miner Res
 , 
2005
, vol. 
20
 (pg. 
1912
-
20
)
5
Persy
V
D'Haese
P
Vascular calcification and bone disease: the calcification paradox
Trends Mol Med
 , 
2009
, vol. 
15
 (pg. 
405
-
16
)
6
Sage
AP
Tintut
Y
Demer
LL
Regulatory mechanisms in vascular calcification
Nat Rev Cardiol
 , 
2010
, vol. 
7
 (pg. 
528
-
36
)
7
Tanko
LB
Bagger
YZ
Christiansen
C
Low bone mineral density in the hip as a marker of advanced atherosclerosis in elderly women
Calcif Tissue Int
 , 
2003
, vol. 
73
 (pg. 
15
-
20
)
8
Hyder
JA
Allison
MA
Wong
N
Papa
A
Lang
TF
Sirlin
C
, et al.  . 
Association of coronary artery and aortic calcium with lumbar bone density: the MESA Abdominal Aortic Calcium Study
Am J Epidemiol
 , 
2009
, vol. 
169
 (pg. 
186
-
94
)
9
Akdeniz
N
Akpolat
V
Kale
A
Erdemoglu
M
Kuyumcuoglu
U
Celik
Y
Risk factors for postmenopausal osteoporosis: anthropometric measurements, age, age at menopause and the time elapsed after menopause onset
Gynecol Endocrinol
 , 
2009
, vol. 
25
 (pg. 
125
-
9
)
10
Broussard
DL
Magnus
JH
Risk assessment and screening for low bone mineral density in a multi-ethnic population of women and men: does one approach fit all?
Osteoporos Int
 , 
2004
, vol. 
15
 (pg. 
349
-
60
)
11
Wuster
C
Duckeck
G
Ugurel
A
Lojen
M
Minne
HW
Ziegler
R
Bone mass of spine and forearm in osteoporosis and in German normals: influences of sex, age and anthropometric parameters
Eur J Clin Invest
 , 
1992
, vol. 
22
 (pg. 
366
-
70
)
12
Shaffer
JR
Kammerer
CM
Rainwater
DL
O'Leary
DH
Bruder
JM
Bauer
RL
, et al.  . 
Decreased bone mineral density is correlated with increased subclinical atherosclerosis in older, but not younger, Mexican American women and men: the San Antonio Family Osteoporosis Study
Calcif Tissue Int
 , 
2007
, vol. 
81
 (pg. 
430
-
41
)
13
Odink
AE
van der Lugt
A
Hofman
A
Hunink
MG
Breteler
MM
Krestin
GP
, et al.  . 
Risk factors for coronary, aortic arch and carotid calcification; The Rotterdam Study
J Hum Hypertens
 , 
2010
, vol. 
24
 (pg. 
86
-
92
)
14
Cassidy
AE
Bielak
LF
Zhou
Y
Sheedy
PF
II
Turner
ST
Breen
JF
, et al.  . 
Progression of subclinical coronary atherosclerosis: does obesity make a difference?
Circulation
 , 
2005
, vol. 
111
 (pg. 
1877
-
82
)
15
Lee
DH
Steffes
MW
Gross
M
Park
K
Holvoet
P
Kiefe
CI
, et al.  . 
Differential associations of weight dynamics with coronary artery calcium vs. common carotid artery intima-media thickness: the CARDIA Study
Am J Epidemiol
 , 
2010
, vol. 
172
 (pg. 
180
-
9
)
16
Elmariah
S
Delaney
JA
O'Brien
KD
Budoff
MJ
Vogel-Claussen
J
Fuster
V
, et al.  . 
Bisphosphonate use and prevalence of valvular and vascular calcification in women MESA (The Multi-Ethnic Study of Atherosclerosis)
J Am Coll Cardiol
 , 
2010
, vol. 
56
 (pg. 
1752
-
9
)
17
Kovacic
JC
Lee
P
Baber
U
Karajgikar
R
Evrard
SM
Moreno
P
, et al.  . 
Inverse relationship between body mass index and coronary artery calcification in patients with clinically significant coronary lesions
Atherosclerosis
 , 
2012
, vol. 
221
 (pg. 
176
-
82
)
18
Chirumamilla
AP
Maehara
A
Mintz
GS
Mehran
R
Kanwal
S
Weisz
G
, et al.  . 
High platelet reactivity on clopidogrel therapy correlates with increased coronary atherosclerosis and calcification: a volumetric intravascular ultrasound study
JACC Cardiovasc Imaging
 , 
2012
, vol. 
5
 (pg. 
540
-
9
)
19
Scott
DS
Arora
UK
Farb
A
Virmani
R
Weissman
NJ
Pathologic validation of a new method to quantify coronary calcific deposits in vivo using intravascular ultrasound
Am J Cardiol
 , 
2000
, vol. 
