Introduction 1 2 3 5 5 6 7 8 9 10 11 13 14 15 17 18 19 The aim of the present study was to test this hypothesis and investigate associations of a wide array of serum lipids and lipoproteins as well as of a genetic determinant of lipid profile with different surrogate measures of osteoporosis (BMD, fractures) and cardiovascular disease (AC, acute events) in 1176 Danish postmenopausal women providing full-set of data for such analyses. Subjects and methods Study population 20 All women gave written their informed consent to participation and the study was carried out in compliance with the Helsinki Declaration II and the European Standards for Good Clinical Practice. The study protocol was approved by the local ethical committee. Demographic characteristics, risk factors, and clinical events 20 Bone mineral density and fat mass measurement 21 22 Fracture diagnosis Lateral X-rays of the thoracic and lumbar spine were performed using standard X-ray equipment. Vertebral deformities from T4 to L4 were assessed by digital measurements of vertebral deformations using the Image Pro Image Analyzer software (version 4.5 for windows, Media Cybernetics Inc., Silver Spring, MD, USA). The ratio of the anterior and posterior heights of each vertebral body was determined and a difference between the anterior and posterior edges exceeding 20% was considered as a radiographic vertebral fracture. Information on prevalent non-vertebral fractures (wrist, hip, humeral fracture, rib, ankle, and foot) was collected during personal interview and later verified by X-rays or hospital discharge summaries. Grading of aorta calcification r n Laboratory parameters Blood samples were collected in the morning after fasting overnight (>12 h). Serum total cholesterol, triglyceride, and high-density lipoprotein (HDL) cholesterol were determined by enzymatic assay using a Vitros-250 (Johnson & Johnson, Taastrup, Denmark). LDL cholesterol was calculated by Friedewald formula (LDL−C = TC−HDL−C−0.45 × TG). Lipoprotein(a), ApoA1, and ApoB were measured by the Cobas Mira (Hoffman-La Roche, Mannheim, Germany) automatic blood analyzer. Genotyping Hae T Afl Statistical analysis Results are presented as mean ± SD, unless otherwise indicated. Confounders of hip BMD and AC were identified by univariate correlation analysis. We classified ApoE genotypes into two categories marked by absence or presence of the ɛ4 allele. General linear models (GLM) tested the differences in selected dependent variables after adjustment for possible confounders. Similar approach was used to identify factors showing differences between women with or without fractures. Logistic regression models were established to assess the age-dependent and age-independent relative risk of wrist, hip, or vertebral fractures in women with advanced aorta calcification or cardiovascular disease. Statistical analyses were carried out using the SPSS data analysis software (version 12, SPSS Inc., Chicago, IL). All statistical tests were two-tailed and differences were considered significant if the p-value was less than 0.05. Results Demographic characteristics 1 equal or below −2.5 Table 1 Demographic characteristics in the study population Characteristics and equivalent values No. of women 1176 Age (years) 69.3 ± 6.4 Age at menopause (years) 48.7 ± 5.0 2 26.4 ± 4.4 Current smoking (%) 24.6% Regular intake of alcohol (%) 55.2% Serum total cholesterol (mmol/l) 6.3 ± 1.0 Serum triglyciders (mmol/l) 1.4 ± 0.7 Serum HDL (mmol/l) 1.7 ± 0.4 Serum LDL (mmol/l) 4.0 ± 0.9 Serum ApoA1 (mmol/l) 203.9 ± 37.1 Serum ApoB (mmol/l) 120.3 ± 24.5 Serum Lp(a) (mmol/l) 31.9 ± 34.6 ApoE ɛ4 allele carrier 367, 31.2% 2 0.80 ± 0.1 2 0.92 ± 0.2 Hip BMD T-score ≤−2.5 (no. of women,%) 147, 12.5% Spine BMD T-score ≤−2.5 (no. of women,%) 179, 15.2% Vertebral fracture (no. of women, %) 204, 17.6% Hip fracture (no. of women,%) 16, 1.4% Aorta calcification 3.0 ± 3.8 Cardiovascular disease (no. of women,%) 50, 4.3% Data shown are mean±SD. 2 22 23 Correlates of hip and spine BMD and AC 2 Table 2 Independent correlates of aorta calcification, hip BMD, and spine BMD (multiple regression models)   Aorta Calcification Hip BMD Spine BMD Age 0.35 (p < 0.001) −0.23 (p < 0.001) −0.03 (p = 0.45) YSM −0.06 (p = 0.16) −0.06 (p = 0.10) −0.08 (p = 0.04) BMI −0.09 (p = 0.07) 0.48 (p < 0.001) 0.45 (p < 0.001) TFM% −0.04 (p = 0.36) −0.10 (p = 0.04) −0.21 (p < 0.001) CFM/PFM ratio 0.08 (p = 0.02) −0.004 (p = 0.85) 0.05 (p = 0.21) Smoking 0.19 (p < 0.001) −0.05 (p = 0.10) −0.01 (p = 0.66) Exercise −0.03 (p = 0.25) 0.04 (p = 0.17) 0.04 (p = 0.21) Treated hypertension 0.08 (p = 0.006) 0.006 (p = 0.81) 0.05 (p = 0.09) Alcohol consumption −0.08 (p = 0.007) 0.06 (p = 0.04) 0.07 (p = 0.02) Apo e4 −0.009 (p = 0.77) 0.01 (p = 0.76) −0.03 (p = 0.22) Triglyceride 0.02 (p = 0.63) 0.03 (p = 0.43) 0.04 (p = 0.32) HDL-C 0.05 (p = 0.48) −0.06 (p = 0.29) 0.02 (p = 0.74) LDL-C Excl. 0.02 (p = 0.73) −0.10 (p = 0.10) Total cholesterol 0.05 (p = 0.43) Excl. Excl. ApoA1 −0.10 (p = 0.04) 0.05 (p = 0.36) 0.05 (p = 0.39) ApoB 0.10 (p = 0.14) −0.03 (p = 0.60) 0.10 (p = 0.85) Aorta Calcification  −0.09 (p = 0.002) 0.04 (p = 0.20)  R = 0.44, SEE = 3.32 P < 0.001 R = 0.55, SEE = 0.09 P < 0.001 R = 0.34, SEE = 0.13 P < 0.001 Excl.: excluded from the model due to strong co-linearity. The independent role of the ApoE ɛ 4 allele 3 3 Table 3 Lipid and bone profile in women with or without an ApoE epsilon 4 allele   Absence of ApoE epsilon 4 allele (n = 809) Presence of ApoE epsilon 4 allele (n = 367) P value Serum total cholersterol* (mmol/l) 6.24 ± 0.03 6.51 ± 0.05 <0.001 Serum triglycides* (mmol/l) 1.36 ± 0.02 1.48 ± 0.03 0.005 Serum HDL* (mmol/l) 1.72 ± 0.01 1.65 ± 0.02 0.006 Serum LDL* (mmol/l) 3.91 ± 0.03 4.19 ± 0.05 <0.001 Serum ApoA1* (mmol/l) 205.8 ± 1.3 201.3 ± 2.0 0.06 Serum ApoB* (mmol/l) 117.3 ± 0.8 127.2 ± 1.3 <0.001 Serum Lp(a)* (mmol/l) 31.9 ± 1.3 29.2 ± 2.0 0.24 2 0.803 ± 0.004 0.804 ± 0.005 0.89 2 0.924 ± 0.005 0.915 ± 0.008 0.32 Hip fracture 1.7% 0.5% 0.17 Vertebral fractures 16.4% 19.3% 0.24 Severity score of AC** 2.81 ± 0.14 2.97 ± 0.21 0.13 Cardiovascular disease 4.0% 4.9% 0.44 Values are mean±SEM. *Adjustment for age, years since menopause, BMI, TFM%, CFM/PFM ratio, current smoking, physical exercise, and regular alcohol consumption. **Adjustment for the aforementioned confounders + treated hypertension. Associations of lipids with BMD and AC 2 2 Table 4 Characteristics of women with or without at least one vertebral fracture   Without vertebral fracture With vertebral fracture p-value Age 68.9 ± 6.4 71.3 ± 6.4 <0.001 YSM 12.7 ± 8.4 15.5 ± 8.6 <0.001 BMI 26.4 ± 4.4 26.3 ± 4.4 NS TFM% 41.3 ± 7.6 40.7 ± 7.8 NS CFM/PFM ratio 0.88 ± 0.23 0.90 ± 0.24 NS Regular exercise 73.8% 67.6% 0.08 ApoE ɛ4 allele + 30.5% 34.8% NS Total cholesterol 6.32 ± 1.00 6.31 ± 1.04 NS Triglycerides 1.42 ± 0.69 1.31 ± 0.54 0.02 HDL-C 1.68 ± 0.43 1.71 ± 0.39 NS LDL-C 3.98 ± 0.93 4.01 ± 0.96 NS ApoA1 203.8 ± 37.8 204.6 ± 33.4 NS ApoB 120.6 ± 24.5 119.1 ± 24.5 