BLOOD DONORS AND BLOOD COLLECTION Predictors of hemoglobin in Danish blood donors: results from the Danish Blood Donor Study Sebastian R. Kotze ,1 Ole B. Pedersen,2 Mikkel S. Petersen,1 Erik Sørensen,3 Lise W. Thørner,3 Cecilie J. Sørensen,1 Andreas S. Rigas,3 Henrik Hjalgrim,4 Klaus Rostgaard,4 Henrik Ullum,3 and Christian Erikstrup1
BACKGROUND: It is well known that blood donors are at increased risk of iron deficiency and subsequent development of iron deficiency anemia. We aimed to investigate the effect of factors influencing hemoglobin (Hb) levels. STUDY DESIGN AND METHODS: Initiated in 2010, the Danish Blood Donor Study is a population-based study and biobank. We performed multivariable linear regression analysis to assess the effects of donation activity, physiologic and lifestyle factors, and diet on Hb levels among 15,197 donors. We also performed multivariable logistic regression to evaluate the effects of these factors on the risk of having low Hb (defined as Hb below the 10th percentile among men and women, respectively) and of a decrease in Hb greater than 0.5 mmol/L (0.8 g/dL) between successive donations. All analyses were performed stratified for sex and smoking status. We also tested a previously used model for the prediction of Hb. RESULTS: The strongest predictors of Hb and risk of low Hb were low ferritin (25), resulting in higher Hb levels. Additionally, high BMI would indicate large dietary intake of iron, protecting against low Hb. Use of oral contraception shortens menstruation and reduces menstrual bleeding.15 Menopause would be expected to be associated with higher Hb, since there is no menstrual blood loss in this group. The aim of this study was to determine the effects of lifestyle factors, as well as previously used predictors such as donation history, previous Hb levels, ferritin, and transferrin on Hb levels. We also aimed to test a previously used model for the prediction of Hb levels,16 and then test whether adjustment for additional factors available to us in this study, for example, smoking, would improve the accuracy of the model. It is hoped that a better understanding of the factors influencing Hb levels among blood donors can contribute toward improved donor management with lower risk to donor health and fewer donor deferrals, providing better tools to blood bank professionals for the assessment of Hb levels and iron status.
blood donors into the Danish Blood Donor Study was initiated March 1, 2010, in the Capital and Central Regions of Denmark. Currently, more than 90,000 blood donors have been included. Blood donors aged 18 to 67 years are eligible for inclusion. All participants are healthy, nonpregnant, and nonlactating adults. During blood donation, donors are informed about the study and invited to participate. Preliminary estimates indicate a participation rate of more than 95% among the invitees.17 Participating blood donors complete a questionnaire and grant permission for routine blood samples to be used for research. A total of 25,877 donors were included in the study between March 1 and December 31, 2010. Hb and ferritin levels were measured for 8289 women and 9225 men. Of these, complete questionnaire data were available for 7141 women and 8056 men. A total of 4203 women and 5308 men had donated again within 6 months of inclusion. The cutoff levels used for the Hb screen in Danish blood banks is 7.8 mmol/L (12.6 g/dL) for women and 8.4 mmol/L (13.6 g/dL) for men for Hb measured in predonation samples. For Hb levels measured in samples taken after donation, the cutoff levels are 7.0 mmol/L (11.3 g/ dL) for women and 7.8 mmol/L (12.6 g/dL) for men. Additionally, a decrease in Hb greater than 1.3 mmol/L (2.1 g/ dL) can result in deferral, and the donor is advised to contact his or her primary physician. Both male and female donors are permitted to donate up to four times per year. At baseline, measurements were performed the day after donation and, thus, did not lead to deferral for the current donation. For the subsequent donation, donors with low Hb values from the baseline visit had their Hb measured before donation. These measurements may have led to deferral and were included in this study.
Questionnaire data The questionnaire ascertained information about smoking status, alcohol consumption, physical activity, diet, supplemental vitamin and iron intake, height, weight, and waist circumference measurements and in female donors, history of child birth, use of oral contraceptives, and menopausal status using previously validated questions from other studies, for example, Short Form 12 for health-related quality of life.17 After completion, the questionnaires were scanned on a high-throughput scanner and processed for automated data capture by optical character recognition software (OpenForms360, NSI Europe GmbH, Wetzlar, Germany). Trained personnel then checked the resulting data before final validation was performed. The final validation of captured data showed an error rate of less than 0.25%.
