Abstract
Reference ranges (RRs) for haematinics need to be independently verified by individual laboratories since it has been shown that those quoted by some manufacturers may be inappropriate. The measurement of serum folate, ferritin and B12 remains a first-line investigation in the assessment of several pathologies. Establishing clearly defined, accurate reference ranges facilitates good interpretation and effective discrimination between health and disease thereby avoiding expensive and needless follow-up. Early morning venous samples were collected into Greiner Vacuette serum tubes (Ref: 456018) from 221 healthy laboratory personnel (F= 159;M = 62) aged 20–63 years for both gender. Age groups were equally represented. Serum vitamin B12, folate and ferritin were measured on all samples on a Beckman-Coulter Access analyser on the same day of collection. NCCLS guidelines (C28-A and H3-A4) were followed throughout. Outliers were excluded and data examined for normal distribution. The following normality checks were applied - Kurtosis, Kolmogorov-Smirnov and Shapiro-Wilk tests of normality, T-values (Skewness/SE Skewness), %diff from mean-median, histogram and normal curve, normal Q-Q probability plot and Box plot. Because all three parameters showed non-normal distribution, RRs were calculated using (1) the 2.5-and 97.5- percentiles, (2) the 2.5- and 97.5- percentiles on the transformed scale. The transformations, log natural, log10 and square root were applied to the variables and tested for normality. The transformation giving the best normal distribution was then selected. The non-parametric Mann-Whitney test was used to examine significant differences between males and females. Significant differences (p values shown below) are indicated by an asterisk.
| . | New Limits . | Historical Limits . | Manufacturers quoted RR . | Mann-Whitney U test (p value) . | |
|---|---|---|---|---|---|
| (a) = Log natural; (b) = Log 10; (c) = square root | |||||
| RR (1) | RR (2) | ||||
| Folate ng/mL | M 2.1 – 14.7 | 2.1 (a) – 14.6 | 2.7–14.0 | >3.0 | 0.016* |
| F 2.7 – 18.1 | 2.7 (b) – 18.1 | 2.7–14.0 | >3.0 | ||
| Ferritin ng/mL | M 11 – 215 | 11 (c) – 215 | 20.0–350 | 24–336 | <0.05* |
| F 5 – 119 | 5 (b) – 119 | 10.0–300 | 11–307 | ||
| Vit B12 pg/mL | M 113 – 567 | 113 (c) – 567 | 180–900 | 180–914 | 0.933 |
| F 136 – 600 | 136 (b) – 600 | 180–900 | 180–914 | ||
| . | New Limits . | Historical Limits . | Manufacturers quoted RR . | Mann-Whitney U test (p value) . | |
|---|---|---|---|---|---|
| (a) = Log natural; (b) = Log 10; (c) = square root | |||||
| RR (1) | RR (2) | ||||
| Folate ng/mL | M 2.1 – 14.7 | 2.1 (a) – 14.6 | 2.7–14.0 | >3.0 | 0.016* |
| F 2.7 – 18.1 | 2.7 (b) – 18.1 | 2.7–14.0 | >3.0 | ||
| Ferritin ng/mL | M 11 – 215 | 11 (c) – 215 | 20.0–350 | 24–336 | <0.05* |
| F 5 – 119 | 5 (b) – 119 | 10.0–300 | 11–307 | ||
| Vit B12 pg/mL | M 113 – 567 | 113 (c) – 567 | 180–900 | 180–914 | 0.933 |
| F 136 – 600 | 136 (b) – 600 | 180–900 | 180–914 | ||
It is vital that investigators use method-specific RRs in their own laboratories since those quoted by some manufacturers are inappropriate. Although our lower limits for serum ferritin appear to be low, they are in keeping with previously published data. A proportion of the normal population have low serum ferritin but are not anaemic nor symptomatic. This confusion between normality and iron deficiency continues to cause difficulties in interpretation. The range of vitamin B12 and folate concentrations in some healthy individuals overlap with those in symptomatic patients. It is useful therefore to quote “indeterminate” ranges. Our data indicates that such a range for vitamin B12 should be 130 – 160 since our distribution histogram (not shown) shows a clearly defined “flattened shoulder” between these values.
Author notes
Corresponding author
