Prevalence and associated factors of anemia in a Russian population: the Ural eye and medical study

Background Although anemia is one of the leading causes of the global burden of disease, information about its prevalence in Russia is mostly missing. We therefore assessed its prevalence and associated factors in a Russian population. Methods The population-based Ural Eye and Medical Study included 5899 (80.5%) out of 7328 eligible individuals (mean age:59.0 ± 10.7 years;range:40–94 years) who underwent a standardized interview and detailed general examination. The definition of anemia was based on the hemoglobin concentration (definition #1:hemoglobin concentration < 140 g/L in men,< 130 g/L in women; definition #2:hemoglobin concentration < 130 g/L in men,< 120 g/L in women [World Health Organization definition]). Results Higher hemoglobin concentration (mean:142.6 ± 14.8 g/L; range:80-171 g/L) was associated (multivariable analysis) with male gender (P < 0.001; standardized regression coefficient beta:-0.20), higher waist-hip circumference ratio (P < 0.001;beta:0.05), higher prevalence of car ownership (P < 0.001;beta:0.05), higher blood concentrations of bilirubin (P < 0.001;beta:0.05) and triglycerides (P < 0.001;beta:0.06), lower erythrocyte sedimentation rate (P < 0.001;beta:-0.32), and shorter blood clotting time (P < 0.001;beta:-0.39). Using definition #1 and #2, anemia was detected in 1385 individuals (23.6%;95% confidence interval CI)CI:22.5,24.7) and in 453 individuals (7.7%;95%CI:7.0,8.4), respectively. Prevalence of moderate anemia (hemoglobin concenttration:110 g/L-80 g/L), detected in 165 individuals (2.8%;95%CI:2.4,3.2), increased with younger age (P = 0.008;odds ratio (OR):0.98;95%CI:0.96,0.99), female gender (P < 0.001;OR:2.52;95%CI:1.47,4.33), higher erythrocyte sedimentation rate (P < 0.001;OR:1.08;95%CI:1.06,1.09), longer blood clotting time (P < 0.001;OR:8.56;95%CI:5.68,12.9), and marginally significantly, with a lower waist-hip circumference ratio (P = 0.058;OR:0.13;95%CI:0.02,1.07). In women, it was significantly (P < 0.001) higher before menopause (8.8%;95%CI:6.4,11.1) than after menopause (3.5%;95%CI:2.8,4.3). Conclusions In this Russian population as compared to populations from countries with a similar socio-demographic index, anemia prevalence was relatively low. As in other populations, higher anemia prevalence was strongly and inversely associated with menopause, and to a minor degree, with lower waist-hip circumference ratio and lower socio-economic background.


Background
The Global Burden of Diseases, Injuries and Risk Factors (GBD) Study estimated in 2010 that the global prevalence of anemia was 32.9%, and that anemia caused 68.4 million YLDs (years lived with disability) or 8.8% of all YLDs [1,2]. That figure was higher than the figures for major depression (63.2 million YLD; 8.2%), chronic respiratory diseases (49.3 million YLDs, 6.3%) and the whole of all injuries (47.2 million YLDs; 6.3%). The anemia-associated YLDs increased in all ages between 1990 and 2010, driven by a combination of increased prevalence and population growth in the young age groups. In adulthood, the increase in the anemia-related YLDs was attributed to population growth and enhanced survival of chronic diseases associated with anemia, such as chronic kidney disease. In elderly groups, the increase in the anemia-associated YLDs was mainly due to population aging while the anemia prevalence in the elderly age groups decreased from all causes except for malaria. In most regions and age groups, anemia prevalence was higher in females than in males [1][2][3].
Despite the general importance of anemia for public health and although Russia is by area the largest, and by population one of the most populous countries worldwide, information about the prevalence of anemia in Russia and factors associated with the occurrence of anemia in Russia has been mostly missing so far. We therefore conducted this study to assess the prevalence of anemia in a population in Russia and to explore its associations with other factors [4,5].

Methods
Designed as a population-based investigation, the Ural Eye and Medical Study (UEMS) was carried out in the Ufa (urban region of Kirovskii), the capital of the Russian Republic of Bashkortostan, and in villages of the Karmaskalinsky District located in a distance of 65 km from Ufa [6]. Bashkortostan (total population: 4.07 million) is located in the west of the southern Ural Mountains. The population of Ufa with 1.1 million inhabitants includes Russians, Tatars, Bashkirs, Ukrainians and other ethnicities. While there were no exclusion criteria, the only inclusion criteria for the participation in the study were an age of 40 or more years and living in the study region.