MATERIALS AND METHODS Study design
Laboratory analysis
The Danish Blood Donor Study is a large, prospective, population-based study and biobank. The inclusion of
Hb levels were measured in venous samples taken before donation (in the Capital Region), using HemoCue
1304 TRANSFUSION Volume 55, June 2015
PREDICTORS OF Hb
€ (HemoCue AB, Angelholm, Sweden), and after donation (in the Central Region) on a Sysmex KX21N (Sysmex, Ramsey, MN). Ferritin was measured in thawed EDTAanticoagulated plasma samples using a commercially available assay on an automated system (Ortho Vitros 5600, Ortho Clinical Diagnostics, Rochester, NY) validated as previously described.9,18 Hb was measured after the donation had taken place, and thus no donors were excluded from the study due to failed Hb screens, and all subsequent Hb measurements were recorded, even if the result was below the level required for donation. During the study period, no donors were placed on iron supplementation on the basis of ferritin levels measured in the study. However, in 2010, the blood banks in the Capital Region were in the process of initiating routine ferritin measurements.19 An unknown number of donors were on this basis offered iron supplements if their ferritin was less than 50 ng/mL.
Statistical analysis Multivariable linear regression analysis was applied to assess the association between Hb levels and age, method of measurement, low ferritin (25 years old) to be associated with lower Hb levels (coef., 20.08 mmol/L; 95% CI, 20.12 to 20.04) and increased risk of low Hb (odds ratio [OR], 1.52; 95% CI, Volume 55, June 2015 TRANSFUSION 1305
ET AL. KOTZE
TABLE 1. Age distribution of participants* Age group (years) 18-30 31-40 41-50 51-60 601 Total
Women
Men
2425 (34.0) 1718 (24.0) 1669 (23.4) 1001 (14.0) 328 (4.6) 7141 (100)
1967 (24.4) 2177 (27.0) 2049 (25.4) 1329 (16.5) 543 (6.7) 8056 (100)
of having a Hb level below the 10th percentile (Tables 2 and 3).
Oral contraceptives, menopause, and childbirth
1.11-2.09) in nonsmoking men, but not in any other groups (data not shown).
The use of oral contraceptives correlated negatively with Hb and was associated with increased risk of having Hb below the 10th percentile, while menopause was positively associated with Hb and was associated with a reduced risk of having a Hb level below the 10th percentile (Tables 2 and 3). There was no effect of contraceptives or menopause on the risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL). Nor was there any effect of childbirth in any of the models.
Smoking status
Diet and iron and vitamin supplementation
The proportion of smokers was 17.6% for women and 16.0% for men. Mean Hb levels were significantly higher in smokers than in nonsmokers (women—0.24 mmol/L [0.39 g/dL], 95% CI 0.21-0.27 mmol/L; men—0.18 mmol/L [0.29 g/dL], 95% CI 0.15-0.21 mmol/L).
We found iron supplementation to have a negative association with Hb in all groups and to be associated with an increased risk of having Hb below the 10th percentile in men, as well as an increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) in smoking women and in nonsmoking men (Tables 2-4). There were no consistently significant dietary factors in any of the models, and the small effects of vitamin supplementation were inconsistent (Tables 2 and 3).
* Age distribution of study participants is presented in absolute numbers (%).
Ferritin Among nonsmokers, Hb levels were lower among participants with low ferritin (women—0.25 mmol/L [0.40 g/dL] lower; men—0.35 mmol/L [0.56 g/dL] lower; Table 2). Low ferritin was also associated with increased risk of having Hb levels below the 10th percentile (women—OR, 3.3; men—OR, 4.8; Table 3). Low ferritin was also associated with increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) between donations (women— OR, 1.7; men—OR, 2.0; Table 4). The ROC plots for ferritin are shown in Figs. 1 and 2. We found area under the curve values to be low, lowest for smoking men (0.51) and highest for nonsmoking women (0.63). The sensitivity and specificity of using a ferritin cutoff of 15 ng/mL as a predictor of low Hb ranged from 0.40 and 0.79 in nonsmoking women to 0.8 and 0.97 in nonsmoking men. Results were similar for smokers.
