TABLE 2.
Ferritin and hemoglobin concentrations in IPD datasets including nonpregnant individuals1.
| Charles et al. 2011 [19] (n = 35) | Charles et al. 2015 [20] (n = 173) | Fischer et al. 2023 [21] (n = 480) | Gallant et al. 2021 [22] (n = 334) | Karakochuk et al. 2015 [12] (n = 420) | Karakochuk et al. 2017 [23] (n = 807) | Makurat et al. 2016 [24] (n = 217) | Rappaport et al. 2017 [26] (n = 326) | Whitfield et al. 2017 [27] (n = 175) | Wieringa et al. 2016 [28] (n = 445) | |
|---|---|---|---|---|---|---|---|---|---|---|
| Unadjusted ferritin (μg/L) | 84.0 [69.0, 102.0] | 39.9 [25.5, 68.6] | 80.7 [43.2, 116.8] | 59.7 [36.0, 93.1] | 90.8 [59.2, 142.3] | 40.0 [17.8, 83.2] | 34.5 [17.6, 62.1] | 62.5 [36.7, 97.7] | 73.4 [49.4, 107.4] | 87.6 [56.6, 136.1] |
| ID, <15 μg/L | 0 (0) | 26 (15) | 22 (5) | 24 (7) | 8 (2) | 169 (21) | 47 (22) | 30 (9) | 2 (1) | 10 (2) |
| Inflammation-adjusted ferritin (μg/L)2 | —3 | —3 | 62.1 [35.0, 93.6] | 46.9 [30.2, 73.4] | 74.7 [45.7, 110.3] | 30.8 [13.7, 57.6] | 29.7 [16.3, 56.8] | 59.6 [33.5, 87.0] | 64.1 [43.8, 87.9] | 55.5 [37.9, 75.9] |
| ID, <15 μg/L | — | — | 33 (7) | 32 (10) | 8 (2) | 221 (27) | 52 (24) | 34 (10) | 2 (1) | 21 (5) |
| Probable ID, <30 μg/L | — | — | 95 (20) | 82 (25) | 39 (9) | 396 (49) | 109 (50) | 68 (21) | 22 (13) | 69 (16) |
| Hemoglobin, measured with an automated analyzer (g/L) | 128.7 ± 14.04 | — | 129.0 ± 11.84 | — | 124.5 ± 10.94 | 115.9 ± 13.34 | — | 117.7 ± 11.14 | 124.3 ± 10.94 | — |
| Anemia, <120 g/L | 8 (23)4 | — | 80 (17)4 | — | 124 (30)4 | 468 (58)4 | — | 164 (50)4 | 52 (30%)4 | — |
| Hemoglobin, measured with a HemoCue (g/L) | — | 121.3 ± 13.54 | — | 118.0 ± 13.44 | 122.0 ± 13.35 | 107.2 ± 9.45 | 124.6 ± 10.64 | 109.1 ± 8.55 | — | 120.5 ± 13.05 |
| Anemia, <120 g/L | — | 76 (44)4 | — | 179 (54)4 | 172 (41)5 | 807 (100)5 | 59 (27)4 | 325 (100)5 | — | 187 (42)5 |
Abbreviations: ID, iron deficiency; IPD, individual participant data.
1
Total n = 3412 individuals from 10 datasets; n = 3377 with eligible data for inclusion in IPD meta-analysis. Values are median [IQR], mean ± SD, or n (%).
2
Ferritin concentrations were adjusted for inflammation based on α-1 acid glycoprotein and C-reactive protein values, as per Biomarkers Reflecting Inflammation and Nutritional Determinants of Anemia (BRINDA) guidelines [11].
3
Ferritin concentrations were not adjusted due to a lack of correlation between ferritin and C-reactive protein values, as per BRINDA guidelines [11].
4
Blood source was venous.
5
Blood source was capillary.