The series of examinations started with a standardized interview which was performed by trained social workers. It consisted of more than 250 questions about socioeconomic variables (e.g., educational level), life style including diet, smoking and alcohol consumption, depression and suicidal ideation, physical activity, and medical history including known diagnosis and therapy of major systemic diseases including presence or history of previous iron-deficiency related anemia. The data were reported using the Guidelines for Accurate and Transparent Health Estimates Reporting: the GATHER statement guidelines [7].
The physical examinations included measurement of blood pressure, pulse rate and body height, body weight and hip and waist circumference. Using a dynamometer, we determined the handgrip strength. We measured hearing loss by hearing loss-related questions in the interview and by conducting Rinne's test and Weber's test. The assessment of blood samples, which were taken under fasting conditions, included a complete blood cell count and determinations of the serum concentrations of glucose, blood lipids, C-reactive protein, bilirubin, urea, creatinine, and other substances. The pulmonary function was tested by spirometry. Diabetes mellitus was defined by a glucose concentration of ≥7.0 mmol/L or a self-reported history of a physician-related diagnosis of diabetes or a history of drug treatment of diabetes. Using the Center for Epidemiologic Studies Depression Scale (CES-D) Scoresheet, we assessed the presence of depression. We applied the State-Trait Anxiety Inventory (STAI) to determine trait and state anxiety.
Anemia was defined by a hemoglobin concentration of less than 140 g/L in men and of less than 130 g/L in women (definition #1), and it was defined by a hemoglobin concentration of less than 130 g/L in men and of less than 120 g/L in women (definition #2). The reason to use two definitions of anemia was that anemia has not yet been generally defined in a consensus meeting or corresponding article. While the second anemia definition applied in our study accorded with the WHO (World Health Organization) definition of anemia, the first definition was based on the several previous population-based studies which addressed the prevalence of anemia and which used different definitions of anemia [1-5, 8, 9]. Anemia was classified into mild anemia (definition #1: hemoglobin concentration of 139 g/L to 110 g/L in men and of 129 g/L to 110 g/L in women; and definition #2: hemoglobin concentration of 129 g/L to 110 g/L in men and of 119 g/L to 110 g/L in women), and into moderate anemia (both definitions #1 and #2) with a hemoglobin concentration ranging between less than 110 g/L and 80 g/L in both gender. Severe anemia was defined by a hemoglobin concentration of less than 80 g/L.
We used a commercially available statistical software program (Statistical Package for Social Science, SPSS, version 25.0; IBM-SPSS Inc., Chicago, USA) for the statistical analysis. In a first step, we determined the mean hemoglobin concentration, expressed as mean ± standard deviation, and the mean prevalence of anemia, presented as mean and its 95% confidence intervals (CI). In a second step, we searched for associations in univariate analysis between the prevalence of anemia and other parameters. In a third step, we conducted a multivariable binary regression analysis with the prevalence of anemia as dependent variable and as independent variables all those parameters which were significantly associated with the prevalence of anemia in the univariate analysis. Finally, we assessed associations between the blood hemoglobin concentration and other parameters in a multivariable linear regression analysis. Odds ratios (OR) and their 95% CIs, the standardized regression coefficient beta and the non-standardized regression coefficient B were calculated. All P-values were two-sided and considered statistically significant when the values were less than 0.05.

Results
Out of 5889 individuals primarily participating in the Ural Eye and Medical Study, the present study included 5864 (99.4%) individuals with biochemical blood examinations. The mean age of the study population (2559 (43.6%) men) was 59.0 ± 10.7 years ( Table 1). The composition of the study population with respect to gender and age corresponded to the gender and age distribution in the Russian population according to the most recent census carried out in 2010 [9]. It showed two constrictions for the birth year groups from 1940 to 1946 and for the birth year groups from 1962 to 1970, directly and indirectly caused by the consequences of World War II.