Donation history The total number of donations was positively associated with Hb levels in nonsmokers, although effects were minimal in the multivariable analysis (Table 2). Baseline Hb was correlated with an increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) at the subsequent donation.
BMI and physical activity Among men, BMI correlated positively with Hb and was associated with reduced risk of having Hb below the 10th percentile, while physical activity during leisure correlated negatively with Hb and was associated with increased risk 1306 TRANSFUSION Volume 55, June 2015
Predictive model Using the model developed by Baart and colleagues in a Dutch study, we found similar effects when using the same predictors. We found greater R2 values for both men and women than in the Dutch study (0.63 vs. 0.45 in men and 0.55 vs. 0.40 in women). However, there was little extra effect of including ferritin and smoking status in the model (R2 values were 0.64 for men and 0.56 for women).
DISCUSSION There was a significant difference in Hb levels between smokers and nonsmokers, and several analyses showed interaction with smoking; we therefore stratified for current smoking status in all subsequent analyses. The strongest predictors of Hb were ferritin and iron supplementation, and these effects were similar to that of smoking. These factors also had the largest effect on the risk of having low Hb. No dietary factors were found to be consistently significant in multivariable models predicting Hb levels, risk of having low Hb, or risk of a decrease in Hb levels of more than 0.5 mmol/L (0.8 g/dL). We previously reported ferritin levels to be strongly correlated with previous donation activity (total number of donations and number of donations in the previous 3 years).12 In this study, we found the association of ferritin with Hb to be substantial (mean Hb, 0.26 mmol/L lower in women and 0.36 mmol/L lower in men) and, indeed, ferritin was the strongest predictor of Hb levels. In
20.14† 0.04† 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02† 20.01
0.02 20.22† 0.06† 0.01 0.00 0.02 20.03† 0.12† 0.00 20.11† 0.06† 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.04† 0.01
to to to to to to to to to to to to to to to to to to to to 20.21† 0.04 0.00 20.01 0.01 20.02 20.05† 0.01 20.02 0.00 20.02
0.02 20.33† 0.02 0.02† 0.00 0.03 20.08† 0.15† 20.06
Coef.
20.28 20.02 20.01 20.02 0.00 20.04 20.08 20.02 20.06 20.04 20.07
to to to to to to to to to to to
to to to to to to to to to 20.14† 0.10 0.01 0.00 0.01 0.00 20.03† 0.03 0.03 0.04 0.04
0.07 20.26† 0.06 0.03† 0.00 0.09 0.00† 0.25† 0.02
95% CI 20.03 20.41 20.02 0.00 0.00 20.03 20.16 0.05 20.14
Coef.
20.49† 20.04† 0.00 20.01 0.00 0.00 20.01 20.03† 20.01 0.01 0.02
20.04† 20.35† 0.02† 0.00 0.00† 20.09†
20.56 20.07 0.00 20.02 0.00 20.01 20.02 20.04 20.03 20.01 20.01
20.05 20.42 0.01 20.01 0.00 20.11
to to to to to to to to to to to
to to to to to to
20.43† 20.02† 0.00 0.00 0.01 0.01 0.00 20.02† 0.00 0.03 0.04
20.03† 20.28† 0.04† 0.00 0.00† 20.06†
95% CI
R2 = 0.18
Nonsmokers
20.75† 20.01 0.00 20.01 0.00 20.01 0.00 20.05† 0.00 20.01 0.03
20.02 20.41† 0.00 0.01 0.01† 20.11†
Coef.
Men (n = 8056) Smokers
20.94 20.07 20.01 20.03 20.02 20.04 20.03 20.08 20.04 20.05 20.02
20.05 20.60 20.03 20.01 0.00 20.18
to to to to to to to to to to to
to to to to to to
20.55† 0.06 0.01 0.01 0.01 0.02 0.03 20.01† 0.04 0.03 0.08
0.02 0.21† 0.04 0.02 0.01† 20.05†
95% CI
R2 = 0.19
* Results of the multivariable linear regression analysis with predictors of Hb levels. In addition to the variables presented in the table, the analysis was also adjusted for method of Hb measurement. † Denotes significant results.
20.16 0.01 0.00 20.01 0.00 20.01 20.01 20.01 20.02 0.01 20.03
0.00 20.25† 0.04† 0.00 0.00 20.01 20.06† 0.08† 20.03
95% CI 20.02 20.28 0.03 0.00 0.00 20.03 20.09 0.04 20.07
Coef.
