J Orthop Sci (2015) 20:522–528 DOI 10.1007/s00776-015-0695-7
ORIGINAL ARTICLE
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years Rowshan Ara Begum · Liaquat Ali · Osamu Takahashi · Tsuguya Fukui · Mahbubur Rahman
Received: 8 August 2014 / Accepted: 30 December 2014 / Published online: 28 January 2015 © The Japanese Orthopaedic Association 2015
Abstract Background Studies examining the reference values of bone mineral density (BMD) and their patterns at different ages are largely based on data generated from developed countries. The objective of this study was to estimate reference values of BMD, along with their correlates, for women living in urban and suburban areas of Bangladesh. Methods Dual-energy X-ray absorptiometry scans were performed on 500 women 16–65 years of age. Reference values at the lumbar spine (LS) and femoral neck (FN) were estimated after adjusting for height and weight. In addition, multiple linear regression analysis was used to examine correlates of BMD at the LS and FN. Results Mean BMD was highest at the LS and FN in women 16–19 (1.001–1.007 g/cm2) and 16–22 years of age (0.880–0.888 g/cm2), respectively, and gradually declined with increasing age thereafter. BMD decreased at an annual rate of 0.0027 g/cm2 at the LS and 0.0046 g/cm2 at the FN among women aged 16–45 years. For women 46–65 years R. A. Begum Department of Reproductive and Child Health, Bangladesh University of Health Science, Mirpur, Dhaka, Bangladesh L. Ali Department of Biochemistry and Cell Biology, Bangladesh University of Health Science, Mirpur, Dhaka, Bangladesh O. Takahashi · T. Fukui Center for Clinical Epidemiology, St. Luke’s International Hospital, Akashi‑cho 9‑1, Chuo‑ku, Tokyo 104‑8560, Japan M. Rahman (*) Department of Obstetrics and Gynecology and the Center for Interdisciplinary Research in Women’s Health, University of Texas Medical Branch, 301 University Blvd., John Sealy Annex, Room No. 3.108.D, Galveston, TX 77555‑0587, USA e-mail: [email protected]
13
of age, the respective figures were 0.0073 and 0.0083 g/ cm2. In both age groups, body weight was positively associated with BMD at both sites, and with height only at the LS. In addition, years of pill use was positively associated with BMD at the LS among women aged 16–45 years. BMD levels at both sites began to decline during the early twenties among Bangladeshi women. Conclusions Age-specific BMD data generated in this study could be useful for interpreting bone densitometry data among women in Bangladesh and other South Asian countries.
Introduction Osteoporosis, characterized by low bone mass and gradual deterioration of bone tissue, is the most common metabolic bone disease among women [1]. A growing elderly population, vitamin D deficiency, and inadequate dietary calcium intake are considered contributing factors to the overall burden of osteoporosis in developing countries [2, 3]. It is expected that within several decades, 75 % of hip fractures worldwide will occur in the developing countries, as this burden has been stabilized, or is even decreasing, in the developed world [2–4]. Accurate measurement of bone density using a dual-energy X-ray absorptiometry (DXA) scan is very important for identifying low bone mineral density (BMD) in assessing future fracture risk. BMD measured using the DXA technique is considered an important predictor of future fractures [5]. One of the key determinants for identifying low bone density situations such as osteopenia and osteoporosis is the reference value of BMD used for DXA scans, which is largely based on values that have been determined for populations in developed or other affluent countries [6–16]. Moreover, the majority of studies
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years
examining BMD patterns at different age levels are also severely limited by the use of data generated from affluent countries, whose sociodemographic characteristics are different from those of South Asian countries. This is a serious limitation in the literature. The objective of this study was to estimate the reference BMD values at the lumbar spine and femoral neck, along with their correlates, based on Bangladeshi women 16–65 years of age old living in urban and suburban areas of the country.
