A good start is HPLC-UV (Saudi Pharmaceutical Journal, 2014, 22, 483-487).

we use almost the same methodology but the problem is hidden in tablet composition. We think that HPMC 100 000 doesn't allow metformin to get into the solution...

Dear Sir. Concerning your issue about the determination of metformin hydrochloride in sustained-release tablets . Adeveloped and validated a suitable method for the assay of metformin hydrochloride (HCl) in tablets containing croscarmellose sodium as an additive was carried out. Methanol and ethanol (99%) were assessed as solvents for sample preparation for the assay of metformin HCl in tablets containing croscarmellose sodium by high performance liquid chromatography (HPLC) and ultra violet spectrophotometric (UV) methods. The proposed method was subjected to validation tests. The recovery of metformin HCl from the placebo-spiked sample was 95.1 to 96.9 % as per BP and USP methods compared with 99.3 to 100.8 % when analyzed by the proposed method. The use of methanol and ethanol as solvents resolved the problem of retention of metformin HCl by croscarmellose sodium in solution during the preparation of sample solution. The modified UV and HPLC methods are suitable for the determination of metformin HCl in tablets both in the presence and absence of croscarmellose sodium. The method is specific, precise, accurate, robust, rugged and gives a linear response for the quantitative estimation of metformin HCl in tablet formulation. I think the following below links may help you in your analysis:

http://www.bioline.org.br/pdf?pr12014

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3465141/

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.407.7134&rep=rep1&type=pdf

Thanks

Andrei Blasko and Isam Eldin Hussein Elgailani, thank you very much for your comments. Increasing organic part in solvent and duration of dissolving process gave us acceptable results.