Age group (18-30, 31-40, 41-50, 51-60, 601 years) Low ferritin (
BACKGROUND: It is well known that blood donors are at increased risk of iron deficiency and subsequent development of iron deficiency anemia. We aimed to investigate the effect of factors influencing hemoglobin (Hb) levels. STUDY DESIGN AND METHODS: Initiated in 2010, the Danish Blood Donor Study is a population-based study and biobank. We performed multivariable linear regression analysis to assess the effects of donation activity, physiologic and lifestyle factors, and diet on Hb levels among 15,197 donors. We also performed multivariable logistic regression to evaluate the effects of these factors on the risk of having low Hb (defined as Hb below the 10th percentile among men and women, respectively) and of a decrease in Hb greater than 0.5 mmol/L (0.8 g/dL) between successive donations. All analyses were performed stratified for sex and smoking status. We also tested a previously used model for the prediction of Hb. RESULTS: The strongest predictors of Hb and risk of low Hb were low ferritin (25), resulting in higher Hb levels. Additionally, high BMI would indicate large dietary intake of iron, protecting against low Hb. Use of oral contraception shortens menstruation and reduces menstrual bleeding.15 Menopause would be expected to be associated with higher Hb, since there is no menstrual blood loss in this group. The aim of this study was to determine the effects of lifestyle factors, as well as previously used predictors such as donation history, previous Hb levels, ferritin, and transferrin on Hb levels. We also aimed to test a previously used model for the prediction of Hb levels,16 and then test whether adjustment for additional factors available to us in this study, for example, smoking, would improve the accuracy of the model. It is hoped that a better understanding of the factors influencing Hb levels among blood donors can contribute toward improved donor management with lower risk to donor health and fewer donor deferrals, providing better tools to blood bank professionals for the assessment of Hb levels and iron status.
blood donors into the Danish Blood Donor Study was initiated March 1, 2010, in the Capital and Central Regions of Denmark. Currently, more than 90,000 blood donors have been included. Blood donors aged 18 to 67 years are eligible for inclusion. All participants are healthy, nonpregnant, and nonlactating adults. During blood donation, donors are informed about the study and invited to participate. Preliminary estimates indicate a participation rate of more than 95% among the invitees.17 Participating blood donors complete a questionnaire and grant permission for routine blood samples to be used for research. A total of 25,877 donors were included in the study between March 1 and December 31, 2010. Hb and ferritin levels were measured for 8289 women and 9225 men. Of these, complete questionnaire data were available for 7141 women and 8056 men. A total of 4203 women and 5308 men had donated again within 6 months of inclusion. The cutoff levels used for the Hb screen in Danish blood banks is 7.8 mmol/L (12.6 g/dL) for women and 8.4 mmol/L (13.6 g/dL) for men for Hb measured in predonation samples. For Hb levels measured in samples taken after donation, the cutoff levels are 7.0 mmol/L (11.3 g/ dL) for women and 7.8 mmol/L (12.6 g/dL) for men. Additionally, a decrease in Hb greater than 1.3 mmol/L (2.1 g/ dL) can result in deferral, and the donor is advised to contact his or her primary physician. Both male and female donors are permitted to donate up to four times per year. At baseline, measurements were performed the day after donation and, thus, did not lead to deferral for the current donation. For the subsequent donation, donors with low Hb values from the baseline visit had their Hb measured before donation. These measurements may have led to deferral and were included in this study.
Questionnaire data The questionnaire ascertained information about smoking status, alcohol consumption, physical activity, diet, supplemental vitamin and iron intake, height, weight, and waist circumference measurements and in female donors, history of child birth, use of oral contraceptives, and menopausal status using previously validated questions from other studies, for example, Short Form 12 for health-related quality of life.17 After completion, the questionnaires were scanned on a high-throughput scanner and processed for automated data capture by optical character recognition software (OpenForms360, NSI Europe GmbH, Wetzlar, Germany). Trained personnel then checked the resulting data before final validation was performed. The final validation of captured data showed an error rate of less than 0.25%.
