Abstract
STUDY OBJECTIVE: Methemoglobinemia after pesticide poisoning is associated with a mortality of 12% in Sri Lanka. Treatment is complicated by the lack of laboratory facilities. We aimed to develop and validate a low-cost bedside test for quantitative estimation of clinically significant methemoglobin to be used in settings of limited resources. METHODS: A method to reliably produce blood samples with 10% to 100% methemoglobin was developed. Freshly prepared methemoglobin samples were used to develop the color chart. One drop (10 microL) of prepared methemoglobin sample was placed on white absorbent paper and scanned using a flatbed Cannon Scan LiDE 25 scanner. The mean red, green, and blue values were measured with ImageJ 1.37v. These color values were used to prepare a color chart to be used at the bedside. Interobserver agreement was assessed against prepared samples. The results from clinical use were compared with formal methemoglobin measurements. RESULTS: The red color value was linearly related to percentage methemoglobin (R(2)=0.9938), with no effect of absolute hemoglobin concentration. Mean interobserver (N=21) agreement and weighted kappa for scanned methemoglobin spots using the color chart were 94% and 0.83, respectively. Mean interobserver (N=9) agreement and weighted kappa for a freshly prepared methemoglobin sample with the chart were 88% and 0.71, respectively. Clinical use of the color chart also showed good agreement with spectrometric measurements. CONCLUSION: A color chart can be used to give a clinically useful quantitative estimate of methemoglobinemia.