85
 (pg. 
37
-
40
)
20
Wang
X
Lu
C
Chen
X
Zhao
X
Xia
D
A new method to quantify coronary calcification by intravascular ultrasound—the different patterns of calcification of acute myocardial infarction, unstable angina pectoris and stable angina pectoris
J Invasive Cardiol
 , 
2008
, vol. 
20
 (pg. 
587
-
90
)
21
Okwuosa
TM
Greenland
P
Lakoski
SG
Ning
H
Kang
J
Blumenthal
RS
, et al.  . 
Factors associated with presence and extent of coronary calcium in those predicted to be at low risk according to Framingham risk score (from the Multi-Ethnic Study of Atherosclerosis)
Am J Cardiol
 , 
2011
, vol. 
107
 (pg. 
879
-
85
)
22
Fadini
GP
Rattazzi
M
Matsumoto
T
Asahara
T
Khosla
S
Emerging role of circulating calcifying cells in the bone-vascular axis
Circulation
 , 
2012
, vol. 
125
 (pg. 
2772
-
81
)
23
Kovacic
JC
Randolph
GJ
Vascular calcification: harder than it looks
Arterioscler Thromb Vasc Biol
 , 
2011
, vol. 
31
 (pg. 
1249
-
50
)
24
Sennerby
U
Melhus
H
Gedeborg
R
Byberg
L
Garmo
H
Ahlbom
A
, et al.  . 
Cardiovascular diseases and risk of hip fracture
JAMA
 , 
2009
, vol. 
302
 (pg. 
1666
-
73
)
25
Foley
RN
Collins
AJ
Herzog
CA
Ishani
A
Kalra
PA
Serum phosphorus levels associate with coronary atherosclerosis in young adults
J Am Soc Nephrol
 , 
2009
, vol. 
20
 (pg. 
397
-
404
)
26
Pai
JK
Pischon
T
Ma
J
Manson
JE
Hankinson
SE
Joshipura
K
, et al.  . 
Inflammatory markers and the risk of coronary heart disease in men and women
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
2599
-
610
)
27
Papanicolaou
DA
Wilder
RL
Manolagas
SC
Chrousos
GP
The pathophysiologic roles of interleukin-6 in human disease
Ann Intern Med
 , 
1998
, vol. 
128
 (pg. 
127
-
37
)
28
Budoff
MJ
Shaw
LJ
Liu
ST
Weinstein
SR
Mosler
TP
Tseng
PH
, et al.  . 
Long-term prognosis associated with coronary calcification: observations from a registry of 25,253 patients
J Am Coll Cardiol
 , 
2007
, vol. 
49
 (pg. 
1860
-
70
)
29
van der Giessen
AG
Gijsen
FJ
Wentzel
JJ
Jairam
PM
van Walsum
T
Neefjes
LA
, et al.  . 
Small coronary calcifications are not detectable by 64-slice contrast enhanced computed tomography
Int J Cardiovasc Imaging
 , 
2011
, vol. 
27
 (pg. 
143
-
52
)
30
Criqui
MH
Kamineni
A
Allison
MA
Ix
JH
Carr
JJ
Cushman
M
, et al.  . 
Risk factor differences for aortic vs. coronary calcified atherosclerosis: the multiethnic study of atherosclerosis
Arterioscler Thromb Vasc Biol
 , 
2010
, vol. 
30
 (pg. 
2289
-
96
)
31
Allison
MA
Michael Wright
C
Body morphology differentially predicts coronary calcium
Int J Obes Relat Metab Disord
 , 
2004
, vol. 
28
 (pg. 
396
-
401
)
32
Kronmal
RA
McClelland
RL
Detrano
R
Shea
S
Lima
JA
Cushman
M
, et al.  . 
Risk factors for the progression of coronary artery calcification in asymptomatic subjects: results from the Multi-Ethnic Study of Atherosclerosis (MESA)
Circulation
 , 
2007
, vol. 
115
 (pg. 
2722
-
30
)
33
Lee
CD
Jacobs
DR
Jr
Schreiner
PJ
Iribarren
C
Hankinson
A
Abdominal obesity and coronary artery calcification in young adults: the Coronary Artery Risk Development in Young Adults (CARDIA) Study
Am J Clin Nutr
 , 
2007
, vol. 
86
 (pg. 
48
-
54
)
34
See
R
Abdullah
SM
McGuire
DK
Khera
A
Patel
MJ
Lindsey
JB
, et al.  . 
The association of differing measures of overweight and obesity with prevalent atherosclerosis: the Dallas Heart Study
J Am Coll Cardiol
 , 
2007
, vol. 
50
 (pg. 