NS Lp(a) 31.7 ± 34.9 32.5 ± 33.5 NS Spine BMD 0.93 ± 0.15 0.87 ± 0.14 <0.001 Hip BMD 0.82 ± 0.11 0.75 ± 0.11 <0.001 Hip fracture 1.0% 2.9% 0.04 Wrist fracture 11.4% 14.6% 0.04 AC score 2.8 ± 3.7 3.5 ± 3.4 0.03 CVD 3.6% 7.4% 0.02 There were no significant differences in the frequency of smoking, regular alcohol, coffee, milk or seasonal vitamin D consumption, type 2 diabetes, treated hyperlipidemia or treated hypertension Comparison of women with or without fractures 4 When comparing the lipid profile of women with or without osteoporotic fractures after adjustment for potential confounders (i.e., age, years since menopause, BMI, TFM%, CFM/PFM ratio, smoking, ApoE ɛ4 allele, physical exercise, alcohol consumption, and severity score of AC), only triglycerides showed significant differences between the two groups (p < 0.05). Thus, triglycerides in women with and without fractures were 1.31 ± 0.04 mmol/l and 1.42 ± 0.03 mmol/l, respectively. When analyzing fracture types separately, the difference between women with or without hip or wrist fractures were not statistically significant. The relative risk of hip or vertebral fracture in women with severe AC or CVD 5 Table 5 Relative risk of different types of osteoporotic fractures in women with severe aorta calcification (score ≥ 6) and/or manifest cardiovascular disease (n = 282)     Odds ratio (95% CI) p-value Hip fracture N = 16 Crude 4.4 (1.6–12.9) 0.001 Age-adjusted 3.0 (1.1–8.8) 0.04 Vertebral fracture N = 204 Crude 1.2 (0.9–1.7) 0.27 Age-adjusted 1.0 (0.7–1.4) 0.93 Wrist fracture N = 145 Crude 1.4 (1.0–2.1) 0.06 Age-adjusted 1.2 (0.8–1.7) 0.46 Discussion In the present study, we investigated whether serum lipid and lipoproteins represent a biological linkage between osteoporosis and cardiovascular disease. The main findings were as follows: 1) neither allelic polymorphism in the ApoE gene nor serum lipids per se were independent correlates of BMD at any skeletal site, 2) ApoA1 levels were independent correlates of AC, which seem to play a role in the determination of hip BMD, 3) decreased triglycerides were independently associated with vertebral fractures, but not with non-vertebral fractures (hip and wrist), and 4) advanced vascular disease was a significant contributor to risk of hip fractures, but not to vertebral or wrist fractures. These observations suggest that the role of lipids as linking factors between cardiovascular disease and osteoporosis is indirect via promotion of atherogenesis, which in turn can contribute to bone loss, especially at skeletal sites with end-arterial blood supply. Antiatherogenic measures can also be expected to contribute to decreasing the burden of hip fractures among the elderly. 24 25 26 27 28 29 30 27 31 32 33 35 11 14 15 17 19 18 15 2 36 5 37 38 39 In summary, the results of the present observational study provide further evidence for the independent association of peripheral vascular disease with osteoporosis in the proximal femur. Since the association of lipids and lipoproteins to BMD and non-vertebral fractures is not independent of the severity of AC, it seems unlikely that these metabolites exert direct and clinically significant effects on bone turnover in postmenopausal women. Their contribution is via promotion of atherogenesis, in which regard ApoA1 levels seem to take a leading role. The remaining issue to be clarified is which genetic or environmental factors underlie the association of low triglycerides levels to vertebral fractures.