MATERIALS AND METHODS Study design
Laboratory analysis
The Danish Blood Donor Study is a large, prospective, population-based study and biobank. The inclusion of
Hb levels were measured in venous samples taken before donation (in the Capital Region), using HemoCue
1304 TRANSFUSION Volume 55, June 2015
PREDICTORS OF Hb
€ (HemoCue AB, Angelholm, Sweden), and after donation (in the Central Region) on a Sysmex KX21N (Sysmex, Ramsey, MN). Ferritin was measured in thawed EDTAanticoagulated plasma samples using a commercially available assay on an automated system (Ortho Vitros 5600, Ortho Clinical Diagnostics, Rochester, NY) validated as previously described.9,18 Hb was measured after the donation had taken place, and thus no donors were excluded from the study due to failed Hb screens, and all subsequent Hb measurements were recorded, even if the result was below the level required for donation. During the study period, no donors were placed on iron supplementation on the basis of ferritin levels measured in the study. However, in 2010, the blood banks in the Capital Region were in the process of initiating routine ferritin measurements.19 An unknown number of donors were on this basis offered iron supplements if their ferritin was less than 50 ng/mL.
Statistical analysis Multivariable linear regression analysis was applied to assess the association between Hb levels and age, method of measurement, low ferritin (25 years old) to be associated with lower Hb levels (coef., 20.08 mmol/L; 95% CI, 20.12 to 20.04) and increased risk of low Hb (odds ratio [OR], 1.52; 95% CI, Volume 55, June 2015 TRANSFUSION 1305
ET AL. KOTZE
TABLE 1. Age distribution of participants* Age group (years) 18-30 31-40 41-50 51-60 601 Total
Women
Men
2425 (34.0) 1718 (24.0) 1669 (23.4) 1001 (14.0) 328 (4.6) 7141 (100)
1967 (24.4) 2177 (27.0) 2049 (25.4) 1329 (16.5) 543 (6.7) 8056 (100)
of having a Hb level below the 10th percentile (Tables 2 and 3).
Oral contraceptives, menopause, and childbirth
1.11-2.09) in nonsmoking men, but not in any other groups (data not shown).
The use of oral contraceptives correlated negatively with Hb and was associated with increased risk of having Hb below the 10th percentile, while menopause was positively associated with Hb and was associated with a reduced risk of having a Hb level below the 10th percentile (Tables 2 and 3). There was no effect of contraceptives or menopause on the risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL). Nor was there any effect of childbirth in any of the models.
Smoking status
Diet and iron and vitamin supplementation
The proportion of smokers was 17.6% for women and 16.0% for men. Mean Hb levels were significantly higher in smokers than in nonsmokers (women—0.24 mmol/L [0.39 g/dL], 95% CI 0.21-0.27 mmol/L; men—0.18 mmol/L [0.29 g/dL], 95% CI 0.15-0.21 mmol/L).
We found iron supplementation to have a negative association with Hb in all groups and to be associated with an increased risk of having Hb below the 10th percentile in men, as well as an increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) in smoking women and in nonsmoking men (Tables 2-4). There were no consistently significant dietary factors in any of the models, and the small effects of vitamin supplementation were inconsistent (Tables 2 and 3).
* Age distribution of study participants is presented in absolute numbers (%).
Ferritin Among nonsmokers, Hb levels were lower among participants with low ferritin (women—0.25 mmol/L [0.40 g/dL] lower; men—0.35 mmol/L [0.56 g/dL] lower; Table 2). Low ferritin was also associated with increased risk of having Hb levels below the 10th percentile (women—OR, 3.3; men—OR, 4.8; Table 3). Low ferritin was also associated with increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) between donations (women— OR, 1.7; men—OR, 2.0; Table 4). The ROC plots for ferritin are shown in Figs. 1 and 2. We found area under the curve values to be low, lowest for smoking men (0.51) and highest for nonsmoking women (0.63). The sensitivity and specificity of using a ferritin cutoff of 15 ng/mL as a predictor of low Hb ranged from 0.40 and 0.79 in nonsmoking women to 0.8 and 0.97 in nonsmoking men. Results were similar for smokers.
Donation history The total number of donations was positively associated with Hb levels in nonsmokers, although effects were minimal in the multivariable analysis (Table 2). Baseline Hb was correlated with an increased risk of a decrease in Hb of more than 0.5 mmol/L (0.8 g/dL) at the subsequent donation.