In univariate analysis, a higher prevalence of anemia (definition #1) was associated with older age (P < 0.001) and female gender (P < 0.001) (Fig. 2). For exploring any other associations, we adjusted the analysis for age and gender. After adjusting for age and gender, a higher prevalence of anemia was associated with lower prevalence of married family status (P = 0.01), higher prevalence of Muslima religion (P = 0.07), lower body weight (P < 0.001) and lower body mass index (P < 0.001), shorter waist circumference (P < 0.001), shorter hip circumference (P = 0.001), and lower waist-to-hip circumference ratio (P < 0.001), lower monthly income (P = 0.04), lower frequency of an ownership of a car (P = 0.001), lower prevalence of a history of arterial hypertension (P = 0.002), tumbling (P = 0.08) and menopause (P < 0.001), higher prevalence of a history of irondeficiency anemia (P < 0.001), lower blood concentration of alanine transferase (P = 0.003), aspartate aminotransferase (P = 0.02), bilirubin (P < 0.001), low-density lipoproteins (P = 0.02), triglycerides (P < 0.001), cholesterol (P < 0.001), higher blood concentration of the rheumatoid factor (P = 0.001), creatinine (P = 0.001), urea (P = 0.001), residual nitrogen (P = 0.09), faster erythrocyte sedimentation rate (P < 0.001), longer blood clotting time (P < 0.001), lower prothrombin time (P = 0.05), lower blood count of leukocytes (P < 0.001), any type of Table 1 Prevalence (%) of any anemia (using two defintiions: defined as a hemoglobin concentration < 140 g/ in men and < 130 g/L in women, or defined as a hemoglobin concentration < 130 g/ in men and < 120 g/L in women), mild anemia (using two defintiions: defined as a hemoglobin concentration < 140 g/L and > 110 g/L in men; < 130 g/L and > 110 g/L in women, or defined as a hemoglobin concentration < 130 g/L and > 110 g/L in men; < 120 g/L and > 110 g/L in women) and moderate anemia (hemoglobin concentration < 110 g/L and > 80 g/L) in the Ural Eye and Medical Study, stratified by sex and age granulocytes (P < 0.001), lymphocytes and monocytes (P < 0.001), lower systolic (P = 0.01) and diastolic (P < 0.001) blood pressure, higher ankle-brachial index (P = 0.07), higher prevalence of not able to stop drinking alcohol once started (P = 0.003), of failing to do what was normally expected due to alcohol consumption (P = 0.02) and of feeling guilty after drinking (P = 0.01), higher total hearing loss score (P = 0.02), and lower manual dynamometric force (P < 0.001) ( Table 3).
In multivariable linear regression analysis, the blood hemoglobin concentration was associated (regression coefficient r 2 = 0.50) with male gender (P < 0.001), lower waist-hip circumference ratio (P < 0.001), lower prevalence of an ownership of a car (P < 0.001), lower blood concentrations of bilirubin (P < 0.001) and triglycerides (P < 0.001), higher erythrocyte sedimentation rate (P < 0.001) and longer blood clotting time (P < 0.001) Table 2 Associations of the prevalence of anemia (defined as a hemoglobin concentration < 140 g/ in men and < 130 g/L in women) in the Ural Eye and Medical Study, after adjusting for gender and age (Continued)  (Table 7). If the parameter of car ownership was replaced by the parameter of level of education, higher educational level was significantly associated with higher hemoglobin concentration (P = 0.007; beta: 0.03). Similar results were obtained if the parameters of erythrocyte sedimentation rate and blood clotting time as factors secondary to anemia were dropped from the model,

Discussion
In our population-based study in Russia, higher hemoglobin concentration was associated with male gender, higher waist-hip circumference ratio, higher level of education, higher blood concentrations of bilirubin and triglycerides, lower erythrocyte sedimentation rate and shorter blood clotting time. The prevalence of anemia defined by a hemoglobin concentration < 140 g/L in men and < 130 g/L in women was 23.6% (95% CI: 22.5, 24.7), and the prevalence of anemia defined by a hemoglobin concentration of < 130 g/L in men and < 120 g/L in women was 7.7% (95%CI: 7.0, 8.4). The prevalence of moderate anemia (hemoglobin concentration: 110 g/L to 80 g/L) was 2.8% (95%CI: 2.4, 3.2) and increased with younger age, female gender, higher erythrocyte sedimentation rate, longer blood clotting time, and marginally significantly, with a lower waist-hip circumference ratio. In women, it was significantly (P < 0.001) higher before menopause (8.8%; 95%CI: 6.4, 11.1) than after menopause (3.5%; 95%CI: 2.8,4.3) (Fig. 2). These prevalence figures for this study population in Russia were relatively low as compared with other populations. As in other populations, the prevalence of anemia was strongly associated with menopause, and to a minor degree, with a decreased lower waist-hip circumference ratio and lower socio-economic and educational background. There was a tendency for a higher anemia prevalence in the urban region of habitation (P = 0.10).
