Validation, transfer and measurement uncertainty estimation of an HPLC-UV method for the quantification of artemisinin in hydro alcoholic extracts of Artemisia annua L.

Malaria is the world’s most important parasitic infection with 500 millions cases annually and almost 2 millions death per year. This disease is more present in Sub-Saharan Africa where 90% of the infections are found. Artemisinin and its semi synthetic derivatives (artemether, artesunate) have actually the most powerful activity on malaria, even in its complicated forms and resistance cases.
Various methods have been proposed for detection and quantification of artemisinin in Artemisia annua L. by HPLC-UV, but the plant extracts used for this quantification were extracts obtained with organic solvents (toluene, petroleum ether, hexane …). To be able to use crude A. annua extracts prepared at low cost to formulate antipaludic drugs, we chose the use of a mixture of water and ethanol as solvent of extraction, but no adequate analytical method for this kind of extracts is published.
The main objectives of this work were first to develop an analytical method for artemisinin quantification in hydro alcoholic extracts of A. annua. Second, this method had to be thoroughly validated by the research and development laboratory and, third, the transfer of this method to the routine laboratory had to be demonstrated. The final aim was to compare the estimation of measurement uncertainty obtained during the method validation with validation standards to measurement uncertainty estimates obtained during the method transfer study with real samples.
The method was validated following the accuracy profile methodology and was found to be accurate in the concentration range of 10.0 to 54.0 µg/ml with CV <8%. Limit of detection and of quantification were 2.73 and 10.0 µg/ml, respectively. The method was then successfully transferred to a laboratory in Benin by showing that the quality of the results that it will generate during routine application of the method is sufficient. Finally, the measurement uncertainty of the method was estimated from the validation experiments as well as from the transfer study with authentic unspiked samples of A. annua. The comparison of these measurement uncertainty estimations showed that they were coherent. It confirmed thus that the estimation of measurement uncertainty from validation experiments predicts well the measurement uncertainty of real routine samples. This analytical method was thus shown to be convenient for routine analysis of hydro alcoholic extracts of A. annua in Benin.