Materials and methods We conducted secondary analyses of data gathered in order to examine the prevalence of osteoporosis and osteopenia among 16–65-year-old Bangladeshi women. As part of a larger study, 500 women between the ages of 16 and 65 years were recruited from the reproductive and gynecology clinics of Mitford Hospital, Dhaka, Bangladesh, between Dec 2010 and May 2011. The methods for the larger study have been reported in detail elsewhere [17]. Briefly, recruitment was designed to achieve a balanced sample for each age bracket. As the total sample size was 500 women, we tried to recruit approximately 10 women per age category. Women were excluded from participating if they were pregnant, had surgically induced menopause, had a known secondary cause of osteoporosis such as a thyroid/endocrine condition (e.g., Cushing’s disease, hyperthyroidism, hyperparathyroidism, malabsorption, or hypogonadism), or prolonged immobilization. Two trained female research assistants approached women attending different reproductive and gynecology clinics to inquire whether they were interested in participating in the study. Those who agreed were asked to provide written informed consent; parental consent was obtained from patients under18 years of age. All procedures were approved by the Institutional Review Board of Bangladesh University of Health Science. After recruitment, women were asked to complete a questionnaire. Data obtained included current age, age at menarche, age at menopause among postmenopausal women, gravidity, education, household income, use of contraceptives, smoking or tobacco chewing habit, weightbearing physical activity, and family history of osteoporosis. We measured tobacco use with questions from the MONICA Project (Multinational Monitoring of Trends and Determinants in Cardiovascular Disease) smoking assessment [18]. Current smokers were defined as those who reported either regular or occasional smoking, while women who were not smoking at the time were classified as nonsmokers, regardless of whether they had smoked in the past. The International Physical Activity Questionnaire (IPAQ) was used to assess the frequency and duration of
523
physical activities [19]. We cleaned the IPAQ questionnaire data using established IPAQ protocols [20]. We calculated physical activities performed in a week in minutes for each participant based on responses obtained from the IPAQ. Kolle and colleagues [21] have reported that the total number of minutes per week should include a medium (121–234 min) to high (≥235 min) level in order to positively influence BMD levels in women. Accordingly, we categorized physical activities into two groups: no physical activities to light physical activities (≤120 min/week) versus medium to high levels of physical activity (≥121 min/ week). Height and weight were measured. BMD measures were obtained using a DXA scanner (Norland XR-46™, CooperSurgical, Inc., Trumbull, CT, USA). The primary outcome was BMD at the lumbar spine (LS) and femoral neck (FN). We calculated a coefficient of variation as a measure to evaluate reproducibility by scanning 11 healthy women twice on the same day, with the scans performed by the same technologist, as recommended [22, 23]. The sitespecific coefficient of variation was 2.34 % for the LS and 1.96 % for the FN. We used a multiple linear regression technique to estimate the BMD at the LS and FN after adjusting for weight and height. Additional robust nonparametric smoothing techniques were used to provide the estimate as a function of age, and the predicted values were obtained [24]. The models were fitted using Stata version 12 statistical software (StataCorp LP, College Station, TX). In addition, we used multiple linear regression analysis to identify the correlates of BMD at the LS and FN separately for women 16–45 and 46–65 years of age from among a set of independent variables (age, age at menarche, weight, height, gravidity, years of Depo-Provera (depot medroxyprogesterone acetate) or pill use, smoking status, and physical activity level). The skewness-kurtosis test and ladder of powers were used to determine whether the dependent variable (BMD) should be transformed and to identify the transformation.
Results The majority of the 500 women recruited for this study were of reproductive age (n = 310, 16–45 years of age; n = 190, 46–65 years of age), currently married, a homemaker, tobacco nonuser, engaged in more than 2 hours of physical activity per week, and had a household income of ≥ BDT 10,000 per month (Table 1). The mean number of schooling years in the sample was 5.3. The mean body mass index (BMI) was 23.9 kg/m2. On average, women had menarche at the age of 13 and a history of four pregnancies, and used contraceptive pills and Depo-Provera injections for 2.5 and 0.8 years, respectively. Among the women
