Alcohol Intake and Bone Turnover in Postmenopausal Women
Alcohol Intake and Bone Turnover in Postmenopausal Women
The physical characteristics of the study population at baseline are presented in Table 1 and reflect a typical healthy population of postmenopausal women. Based on the Block Food Frequency Questionnaire, the mean daily intakes of calcium were 814 ± 46 mg/day as dietary calcium and 748 ± 76 mg/day as supplemental calcium, and the mean daily intakes of vitamin D were 128 ± 12 IU/day as dietary vitamin D and 383 ± 41 IU/day as supplemental vitamin D.
Based on 1-week alcohol records before study initiation, the amount of alcohol consumed by the participants averaged 1.4 standard (American) drinks/day or ethanol at 19 ± 1 g/day.
T scores (comparison to young adults) for BMD ranged from -0.50 to -1.00, whereas Z scores (comparison to same age adults) ranged from 0.12 to 0.29 ( Table 2 ).
There was a positive association at baseline between alcohol consumption and total hip BMD (Figure 1A) and between alcohol consumption and trochanter BMD (Figure 1B). Significant associations were not detected between alcohol intake and BMD at the femoral neck (Figure 1C) or the lumbar spine (Figure 1D). Femur measurements were not available for one participant owing to a double hip replacement.
(Enlarge Image)
Figure 1.
Total Hip (A), Trochanter (B), Femoral Neck (C), and Lumbar Spine (D) BMD Versus Quantity of Alcohol Intake in Early Postmenopausal Women. Please note the significant positive association between alcohol intake and total hip (A) and trochanter (B) BMD. n = 39 for femur and n = 40 for lumbar spine. BMD, bone mineral density.
A significant linear association between osteocalcin and CTx was observed at baseline (r = 0.36; P < 0.01). However, no relationship between baseline levels of alcohol consumption and bone turnover markers was noted.
Serum osteocalcin and CTx increased after 14 days of abstinence (P = 0.01 and P = 0.02, respectively; Figure 2). When alcohol was administered on the evening of day 14 and when blood was drawn on the morning of day 15, serum osteocalcin and CTx decreased compared with day 14 (P = 0.01 and P = 0.05, respectively). Day 15 levels of serum osteocalcin and CTx did not differ from baseline (P = 0.94 and P = 0.57, respectively).
(Enlarge Image)
Figure 2.
Mean Percentage Change in Serum Osteocalcin (A), a Marker of Bone Formation, and in Serum Ctx (B), a Marker of Bone Resorption, After 14 Days of Abstinence From Alcohol and Subsequent Restoration of Alcohol Intake to Preabstinence Levels. Please note that a relatively brief interval of abstinence (14 d) resulted in significant increases in serum osteocalcin and CTx, whereas resumption of alcohol consumption resulted in an overnight decrease in serum osteocalcin and CTx to levels that did not differ from baseline (day 0). Values are expressed as mean +/- SE (n = 40 for each time interval). P ≤ 0.05 versus day 0. P ≤ 0.05 versus day 14. CTx, C-terminal telopeptide.
Serum estradiol was evaluated in 39 of 40 participants. When all values were included in the calculation of mean ± SE, estradiol levels were found to be 7.0 ± 1.0 pg/mL on day 0, 7.2 ± 1.1 pg/mL on day 14, and 6.6 ± 1.2 pg/mL on day 15. However, 33 of 117 samples (each participant assayed at three time points) were below the reliable detection limit for the assay (4 pg/mL). A preliminary analysis set to 0 all values that fell below the detection limit and fitted a general linear model with a compound symmetric correlation structure using maximum likelihood methods. The null hypothesis of equal mean estradiol levels at baseline, day 14, and day 15 had a likelihood ratio of P = 0.67. There also was no evidence of a difference in mean estradiol levels over time when all nonmeasurable levels were set equal to 4 (likelihood ratio P = 0.65). In addition, a linear mixed model was fitted using clustered normal data, with undetectable values treated as intervals censored between 0 and 4. There was no evidence of a difference in the mean estradiol levels from days 0 to 14 (P = 0.52) or between days 14 and 15 (P = 0.40). The very low serum estradiol levels observed in this study are consistent with the expected low levels of estradiol (0-30 pg/mL) in postmenopausal women who are not on HT.
