Effects of Abdominal Aortic Calcication and Facet Joint Arthritis On Lumbar Bone Mineral Density Using Dual-Energy X-Ray Absorptiometry

Abdominal aortic calcification (AAC) may overestimate lumbar bone mineral density (BMD) examined 23 by dual-energy X-ray absorptiometry (DXA); however, the degree of effect of AAC on lumbar BMD has 24 not been quantified. In particular, no study has quantitatively compared and analysed segmental BMD and 25 AAC using computed tomography (CT) scan. Thus, this study aimed to quantify the effect of AAC on 26 BMD measurements using DXA via multiple linear regression analysis. the BMD of each lumbar segment using 31 CT.


absorptiometry; computed tomography 45
Background 46 Dual-energy X-ray absorptiometry (DXA) is a widely used diagnostic tool for osteoporosis, which can evaluate fracture risk and monitor treatment response [1]. However, confounding factors can 48 lead to measurement errors when DXA is used to evaluate lumbar bone mineral density (BMD); these 49 factors include vertebral size and shape variations, bone marrow fat, soft tissue calcification, and 50 degenerative lumbar spine changes.[2, 3] Among them, common degenerative changes in the elderly 51 include facet joint arthritis (FJA) and abdominal aortic calcification (AAC) [4,5], which have been 52 associated with BMD. Previous studies agree that FJA overestimates BMD measurement [6,7]. In 53 contrast, the effect of AAC on BMD remains controversial. One case report found that AAC 54 overestimates lumbar BMD measurement on DXA due to beam path calcification [8]. On the other hand, 55 some studies have suggested that AAC is not related to BMD [9][10][11]. Other studies have reported a 56 negative correlation between AAC and BMD [12,13]. Based on existing literature, the impact of AAC on 57 BMD measurements seems minimal. However, the additive effect of calcified aortic tissue on DXA 58 measurements remain unclear. To the best of our knowledge, no previous studies have quantified AAC 59 using computed tomography (CT) scan and conducted simultaneous evaluation of each spinal segment. 60 Therefore, this study aimed to quantify the effect of AAC on BMD by evaluating AAC using CT scans 61 and to examine the effect of AAC on BMD at each spine level. 62

Evaluation of FJA and AAC 96
FJA was evaluated in four stages from Grade 0 to Grade 3 via CT scan. Namely, grade 0 was 97 normal; grade 1, joint space narrowing; grade 2, narrowing with sclerosis or hypertrophy; and grade 3, 98 severe osteoarthritis with osteophyte and sclerosis. [15] For ease of analysis, the FJA grading was 99 converted to an ordered variable from 0 to 3, corresponding to the grade. 100 The AAC volume was obtained by measuring the volume of abdominal aortic calcification, 101 which showed attenuation of >90 Hounsfield units on CT. Between the L1-L4 vertebra, measurements 102 were performed for aortic calcification located anterior to each vertebra. To calculate the volume, a cross-103 sectional area above 90 HU was identified in the axial cut of the CT between the upper and the lower 104 endplate of each lumbar segment and multiplied by the number of cuts of the cross-sectional area. [16] 105 The volume of aortic calcification corresponding to each spinal segment was recorded in cubic 106 millimeters. 107

Vertebral compression fracture and previous spine surgery 108
Vertebral compression fractures and previous spine surgeries were investigated using medical 109 records, radiography, and CT scans. The presence of a vertebral compression fracture was evaluated and 110 recorded for each segment. In this study, any spine surgery included fusion surgery, partial laminectomy, 111 diskectomy, vertebroplasty, and kyphoplasty performed between L1 and L4. Fusion surgery was defined 112 as any interbody fusion surgery or posterolateral fusion surgery of the lumbar spine that required 113 instrumentation. The BMD of the vertebral segment with compression fracture was included in the 114 analysis, but not BMD of vertebral segments treated with surgery. variables, and their significance was verified. 122 In our multiple linear regression model, sex was assigned a value of 0 and 1 for women and 123 men, respectively. Age was assigned a value of years. BMI was used as a continuous variable, which was 124 measured as kg/m 2 ). FJA was introduced as an ordinal scale with values ranging from 0 to 3 according to 125 the grading system. Other factors such as the presence of vertebral fracture, fusion surgery [17], 126 osteoporosis medication, and steroid use were assigned a value of 1 and their absence was given a value 127 of 0. Additionally, the independence, multicollinearity, and homoscedasticity of the standard residuals of 128 the independent variables were examined. All data were analyzed using SPSS Statistics version 20 (IBM, 129 Armonk, NY, USA). 130