752
-
9
)
35
Hsu
CH
Chang
SG
Hwang
KC
Chou
P
Impact of obesity on coronary artery calcification examined by electron beam computed tomographic scan
Diabetes Obes Metab
 , 
2007
, vol. 
9
 (pg. 
354
-
9
)
36
Kramer
CK
von Muhlen
D
Gross
JL
Barrett-Connor
E
A prospective study of abdominal obesity and coronary artery calcium progression in older adults
J Clin Endocrinol Metab
 , 
2009
, vol. 
94
 (pg. 
5039
-
44
)
37
Riggs
BL
Melton
LJ
Robb
RA
Camp
JJ
Atkinson
EJ
McDaniel
L
, et al.  . 
A population-based assessment of rates of bone loss at multiple skeletal sites: evidence for substantial trabecular bone loss in young adult women and men
J Bone Miner Res
 , 
2008
, vol. 
23
 (pg. 
205
-
14
)
38
Shen
C
Deng
J
Zhou
R
Chen
J
Fan
S
Li
Z
, et al.  . 
relation between bone mineral density, bone loss and the risk of cardiovascular disease in a Chinese Cohort
Am J Cardiol
 , 
2012
, vol. 
110
 (pg. 
1138
-
42
)
39
Kuro-o
M
Matsumura
Y
Aizawa
H
Kawaguchi
H
Suga
T
Utsugi
T
, et al.  . 
Mutation of the mouse klotho gene leads to a syndrome resembling ageing
Nature
 , 
1997
, vol. 
390
 (pg. 
45
-
51
)
40
Arking
DE
Becker
DM
Yanek
LR
Fallin
D
Judge
DP
Moy
TF
, et al.  . 
KLOTHO allele status and the risk of early-onset occult coronary artery disease
Am J Hum Genet
 , 
2003
, vol. 
72
 (pg. 
1154
-
61
)
41
Bucay
N
Sarosi
I
Dunstan
CR
Morony
S
Tarpley
J
Capparelli
C
, et al.  . 
Osteoprotegerin-deficient mice develop early onset osteoporosis and arterial calcification
Genes Dev
 , 
1998
, vol. 
12
 (pg. 
1260
-
8
)
42
Coen
G
De Paolis
P
Ballanti
P
Pierantozzi
A
Pisano
S
Sardella
D
, et al.  . 
Peripheral artery calcifications evaluated by histology correlate to those detected by CT: relationships with fetuin-A and FGF-23
J Nephrol
 , 
2011
, vol. 
24
 (pg. 
313
-
21
)
43
Linefsky
JP
O'Brien
KD
Katz
R
de Boer
IH
Barasch
E
Jenny
NS
, et al.  . 
Association of serum phosphate levels with aortic valve sclerosis and annular calcification: the cardiovascular health study
J Am Coll Cardiol
 , 
2011
, vol. 
58
 (pg. 
291
-
7
)
44
Oreopoulos
A
Padwal
R
Norris
CM
Mullen
JC
Pretorius
V
Kalantar-Zadeh
K
Effect of obesity on short- and long-term mortality postcoronary revascularization: a meta-analysis
Obesity (Silver Spring)
 , 
2008
, vol. 
16
 (pg. 
442
-
50
)
45
Lavie
CJ
Osman
AF
Milani
RV
Mehra
MR
Body composition and prognosis in chronic systolic heart failure: the obesity paradox
Am J Cardiol
 , 
2003
, vol. 
91
 (pg. 
891
-
4
)
46
Horwich
TB
Hamilton
MA
Fonarow
GC
B-type natriuretic peptide levels in obese patients with advanced heart failure
J Am Coll Cardiol
 , 
2006
, vol. 
47
 (pg. 
85
-
90
)
47
Vaduganathan
M
Lee
R
Beckham
AJ
Andrei
AC
Lapin
B
Stone
NJ
, et al.  . 
Relation of body mass index to late survival after valvular heart surgery
Am J Cardiol
 , 
2012
, vol. 
110
 (pg. 
1667
-
78
)
48
Agarwal
R
Body mass index-mortality paradox in hemodialysis: can it be explained by blood pressure?
Hypertension
 , 
2011
, vol. 
58
 (pg. 
1014
-
20
)
49
Ovbiagele
B
Bath
PM
Cotton
D
Vinisko
R
Diener
HC
Obesity and recurrent vascular risk after a recent ischemic stroke
Stroke
 , 
2011
, vol. 
42
 (pg. 
3397
-
402
)
50
Vemmos
K
Ntaios
G
Spengos
K
Savvari
P
Vemmou
A
Pappa
T
, et al.  . 
Association between obesity and mortality after acute first-ever stroke: the obesity-stroke paradox
Stroke
 , 
2011
, vol. 
42
 (pg. 
30
-
6
)

Comments

0 Comments