Results
Study Population
The physical characteristics of the study population at baseline are presented in Table 1 and reflect a typical healthy population of postmenopausal women. Based on the Block Food Frequency Questionnaire, the mean daily intakes of calcium were 814 ± 46 mg/day as dietary calcium and 748 ± 76 mg/day as supplemental calcium, and the mean daily intakes of vitamin D were 128 ± 12 IU/day as dietary vitamin D and 383 ± 41 IU/day as supplemental vitamin D.
Alcohol Intake
Based on 1-week alcohol records before study initiation, the amount of alcohol consumed by the participants averaged 1.4 standard (American) drinks/day or ethanol at 19 ± 1 g/day.
BMD
T scores (comparison to young adults) for BMD ranged from -0.50 to -1.00, whereas Z scores (comparison to same age adults) ranged from 0.12 to 0.29 ( Table 2 ).
Alcohol Consumption and BMD
There was a positive association at baseline between alcohol consumption and total hip BMD (Figure 1A) and between alcohol consumption and trochanter BMD (Figure 1B). Significant associations were not detected between alcohol intake and BMD at the femoral neck (Figure 1C) or the lumbar spine (Figure 1D). Femur measurements were not available for one participant owing to a double hip replacement.
(Enlarge Image)
Figure 1.
Total Hip (A), Trochanter (B), Femoral Neck (C), and Lumbar Spine (D) BMD Versus Quantity of Alcohol Intake in Early Postmenopausal Women. Please note the significant positive association between alcohol intake and total hip (A) and trochanter (B) BMD. n = 39 for femur and n = 40 for lumbar spine. BMD, bone mineral density.
Alcohol Consumption and Biomarkers of Bone Turnover
A significant linear association between osteocalcin and CTx was observed at baseline (r = 0.36; P < 0.01). However, no relationship between baseline levels of alcohol consumption and bone turnover markers was noted.
Serum osteocalcin and CTx increased after 14 days of abstinence (P = 0.01 and P = 0.02, respectively; Figure 2). When alcohol was administered on the evening of day 14 and when blood was drawn on the morning of day 15, serum osteocalcin and CTx decreased compared with day 14 (P = 0.01 and P = 0.05, respectively). Day 15 levels of serum osteocalcin and CTx did not differ from baseline (P = 0.94 and P = 0.57, respectively).
(Enlarge Image)
Figure 2.
Mean Percentage Change in Serum Osteocalcin (A), a Marker of Bone Formation, and in Serum Ctx (B), a Marker of Bone Resorption, After 14 Days of Abstinence From Alcohol and Subsequent Restoration of Alcohol Intake to Preabstinence Levels. Please note that a relatively brief interval of abstinence (14 d) resulted in significant increases in serum osteocalcin and CTx, whereas resumption of alcohol consumption resulted in an overnight decrease in serum osteocalcin and CTx to levels that did not differ from baseline (day 0). Values are expressed as mean +/- SE (n = 40 for each time interval). P ≤ 0.05 versus day 0. P ≤ 0.05 versus day 14. CTx, C-terminal telopeptide.
Alcohol Consumption and Serum Estradiol
Serum estradiol was evaluated in 39 of 40 participants. When all values were included in the calculation of mean ± SE, estradiol levels were found to be 7.0 ± 1.0 pg/mL on day 0, 7.2 ± 1.1 pg/mL on day 14, and 6.6 ± 1.2 pg/mL on day 15. However, 33 of 117 samples (each participant assayed at three time points) were below the reliable detection limit for the assay (4 pg/mL). A preliminary analysis set to 0 all values that fell below the detection limit and fitted a general linear model with a compound symmetric correlation structure using maximum likelihood methods. The null hypothesis of equal mean estradiol levels at baseline, day 14, and day 15 had a likelihood ratio of P = 0.67. There also was no evidence of a difference in mean estradiol levels over time when all nonmeasurable levels were set equal to 4 (likelihood ratio P = 0.65). In addition, a linear mixed model was fitted using clustered normal data, with undetectable values treated as intervals censored between 0 and 4. There was no evidence of a difference in the mean estradiol levels from days 0 to 14 (P = 0.52) or between days 14 and 15 (P = 0.40). The very low serum estradiol levels observed in this study are consistent with the expected low levels of estradiol (0-30 pg/mL) in postmenopausal women who are not on HT.