Anemia as a major public health problem has been addressed in numerous studies before, estimating the total number of individuals affected by anemia to be 1.62 billion people globally [10]. The prevalence of anemia ranged between 9% in high-income countries and 43% in countries with a relatively low degree of development [10]. Anemia prevalence was usually higher in pregnant women, women of reproductive age and in children than in men or elderly women. Global estimates for the prevalence of anemia were as high as 47% for children with an age of less than 5 years, 42% for pregnant women, and 30% for non-pregnant women with an age of 15 to 49 years [10]. Africa and Asia were the continents bearing the load of about 85% of the absolute  anemia burden in the high-risk groups. It has been estimated that globally and per year, anemia contributed to more than 591,000 perinatal deaths and 115,000 maternal deaths [11]. The figures found for our study population from Russia show that the prevalence of anemia in Russia is similar to the anemia prevalence found in highincome countries, despite the marked difference between Russia and high-income countries in the social development index (SDI) as calculated by the Global Burden of Disease Study [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]. Using the WHO definition of anemia, the prevalence of anemia in our study population was 7.7% (95%CI: 7.0, 8.4). Correspondingly, none of the study participants had a severe anemia defined by a hemoglobin concentration of less than 80 mg/L. These results are in contrast to anemia prevalences found in countries with a similar SDI as Russia but which show a markedly higher anemia prevalence including severe anemia [10,11,[13][14][15][16][17][18][19][20][21][22][23][24][25]. It suggests that the public health care system in Russia is better than it would be expected based on its SDI. It agrees with the finding of a relatively low unawareness rate for diabetes of 25% in the Russian population (Ural Eye and Medical Study; own data), a figure similar or even lower than the rate found in highincome countries.
The results obtained in our study on associations of anemia with other systemic parameters are in agreement with findings obtained in previous investigations. The main factor associated with anemia was female gender in the reproductive age (Fig. 2), in addition to a low waist-hip circumference ratio or low body mass index and a low socio-economic and educational background [1-3, 10, 11, 13-25].
The univariate analysis included many diverse parameters which were either the determinants of anemia or the sequels of anemia. Subsequently, the multivariate analysis consisted of causes and of consequences of anemia. To cite examples, the parameters of male gender, higher waist-hip circumference ratio, higher prevalence of car ownership and status before menopause were likely variables associated with the causation of anemia, while the variable of a higher blood concentration of bilirubin was probably a consequence of anemia. The diverse nature of the independent parameters in the multivariable analysis should therefore be taken into account when the results of the statistical analysis are discussed.
When the results of our study are discussed, its limitations should be taken into account. First, as for any population-based study, the participation rate is critical to assure the representativeness of the study population. With more than 80% of the eligible population taking part in our investigation, a pronounced bias in the inclusion of participants might have been unlikely. Second, our study population was composed of various ethnicities. While this multi-ethnic composition was typical for Southern Russia, the population of North-Western  Russia and Central Russia usually shows a higher percentage of Russians. To overcome this potential limitation of our study, we assessed the prevalence of anemia in dependence of the ethnic background and found that the prevalence did not differ significantly between the Russian group and the non-Russian group. The age and gender distribution in our study population was comparable to the results of the Russian census 2010 [12]. Third, there were several definitions of anemia, and we used the anemia definition based on the blood hemoglobin concentration. It lacks, however, specificity for establishing the iron status. Measurement of the concentrations of serum ferritin and transferrin receptor in combination with determinations of the hemoglobin concentrations would have been better measures. Fourth, in the statistical analysis we did not clearly differentiate between biomarkers which were probably due to anemia and determinants of anemia.

Conclusions
In conclusion, in this ethically mixed urban and rural Russian population aged 40+ years, the prevalence of anemia (defined by a hemoglobin concentration of < 130 g/L in men and of < 120 g/L in women) was relatively low (7.7%;95%CI:7.0,8.4), and it was similar to the anemia prevalence in high-income countries. As in other populations, the anemia prevalence was strongly associated with menopause, and to a minor degree, with a lower waist-hip circumference ratio and lower socioeconomic and educational background. There was a tendency for a higher anemia prevalence in the urban region of habitation (P = 0.10). These data may be of interest for assessing the burden of anemia in Russia.