Demographic data 133
During the study period, 685 subjects were considered for the study. However, 65 patients were 134 excluded; among them, 37 were excluded due to fractures of >3 vertebral segments between L1-L4, 15 135 underwent spinal surgery involving ≥3 surgical segments between L1-L4, 12 had ≥3 fractures in surgical 136 segments between L1-L4, and one had multiple myeloma. A total of 620 subjects (153 men and 467 137 women) were enrolled in the analysis. 138 The 620 patients had a mean age of 71.6 ± 9.1 years (range, 31 to 89 years) and mean BMI of 139 24.9 ± 4.1 kg/m 2 . Among the 620 patients, 181 (29.2%) had fractures in vertebral bodies between L1 and 140 L4, 66 (10.6%) had history of spine surgery, and 50 (8.1%) had fusion surgery involving L1 through L4. 141 (Table 1) More detailed information regarding fracture and spine surgery is provided in Supplementary 142 Table 1. 143

Lumbar BMD and T-score 144
In all enrolled study subjects, the mean BMD of women and men was 0.900 ± 0.175 and 1.063 145 1.72), respectively. The mean and standard deviation of the BMD and T-score for each sex and segment 147 are summarized in Table 2. 148

Abdominal aortic calcification 149
The overall prevalence of AAC measured by CT was 60.4% in women and 68.0% in men. 150 Depending on the spinal segment, the prevalence and volume of AAC tended to increase from L1 to L4 in 151 both men and women. Among L1 to L4, L3 (45.2%) had the highest prevalence of AAC, followed by L4 152 (41.1%), L2 (23.4%), and L1 (11.3%). The average AAC volume was the highest in L4 at 213.67 ± 153 443.82 mm 3 , followed by L3 at 161.95 ± 338.09 mm 3 . The mean volume of AAC was 27.39 ± 130.27 and 154 60.37 ± 218.84 mm 3 in L1 and L2, respectively, which were less than half compared to L3 or L4. 155 Table 2) 156

(Supplementary
In addition, AAC increased with age in both men and women, which was observed in 72.3% of 157 the patients >70 years of age. Furthermore, AAC was found in 79.2% of women and 76.2% of men over 158 80 years of age. In contrast, only 27.7% of women and 47.6% of men aged <60 years demonstrated AAC. 159 Similarly, the mean AAC volume increased gradually with age and toward the lower lumbar spine. In 160 patients >80 years old, the maximum volume of AAC in L4 were 372.4 ± 461.8 mm 3 and 509.5 ± 763.7 161 mm 3 in women and men, respectively. (Table 3) 162

Multiple linear regression analysis by gender 185
The regression model for female confirmed in this study was as follows. This study has several findings. First, AAC was common in elderly men and women (72.3% 211 over 70), which tended to increase with age, regardless of sex, prevalence, and volume. Second, the 212 prevalence and volume of AAC were significantly higher at the L3-4 level compared to the L1-2 level. 213 Third, grade 3 and grade 1 FJA were most common in L4 and L1, respectively. Fourth, females 214 demonstrated that the BMD of L3 and L4 were significantly positively correlated with BMI, FJA, and 215 AAC volume. 216 AAC and FJA are common degenerative changes found in the elderly population [4, 5] and may 217 act as confounding factors for BMD measurements. Vascular calcification is an actively regulated process 218 affected by the balance of several factors. It has been suggested that chronic kidney disease, diabetes 219 mellitus, atherosclerosis, and aging potentially affect vascular calcification formation and progression. 220 analysing the effect on each segment's BMD. 254 In this study, we obtained consistent results that AAC was positively correlated with BMD 255 measurements in L3 and L4 in women. The regression coefficients seemed to be lower than the other 256 factors. However, in our study, 14.5% of AACs found at the L4 level in women had >1000 mm 3 in 257 volume, with an average volume of 1615 mm 3 . This indicates that L4 BMC can be measured 10.7% larger 258 on average, which is sufficient to change the T-score. Thus, in female subjects, the BMD of L3/L4 can be 259 overestimated by AAC, especially in those older than 70 years. In contrast, the BMD of males were not 260 associated with AAC, FJA, and age. This may be due to the lower number of males or the presence of 261 other significant factors. Currently, no study has found the effects of AAC on the BMD of males. 262 Therefore, further investigation of BMD in men is required. This study has some limitations. First, the rate of BMD decline was not constant, but this was 272 not reflected in our model. Second, the grading of each facet joint was analysed as a continuous variable.

Ethics approval and consent to participate 285
The protocol of this study was approved by the Institutional Review Board of Seoul 286 Metropolitan Government-Seoul National University (SMG-SNU) Boramae Medical Center. 287

Consent for publication 288
Not applicable 289

Availability of Data and Materials 290
The datasets generated during and/or analysed during the current study are available from the 291 corresponding author on reasonable request. 292

Competing interests 293
The authors declare that they have no known competing financial interests or personal relationships that 294 could have appeared to influence the work reported in this paper. 295              Figure 1 Facet joint arthritis from L1 to L4