BMI and physical activity Among men, BMI correlated positively with Hb and was associated with reduced risk of having Hb below the 10th percentile, while physical activity during leisure correlated negatively with Hb and was associated with increased risk 1306 TRANSFUSION Volume 55, June 2015
Predictive model Using the model developed by Baart and colleagues in a Dutch study, we found similar effects when using the same predictors. We found greater R2 values for both men and women than in the Dutch study (0.63 vs. 0.45 in men and 0.55 vs. 0.40 in women). However, there was little extra effect of including ferritin and smoking status in the model (R2 values were 0.64 for men and 0.56 for women).
DISCUSSION There was a significant difference in Hb levels between smokers and nonsmokers, and several analyses showed interaction with smoking; we therefore stratified for current smoking status in all subsequent analyses. The strongest predictors of Hb were ferritin and iron supplementation, and these effects were similar to that of smoking. These factors also had the largest effect on the risk of having low Hb. No dietary factors were found to be consistently significant in multivariable models predicting Hb levels, risk of having low Hb, or risk of a decrease in Hb levels of more than 0.5 mmol/L (0.8 g/dL). We previously reported ferritin levels to be strongly correlated with previous donation activity (total number of donations and number of donations in the previous 3 years).12 In this study, we found the association of ferritin with Hb to be substantial (mean Hb, 0.26 mmol/L lower in women and 0.36 mmol/L lower in men) and, indeed, ferritin was the strongest predictor of Hb levels. In
20.14† 0.04† 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02† 20.01
0.02 20.22† 0.06† 0.01 0.00 0.02 20.03† 0.12† 0.00 20.11† 0.06† 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.04† 0.01
to to to to to to to to to to to to to to to to to to to to 20.21† 0.04 0.00 20.01 0.01 20.02 20.05† 0.01 20.02 0.00 20.02
0.02 20.33† 0.02 0.02† 0.00 0.03 20.08† 0.15† 20.06
Coef.
20.28 20.02 20.01 20.02 0.00 20.04 20.08 20.02 20.06 20.04 20.07
to to to to to to to to to to to
to to to to to to to to to 20.14† 0.10 0.01 0.00 0.01 0.00 20.03† 0.03 0.03 0.04 0.04
0.07 20.26† 0.06 0.03† 0.00 0.09 0.00† 0.25† 0.02
95% CI 20.03 20.41 20.02 0.00 0.00 20.03 20.16 0.05 20.14
Coef.
20.49† 20.04† 0.00 20.01 0.00 0.00 20.01 20.03† 20.01 0.01 0.02
20.04† 20.35† 0.02† 0.00 0.00† 20.09†
20.56 20.07 0.00 20.02 0.00 20.01 20.02 20.04 20.03 20.01 20.01
20.05 20.42 0.01 20.01 0.00 20.11
to to to to to to to to to to to
to to to to to to
20.43† 20.02† 0.00 0.00 0.01 0.01 0.00 20.02† 0.00 0.03 0.04
20.03† 20.28† 0.04† 0.00 0.00† 20.06†
95% CI
R2 = 0.18
Nonsmokers
20.75† 20.01 0.00 20.01 0.00 20.01 0.00 20.05† 0.00 20.01 0.03
20.02 20.41† 0.00 0.01 0.01† 20.11†
Coef.
Men (n = 8056) Smokers
20.94 20.07 20.01 20.03 20.02 20.04 20.03 20.08 20.04 20.05 20.02
20.05 20.60 20.03 20.01 0.00 20.18
to to to to to to to to to to to
to to to to to to
20.55† 0.06 0.01 0.01 0.01 0.02 0.03 20.01† 0.04 0.03 0.08
0.02 0.21† 0.04 0.02 0.01† 20.05†
95% CI
R2 = 0.19
* Results of the multivariable linear regression analysis with predictors of Hb levels. In addition to the variables presented in the table, the analysis was also adjusted for method of Hb measurement. † Denotes significant results.
20.16 0.01 0.00 20.01 0.00 20.01 20.01 20.01 20.02 0.01 20.03
0.00 20.25† 0.04† 0.00 0.00 20.01 20.06† 0.08† 20.03
95% CI 20.02 20.28 0.03 0.00 0.00 20.03 20.09 0.04 20.07
Coef.
Age group (18-30, 31-40, 41-50, 51-60, 601 years) Low ferritin (
