Low haemoglobin density for detecting iron deficiency across a large population, including pregnancy.

Introduction: Low haemoglobin density (LHD%) from Coulter counters has been suggested as a means to detect iron deficiency. Its performance in a broad population group, including pregnancy, has not been evaluated.

Methods: A retrospective study of adult and paediatric (under 12 years old) patient samples referred for blood counts and iron studies between October 2013 and March 2015. Receiver operator characteristic (ROC) curves were constructed to evaluate the performance of LHD% adults, children, and in the antenatal subgroup.

Results: Using a strict definition for iron deficiency, compared with a selected normal cohort, LHD% had a ROC area under the curve (AUC) of 0.90 (0.89-0.91), but in an unselected cohort, the AUC fell to 0.74 (0.73-0.75) with a sensitivity of 74% and specificity of 60% at a cut-off value of 5.9%. In the paediatric cohort, the AUC was 0.79(0.73-0.85), giving a sensitivity and specificity of 75% and 68%, respectively. LHD% did not effectively identify iron deficiency in pregnancy with an AUC of 0.60 (0.54-0.65) and was no better than MCV at detecting iron deficiency.

Conclusion: LHD% detects iron deficiency in adult and paediatric populations, but not in the antenatal setting, and does not appear superior to MCV.