In this clinic-based study conducted in 2021, about one-fifth of the Gorgan diabetic women and men were found to have anemia. Obesity, T2DM duration more than five years, albuminuria, D DKD, high triglyceride, as well as using insulin both with and without oral GLDs were independently associated with prevalence of anemia among our participants.
Several studies have reported the prevalence of anemia among patients with T2DM to be common, especially among developing countries. Our result were similar to the previous study from Iran( Mashhad city) in 2011(19.6%) and lower than another one( Tehran city) in 2014 (30.4%) . A recent meta-analysis estimated the prevalence of 35% among patients with T2DM in Africa. Other studies conducted in Kuwait, Malaysia, Brazil , California (DISTANCE study) , Greece (patients with DKD stages 2-4) , Saudi, England  and Pakistan , revealed the higher prevalence of 29.7%(5,655 /19,059), 31.7% (256/808), 34.2% (50/146), 34.7% (22,812/65), 47.8% (88/184), 55.5% (126/227), 59% (68/115) and 63% (126/200) among patients with T2DM. However, a study from India  and a community-based cohort from Australia  showed a lower prevalence of 12.13% (174 /1414) and 11.5% (178/1551), respectively. These differences may be due to the quality of health care services such as accessibility of patients to visit specialist and laboratory testing besides, selection bias could be the underlie the difference; for instance, in countries with high quality of health care services, the participants are in close follow up in specialized centers therefore were not representative of the whole population diabetic patients.
The differences observed could be the result of variations in studies' methodology and participant characteristics such as lifestyle, feeding habits, type of GLD usage, duration of T2DM, ethnicity, and mean age [1, 21].
The current study showed that T2DM duration for more than five years (regardless of glycemic indices and nephropathy) had a strong independent association with anemia, in line with some other studies [8, 26]. It seems that chronic hyperglycemia could decrease erythropoiesis and increase red blood cells (RBCs) destruction due to more exposure to inflammation and oxidative stress, and bone marrow impairment [27, 28].
In this study, although the levels of FPG and HbA1c were higher among diabetics with anemia, but statistically they were not significantly associated with anemia. However, some studies showed lower HbA1c among patients with anemia [29, 30]. They explained that deceased hemoglobin concentration and the enhanced RBC turnover in anemia of chronic disease could reduce the glycation process and consequently lead to falsely reported lower HbA1c levels .
Despite the lack of significant association between glucose indices and prevalence of anemia, using insulin had a significant association in the present study. Insulin users had potentially worse baseline characteristics rather than non-insulin users; also, using insulin could be a representative of poor control DM (allocation bias) [32, 33]. So this result is in agreement with some other studies that mentioned poor control DM has an association with anemia [12, 34]. Since neuropathy is common in patients with poor glycemic control, one of the reasons for the increased risk of anemia is the impairment of production and release of erythropoietin stimulated by the autonomic nervous system . DM could negatively affect interstitial and peritubular structures (where erythropoietin is produced), and anemia could be the result of Decreased erythropoietin production by the failing kidney. Moreover, the exposure of erythroblasts or the mature erythrocytes to oxidative stress (due to glucose toxicity) could cause erythrocyte dysfunction . Furthermore, metformin is the first-line choice for T2DM management unless in the patients with e-GFR <45 ml/min/1.73 m2, according to ADA 2021 guidelines. It has been reported that metformin interferes with cyanocobalamin absorption and is associated with vitamin B12 deficiency, resulting in an increased risk of anemia among patients with T2DM . We still doubt whether the long duration of poor-control T2DM could increase the risk of anemia or not.
We found that obesity and high triglyceride were independently associated with prevalence of anemia. There are controversial results in this area as some studies reported that obesity is positively associated with anemia. They believe that obesity could lead to insulin resistance which could result in a hyperglycemic state. The adipose tissue is the source of different cytokines. The increase in inflammatory activity of adipose tissue in obese patients could cause an increase in hepcidin levels, causing a reduction in serum iron and limiting the availability of iron . In this line, studies reported an obesity paradox in anemia and observed that normal/overweight T2DM patients were more anemic than obese patients . However, some studies could not show any independent association between anemia and BMI among T2DM patients [38, 39]. Over nutrition may be associated with increased consumption of protein, iron, and other micronutrients with a protective effect against iron and/ or B12 deficiency in.diabetic population.
We found that albuminuria> 30 mg/24hr and e-GFR 0 ml/min/1.73 m2 as two important parameters of Diabetic kidney disease (DKD) had a significant independent association with anemia among patients with T2DM. Our findings were in accordance with other studies showing a higher frequency of anemia in patients with T2DM and nephropathy [41-43]. A large multicenter US study of the kidney early evaluation program (KEEP) demonstrated that development and aggravation of anemia in patients with DKD had a statistical relationship with severity of albuminuria and e-GFR . DM could damage tubulointerstitial tissue (associated with the degree of albuminuria) in the early stage, even before any reduction in e-GFR. It could cause decreased erythropoietin production and iron metabolism impairment leading to reduced production of RBCs . Blood urea may increase due to renal dysfunction and could negatively affect RBC’s lifespan .
We found no significant difference in the prevalence of anemia among T2DM regarding gender in contrast to other studies [12, 46]. However, Bekele et al. showed a higher prevalence of anemia among men . As the mean age of our participants was 57.8 years, men and women were similar due to decreased occurrence of anemia in women due to menopause. In the Iranian diabetic society, men and women have identical diets, education, and health awareness [47, 48].
AlDallal and Idris et al. [6, 19] showed a positive association of age and prevalence of anemia, which is in contrast with our results that could be due to bone marrow disorders occurring in older age, deficiencies in vitamins such as cyanocobalamin and folate, and a higher number of comorbidities .
The strengths of the research were: first, we assess various baseline characteristics such as various laboratory and clinical data in the study. We considered different potential associated factors in our model in a sufficient sample size. Thus, we could discuss some important independent associated factors in detail. Second, we used a random sampling method that could mitigate selection bias and enhance the ability to generalize our results to the Iranian population (due to different ethnic groups living in Gorgan, Indigenous inhabitants, Turkman, Sistani, and Baloch). Third, to the best of our knowledge, there were no studies concerning the prevalence of anemia among diabetic population in the north of Iran. This research had several limitations. First, as it was a cross-sectional study, we could not show the casualty, so a longitudinal study is needed to assess the relationship over time. Second, we did not consider some drug usage including iron supplements, erythropoietin (EPO) in case of DKD. Third, we did not record dietary patterns, particularly iron intake was not considered in our study, although all patients are routinely educated about diet and health care in the Gorgan diabetes clinic. Fourth, we did not measure erythropoietin, B12, and folate levels in our participants, so lack of definition of anemia etiology was another limitation.