13
524
R. A. Begum et al.
Table 1 Characteristics of study population (n = 500) Characteristics
N (%) or mean ± SD
Age, years, n (%) 16–25 26–35 36–45 46–55 56–65 School years, mean ± SD
95 (19.0) 106 (21.2) 109 (21.8) 104 (20.8) 85 (17.0) 5.3 ± 4.4
Income (monthly), n (%) < BDT 10,000 ≥ BDT 10,000
205 (41.0) 295 (59.0)
Marital status, n (%) Single/never married Married Widowed/divorced/separated Occupation Homemaker Student White collar job Blue collar job Height, cm, mean ± SD
412 (82.7) 31 (6.2) 16 (3.2) 39 (7.8) 151.1 ± 5.8
Weight, kg, mean ± SD Age at menarche, years, mean ± SD BMI, kg/m2, mean ± SD Gravidity, mean ± SD
54.5 ± 10.6 13.2 ± 1.4 23.9 ± 4.2 3.8 ± 2.8
Menopause, n (%) Yes No Years of pill use, mean ± SD Years of Depo-Provera use, mean ± SD
180 (36.0) 320 (64.0) 2.5 ± 3.9 0.8 ± 2.0
Relative with shortened height, n (%)
65 (13.0)
Relative with hip fracture history, n (%)
58 (11.6)
Current smoker, n (%)
2 (0.4)
Dried tobacco chewer, n (%)
97 (19.4)
Physical activities >120 min/week, n (%)
336 (67.2)
52 (10.4) 381 (76.2) 67 (13.4)
BDT = Bangladeshi Taka (currency), BDT 77 = US $1 SD standard deviation
in the sample, 13 % and 11.6 % had close relatives with a history of loss of height and hip fractures, respectively. Table 2 and Fig. 1 (a, b) show the relationship between age and BMD at the LS and FN after adjusting for weight and height using linear equation and smoothing techniques. The mean adjusted BMD was highest at the LS and FN for women 16–19 (1.001–1.007 g/cm2) and 16–22 years of age (0.880–0.888 g/cm2), respectively, and gradually declined with increasing age thereafter. Actual values of BMD were used for analysis, as they exhibited nearly normal distribution. Multiple linear
13
Table 2 Bone mineral density at lumbar spine and femoral neck (by age) adjusted by weight and height Age
Number of women
Lumbar spine BMD (g/cm2)
Femoral neck BMD (g/cm2)
16 17 18 19 20 21 22 23 24
4 13 9 10 9 11 11 11 9
1.008 1.005 1.003 1.001 0.998 0.995 0.992 0.989 0.985
0.888 0.887 0.886 0.885 0.884 0.882 0.880 0.878 0.876
25 26 27 28 29 30 31 32 33 34
8 11 11 10 10 12 10 11 12 9
0.982 0.978 0.974 0.971 0.967 0.963 0.959 0.955 0.951 0.946
0.874 0.871 0.869 0.866 0.863 0.860 0.857 0.854 0.851 0.848
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
12 11 10 11 12 10 13 10 10 11 12 11 9 12 11 10 11 9 9 11 11 11 10 8 8 7 7 7
0.942 0.937 0.932 0.927 0.922 0.917 0.911 0.906 0.900 0.893 0.888 0.882 0.876 0.869 0.862 0.855 0.848 0.840 0.833 0.825 0.817 0.810 0.802 0.794 0.786 0.777 0.769 0.761
0.844 0.840 0.836 0.832 0.828 0.823 0.819 0.813 0.808 0.803 0.798 0.793 0.788 0.782 0.776 0.770 0.763 0.757 0.750 0.744 0.737 0.730 0.724 0.717 0.710 0.703 0.696 0.689
525
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years Table 2 continued Age
Number of women
Lumbar spine BMD (g/cm2)
Femoral neck BMD (g/cm2)
63 64
8 9
0.753 0.745
0.682 0.676
65
10
0.737
0.669
1 .75 .5
BMD g/cm2
1.25
a
1.5
BMD bone mineral density
regression analysis showed that BMD decreased at a rate of 0.0027 g/cm2 per year (p = 0.015) at the LS and 0.0046 g/ cm2 per year (p BDT 10000/month Postmenopausal Weight (kg) Height (cm) Gravidity Contraceptive pill use, years Depo-Provera use, years Tobacco use Physical activity >120 min/week Relative with hip fracture or height loss
Lumbar spine bone mineral density
Femoral neck bone mineral density
Coefficient
p value
R2
Coefficient
p value
R2
−0.0027 0.0008 −0.0080 0.0218 0.0008 0.0037 0.0031 0.0012 0.0059 −0.0022 −0.0119 0.0127 −0.0154
0.015 0.658 0.560 0.325 0.861
ORIGINAL ARTICLE
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years Rowshan Ara Begum · Liaquat Ali · Osamu Takahashi · Tsuguya Fukui · Mahbubur Rahman
Received: 8 August 2014 / Accepted: 30 December 2014 / Published online: 28 January 2015 © The Japanese Orthopaedic Association 2015
