Reliability of transcutaneous hemoglobin measurements in an outpatient pediatric clinic

Objective: Our study assessed the reliability of a transcutaneous hemoglobin (tcHgb) measurement as compared to a standard capillary hemoglobin (cHgb) measurement in screening for iron deficiency anemia in a single university-based pediatric outpatient clinic. Methods: Study participants included all pediatric patients requiring a hemoglobin (Hgb) assessment from July 2019 to June 2020. A tcHgb measurement was attempted on all children who received a cHgb measurement. Additional variables evaluated were age, visit type, gender, insurance type, weight, BMI percentile and presence of comorbid conditions. Results: Of 777 attempts, both cHgb and tcHgb were obtained in 196 children aged 9 months to 21 years. Attempts were most successful in children > 2 years of age due to finger size and ability to remain still for one to two minutes. The mean cHgb was 12.5 ± 1.5 g/dL, mean tcHgb value 13.1 ± 2.1 g/dL, and the mean difference was 0.6 ± 2.1 g/dL (tcHgb-cHgb). An intraclass correlation coefficient was 0.29. There were no differences with regards to age, visit type, gender, insurance type, weight, BMI percentile and presence of comorbid conditions. Bland-Altman analysis displayed a lack of agreement between the tcHgb and cHgb measurements and tcHgb tended to over-estimate Hgb values when the cHgb was low. Conclusions: The measurement of a transcutaneous Hgb is less invasive for pediatric patients but has significant limitations. Smaller children (probe specifications) and movement limited the ability to obtain a tcHgb measurement in ~75% of children tested. Falsely normal tcHgb values occurred due to overestimation of hemoglobin when compared to the traditional cHgb device. The suboptimal sensitivity of the tcHgb device may cause providers to miss a diagnosis of anemia. Future research should compare both methods of Hgb assessment to the gold standard laboratory-analyzed complete blood count and use a smaller probe for children under 2, once available.