BACKGROUND: Dried blood may be a convenient method of sample collection in epidemiological studies; however, the method needs evaluation in a field settings. In the present study, feasibility of using dried blood for measurement of cholesterol and triglycerides was evaluated in multicenter surveillance study for noncommunicable disease (NCD). METHODS: Samples were collected in a cross-sectional study for NCD risk factor surveillance conducted in six centers in India. For every tenth subject recruited, a blood sample was also collected on filter paper. These 10% serum samples and dried blood spots were analyzed for cholesterol and triglycerides. RESULTS: The mean coefficient of variation (CV) for cholesterol was less than 10% between dried blood and serum in five of the six participating centers. Only one center showed a high CV of 14%. Similarly, the mean bias was less than 10% in five centers. The intraclass correlation between cholesterol values in dried blood and serum were greater than 0.638 in all centers, which suggests a good homogeneity of results. The mean CV for triglycerides ranged from 0.36% to 17.97%. The intraclass correlation between triglyceride values in dried blood and serum ranged from 0.756 to 0.880 in the six centers. CONCLUSION: In conclusion, dried blood would be a good method for collection of blood for measurement of cholesterol and triglycerides for population health surveys. However, the benefits of blood spot analysis should be weighed against potential sources of errors attributable to sampling and other factors, such as temperature and humidity, in a country like India.
BACKGROUND: Dried blood may be a convenient method of sample collection in epidemiological studies; however, the method needs evaluation in a field settings. In the present study, feasibility of using dried blood for measurement of cholesterol and triglycerides was evaluated in multicenter surveillance study for noncommunicable disease (NCD). METHODS: Samples were collected in a cross-sectional study for NCD risk factor surveillance conducted in six centers in India. For every tenth subject recruited, a blood sample was also collected on filter paper. These 10% serum samples and dried blood spots were analyzed for cholesterol and triglycerides. RESULTS: The mean coefficient of variation (CV) for cholesterol was less than 10% between dried blood and serum in five of the six participating centers. Only one center showed a high CV of 14%. Similarly, the mean bias was less than 10% in five centers. The intraclass correlation between cholesterol values in dried blood and serum were greater than 0.638 in all centers, which suggests a good homogeneity of results. The mean CV for triglycerides ranged from 0.36% to 17.97%. The intraclass correlation between triglyceride values in dried blood and serum ranged from 0.756 to 0.880 in the six centers. CONCLUSION: In conclusion, dried blood would be a good method for collection of blood for measurement of cholesterol and triglycerides for population health surveys. However, the benefits of blood spot analysis should be weighed against potential sources of errors attributable to sampling and other factors, such as temperature and humidity, in a country like India.
Authors: Parisa Tehranifar; Angeline Protacio; Karen M Schmitt; Elise Desperito; Sabine Oskar; Alan J Potter; Natalie J Engmann; Mary Beth Terry Journal: Cancer Causes Control Date: 2015-07-14 Impact factor: 2.506