Abstract Background Studies examining the reference values of bone mineral density (BMD) and their patterns at different ages are largely based on data generated from developed countries. The objective of this study was to estimate reference values of BMD, along with their correlates, for women living in urban and suburban areas of Bangladesh. Methods Dual-energy X-ray absorptiometry scans were performed on 500 women 16–65 years of age. Reference values at the lumbar spine (LS) and femoral neck (FN) were estimated after adjusting for height and weight. In addition, multiple linear regression analysis was used to examine correlates of BMD at the LS and FN. Results Mean BMD was highest at the LS and FN in women 16–19 (1.001–1.007 g/cm2) and 16–22 years of age (0.880–0.888 g/cm2), respectively, and gradually declined with increasing age thereafter. BMD decreased at an annual rate of 0.0027 g/cm2 at the LS and 0.0046 g/cm2 at the FN among women aged 16–45 years. For women 46–65 years R. A. Begum Department of Reproductive and Child Health, Bangladesh University of Health Science, Mirpur, Dhaka, Bangladesh L. Ali Department of Biochemistry and Cell Biology, Bangladesh University of Health Science, Mirpur, Dhaka, Bangladesh O. Takahashi · T. Fukui Center for Clinical Epidemiology, St. Luke’s International Hospital, Akashi‑cho 9‑1, Chuo‑ku, Tokyo 104‑8560, Japan M. Rahman (*) Department of Obstetrics and Gynecology and the Center for Interdisciplinary Research in Women’s Health, University of Texas Medical Branch, 301 University Blvd., John Sealy Annex, Room No. 3.108.D, Galveston, TX 77555‑0587, USA e-mail: [email protected]
13
of age, the respective figures were 0.0073 and 0.0083 g/ cm2. In both age groups, body weight was positively associated with BMD at both sites, and with height only at the LS. In addition, years of pill use was positively associated with BMD at the LS among women aged 16–45 years. BMD levels at both sites began to decline during the early twenties among Bangladeshi women. Conclusions Age-specific BMD data generated in this study could be useful for interpreting bone densitometry data among women in Bangladesh and other South Asian countries.
Introduction Osteoporosis, characterized by low bone mass and gradual deterioration of bone tissue, is the most common metabolic bone disease among women [1]. A growing elderly population, vitamin D deficiency, and inadequate dietary calcium intake are considered contributing factors to the overall burden of osteoporosis in developing countries [2, 3]. It is expected that within several decades, 75 % of hip fractures worldwide will occur in the developing countries, as this burden has been stabilized, or is even decreasing, in the developed world [2–4]. Accurate measurement of bone density using a dual-energy X-ray absorptiometry (DXA) scan is very important for identifying low bone mineral density (BMD) in assessing future fracture risk. BMD measured using the DXA technique is considered an important predictor of future fractures [5]. One of the key determinants for identifying low bone density situations such as osteopenia and osteoporosis is the reference value of BMD used for DXA scans, which is largely based on values that have been determined for populations in developed or other affluent countries [6–16]. Moreover, the majority of studies
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years
examining BMD patterns at different age levels are also severely limited by the use of data generated from affluent countries, whose sociodemographic characteristics are different from those of South Asian countries. This is a serious limitation in the literature. The objective of this study was to estimate the reference BMD values at the lumbar spine and femoral neck, along with their correlates, based on Bangladeshi women 16–65 years of age old living in urban and suburban areas of the country.
