Skip to main content

Table 3 Association between body iron stores and type 2 diabetes mellitus (T2DM) in the included studies

From: Dietary iron intake, body iron stores, and the risk of type 2 diabetes: a systematic review and meta-analysis

Source Gender Comparison Models RR (95% CI) Matched or adjusted covariates
Ferritin as indicators of body iron stores (n = 5)   
Jiang et al, 2004, USA [10] Women Highest (≥107.2 ng/ml) versus lowest (<21.1 ng/ml) quintile Model 1a 2.68 (1.75 to 4.11) Age, ethnicity, fasting status, BMI, FH, PA, smoking, alcohol consumption, menopausal status, glycemic load, intake of total energy, cereal fiber, magnesium, and trans fat, and ratio of polyunsaturated fat to saturated fat
    Model 2b 2.61 (1.68 to 4.07) Additional adjustment for CRP
Forouhi et al, 2007, UKc [11] Men Highest (≥135.7 ng/ml) versus lowest (<34.4 ng/ml) quintile Model 1a 1.97 (1.12 to 3.45) Age, sex, BMI, FH, PA, smoking, dietary factors (TEI, alcohol consumption, intake of dietary iron, magnesium, and red meat and processed meat, plasma vitamin C)
    Model 2b 1.78 (0.99 to 3.19) Additional adjustment for CRP, fibrinogen, and IL-6
    Model 3 1.13 (0.58, 2.19) Additional adjustment for ALT, GGT, and adiponectin
  Women Highest (≥71.7 ng/ml) versus lowest (<17.8 ng/ml) quintile Model 1a 2.55 (1.22 to 5.34) Age, sex, BMI, FH, PA, smoking, dietary factors (TEI, alcohol consumption, intake of dietary iron, magnesium, red meat and processed meat and plasma vitamin C)
    Model 2b 2.11 (0.98 to 4.56) Additional adjustment for CRP, fibrinogen, and IL-6
    Model 3 1.08 (0.44, 2.62) Additional adjustment for ALT, GGT, and adiponectin
Jehn et al, 2007, USA [15] Both Highest (≥235.4 ng/ml; median, 354.5 ng/ml) versus lowest (<40.0 ng/ml; median, 20.0 ng/ml)quintile Model 1a 1.51 (0.98 to 2.31) Age, study center, ethnicity, smoking, alcohol consumption, and BMI
    Model 2 0.81 (0.49 to 1.34) Additional adjustment for metabolic syndrome components (HDL-C, WC, hypertension, FPG, and TG)
    Model 3b 0.79 (0.48 to 1.32) Additional adjustment for FPI and inflammation score
Le et al, 2009, USA [16] Men Highest (>188 ng/ml) versus lowest (<80 ng/ml) quartile Model 1a 1.79 (1.13 to 2.82) Age, ethnicity, and BMI
  Women Highest (premenopausal, >60 ng/ml, postmenopausal, >90 ng/ml) versus lowest (premenopausal,<21 ng/ml, postmenopausal,<37 ng/ml) quartile Model 1a 0.87 (0.37 to 2.03) Age, ethnicity, and BMI
Rajpathak et al, 2009, USA [17] Both Highest (median, 203.7 ng/ml) versus lowest (median, 20.1 ng/ ml) quartile Model 1a 1.02 (0.60 to 1.74) Age, sex, ethnicity, and BMI
    Model 2 1.65 (0.90 to 3.02) Additional adjustment for FH, PA, HbA1c, and sTfR
    Model 3b 1.53 (0.83 to 2.82) Additional adjustment for CRP
    Model 4 1.61 (0.85 to 3.02) Additional adjustment for HOMA-IR
Ratio of sTfR to ferritin as indicators of body iron stores (n = 2)   
Salonen et al, 1998, Finland [9] Men Highest (< 9.4) versus quartile (no data available) Model 1 2.40 (1.03 to 5.50) Age, time of examination, place of residence, cigarette smoking, exercise, maximal oxygen uptake, socioeconomic status, height, weight, hip and waist circumferences, glucose, insulin, vitamin E, and serum SFA to (PUFA + MUFA) ratio
Jiang et al, 2004, USA [10] Women Highest (<26.7) versus lowest (≥149.4) quintile Model 1 2.44 (1.61 to 3.71) Age, ethnicity, fasting status, BMI, FH, PA, smoking, alcohol consumption, menopausal status, glycemic load, intake of total energy, cereal fiber, magnesium, and trans fat, ratio of polyunsaturated fat to saturated fat
    Model 2 2.40 (1.55 to 3.71) Additional adjustment for CRP
sTfR as indicators of body iron stores (n = 1)   
Rajpathak et al, 2009, USA [17] Both Highest (median, 4.4 mg/l) versus lowest (median, 2.3 mg/l) quartile Model 1 1.55 (0.93 to 2.57) Age, sex, ethnicity, and BMI
    Model 2 2.26 (1.27 to 4.01) Additional adjustment for FH, PA, HbA1c, and sTfR
    Model 3 2.39 (1.34 to 4.28) Additional adjustment for CRP
    Model 4 2.23 (1.22 to 4.06) Additional adjustment for HOMA-IR
  1. Abbreviations: ALT, alanine aminotransferase; BMI, body mass index; CRP, C-reactive protein; FH, family history; FPG, fasting plasma glucose; FPI, fasting plasma insulin; GGT, γ-glutamyltransferase; HbA1c, glycated hemoglobin; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment for insulin resistance; IL-6, interleukin-6; MUFA, monounsaturated fatty acids; PA, physical activity; PUFA, polyunsaturated fatty acids; RR, relative risk; SFA, saturated fatty acids; sTfR, soluble transferrin receptor; TEI, total energy intake; TG, triglycerides; WC, waist circumference.
  2. a Estimates used in the meta-analysis of ferritin and T2DM risk in multivariate-adjusted models.
  3. b Estimates used in the meta-analysis of ferritin and T2DM risk in multivariate-adjusted models including inflammatory markers.
  4. c From an ad hoc analysis by Forouhi et al. [11], which used quintiles of ferritin levels as exposure and separately reported for men and women.