Materials and methods We conducted secondary analyses of data gathered in order to examine the prevalence of osteoporosis and osteopenia among 16–65-year-old Bangladeshi women. As part of a larger study, 500 women between the ages of 16 and 65 years were recruited from the reproductive and gynecology clinics of Mitford Hospital, Dhaka, Bangladesh, between Dec 2010 and May 2011. The methods for the larger study have been reported in detail elsewhere [17]. Briefly, recruitment was designed to achieve a balanced sample for each age bracket. As the total sample size was 500 women, we tried to recruit approximately 10 women per age category. Women were excluded from participating if they were pregnant, had surgically induced menopause, had a known secondary cause of osteoporosis such as a thyroid/endocrine condition (e.g., Cushing’s disease, hyperthyroidism, hyperparathyroidism, malabsorption, or hypogonadism), or prolonged immobilization. Two trained female research assistants approached women attending different reproductive and gynecology clinics to inquire whether they were interested in participating in the study. Those who agreed were asked to provide written informed consent; parental consent was obtained from patients under18 years of age. All procedures were approved by the Institutional Review Board of Bangladesh University of Health Science. After recruitment, women were asked to complete a questionnaire. Data obtained included current age, age at menarche, age at menopause among postmenopausal women, gravidity, education, household income, use of contraceptives, smoking or tobacco chewing habit, weightbearing physical activity, and family history of osteoporosis. We measured tobacco use with questions from the MONICA Project (Multinational Monitoring of Trends and Determinants in Cardiovascular Disease) smoking assessment [18]. Current smokers were defined as those who reported either regular or occasional smoking, while women who were not smoking at the time were classified as nonsmokers, regardless of whether they had smoked in the past. The International Physical Activity Questionnaire (IPAQ) was used to assess the frequency and duration of
523
physical activities [19]. We cleaned the IPAQ questionnaire data using established IPAQ protocols [20]. We calculated physical activities performed in a week in minutes for each participant based on responses obtained from the IPAQ. Kolle and colleagues [21] have reported that the total number of minutes per week should include a medium (121–234 min) to high (≥235 min) level in order to positively influence BMD levels in women. Accordingly, we categorized physical activities into two groups: no physical activities to light physical activities (≤120 min/week) versus medium to high levels of physical activity (≥121 min/ week). Height and weight were measured. BMD measures were obtained using a DXA scanner (Norland XR-46™, CooperSurgical, Inc., Trumbull, CT, USA). The primary outcome was BMD at the lumbar spine (LS) and femoral neck (FN). We calculated a coefficient of variation as a measure to evaluate reproducibility by scanning 11 healthy women twice on the same day, with the scans performed by the same technologist, as recommended [22, 23]. The sitespecific coefficient of variation was 2.34 % for the LS and 1.96 % for the FN. We used a multiple linear regression technique to estimate the BMD at the LS and FN after adjusting for weight and height. Additional robust nonparametric smoothing techniques were used to provide the estimate as a function of age, and the predicted values were obtained [24]. The models were fitted using Stata version 12 statistical software (StataCorp LP, College Station, TX). In addition, we used multiple linear regression analysis to identify the correlates of BMD at the LS and FN separately for women 16–45 and 46–65 years of age from among a set of independent variables (age, age at menarche, weight, height, gravidity, years of Depo-Provera (depot medroxyprogesterone acetate) or pill use, smoking status, and physical activity level). The skewness-kurtosis test and ladder of powers were used to determine whether the dependent variable (BMD) should be transformed and to identify the transformation.
Results The majority of the 500 women recruited for this study were of reproductive age (n = 310, 16–45 years of age; n = 190, 46–65 years of age), currently married, a homemaker, tobacco nonuser, engaged in more than 2 hours of physical activity per week, and had a household income of ≥ BDT 10,000 per month (Table 1). The mean number of schooling years in the sample was 5.3. The mean body mass index (BMI) was 23.9 kg/m2. On average, women had menarche at the age of 13 and a history of four pregnancies, and used contraceptive pills and Depo-Provera injections for 2.5 and 0.8 years, respectively. Among the women
13
524
R. A. Begum et al.
Table 1 Characteristics of study population (n = 500) Characteristics
N (%) or mean ± SD
Age, years, n (%) 16–25 26–35 36–45 46–55 56–65 School years, mean ± SD
95 (19.0) 106 (21.2) 109 (21.8) 104 (20.8) 85 (17.0) 5.3 ± 4.4
Income (monthly), n (%) < BDT 10,000 ≥ BDT 10,000
205 (41.0) 295 (59.0)
Marital status, n (%) Single/never married Married Widowed/divorced/separated Occupation Homemaker Student White collar job Blue collar job Height, cm, mean ± SD
412 (82.7) 31 (6.2) 16 (3.2) 39 (7.8) 151.1 ± 5.8
Weight, kg, mean ± SD Age at menarche, years, mean ± SD BMI, kg/m2, mean ± SD Gravidity, mean ± SD
54.5 ± 10.6 13.2 ± 1.4 23.9 ± 4.2 3.8 ± 2.8
Menopause, n (%) Yes No Years of pill use, mean ± SD Years of Depo-Provera use, mean ± SD
180 (36.0) 320 (64.0) 2.5 ± 3.9 0.8 ± 2.0
Relative with shortened height, n (%)
65 (13.0)
Relative with hip fracture history, n (%)
58 (11.6)
Current smoker, n (%)
2 (0.4)
Dried tobacco chewer, n (%)
97 (19.4)
Physical activities >120 min/week, n (%)
336 (67.2)
52 (10.4) 381 (76.2) 67 (13.4)
BDT = Bangladeshi Taka (currency), BDT 77 = US $1 SD standard deviation
in the sample, 13 % and 11.6 % had close relatives with a history of loss of height and hip fractures, respectively. Table 2 and Fig. 1 (a, b) show the relationship between age and BMD at the LS and FN after adjusting for weight and height using linear equation and smoothing techniques. The mean adjusted BMD was highest at the LS and FN for women 16–19 (1.001–1.007 g/cm2) and 16–22 years of age (0.880–0.888 g/cm2), respectively, and gradually declined with increasing age thereafter. Actual values of BMD were used for analysis, as they exhibited nearly normal distribution. Multiple linear
13
Table 2 Bone mineral density at lumbar spine and femoral neck (by age) adjusted by weight and height Age
Number of women
Lumbar spine BMD (g/cm2)
Femoral neck BMD (g/cm2)
16 17 18 19 20 21 22 23 24
4 13 9 10 9 11 11 11 9
1.008 1.005 1.003 1.001 0.998 0.995 0.992 0.989 0.985
0.888 0.887 0.886 0.885 0.884 0.882 0.880 0.878 0.876
25 26 27 28 29 30 31 32 33 34
8 11 11 10 10 12 10 11 12 9
0.982 0.978 0.974 0.971 0.967 0.963 0.959 0.955 0.951 0.946
0.874 0.871 0.869 0.866 0.863 0.860 0.857 0.854 0.851 0.848
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
12 11 10 11 12 10 13 10 10 11 12 11 9 12 11 10 11 9 9 11 11 11 10 8 8 7 7 7
0.942 0.937 0.932 0.927 0.922 0.917 0.911 0.906 0.900 0.893 0.888 0.882 0.876 0.869 0.862 0.855 0.848 0.840 0.833 0.825 0.817 0.810 0.802 0.794 0.786 0.777 0.769 0.761
0.844 0.840 0.836 0.832 0.828 0.823 0.819 0.813 0.808 0.803 0.798 0.793 0.788 0.782 0.776 0.770 0.763 0.757 0.750 0.744 0.737 0.730 0.724 0.717 0.710 0.703 0.696 0.689
525
Bone mineral density: reference values and correlates for Bangladeshi women aged 16–65 years Table 2 continued Age
Number of women
Lumbar spine BMD (g/cm2)
Femoral neck BMD (g/cm2)
63 64
8 9
0.753 0.745
0.682 0.676
65
10
0.737
0.669
1 .75 .5
BMD g/cm2
1.25
a
1.5
BMD bone mineral density
regression analysis showed that BMD decreased at a rate of 0.0027 g/cm2 per year (p = 0.015) at the LS and 0.0046 g/ cm2 per year (p BDT 10000/month Postmenopausal Weight (kg) Height (cm) Gravidity Contraceptive pill use, years Depo-Provera use, years Tobacco use Physical activity >120 min/week Relative with hip fracture or height loss
Lumbar spine bone mineral density
Femoral neck bone mineral density
Coefficient
p value
R2
Coefficient
p value
R2
−0.0027 0.0008 −0.0080 0.0218 0.0008 0.0037 0.0031 0.0012 0.0059 −0.0022 −0.0119 0.0127 −0.0154
0.015 0.658 0.560 0.325 0.861
