Characterization of currently marketed heparin products: composition analysis by 2D-NMR

Each type of heparin has a unique pattern of signals in their NMR spectra because of the effects of the process specific chemical reactions used to manufacture them. Thus, for heparin sodium, 1D-H-1-NMR identification tests are part of the United States Pharmacopeia (USP) and European Pharmacopeia (EP) monographs. Previous work has shown that two-dimensional (2D) NMR data can be used to identify heparin types and be used to calculate heparin composition. Therefore, in this study, we applied a 2D H-1-C-13-heteronuclear single quantum coherence (HSQC) spectroscopy experiment to characterize the normal variation of intact heparin sodium or low molecular weight heparin (LMWH) molecular composition in USP grade sample lots which were available on the US market in the summer of 2009. We tested 7 heparin sodium active pharmaceutical ingredient (API) samples by the 2D method representing lots from 6 manufacturers and the USP heparin sodium identification standard. In addition, we tested 10 LMWH API samples from 3 manufactures of different types of LMWH and the USP enoxaparin identification standard. Using the integrated volumes of HSQC cross-peaks assigned to specific heparin species the monosaccharide or disaccharide percent compositions were calculated. These data establish the composition and normal range of variability for each of the heparin types in the 2D assay across manufacturers supplying the US market. The values obtained from the NMR test were similar to those obtained from mass spectrometric analysis of heparin digests on the same sample set. The robustness of the assay was tested by varying the acquisition time from similar to 3 to similar to 68 h by changing the number of transients co-added or the relaxation delay; across these changes the percent composition values obtained did not vary significantly. We conclude that the manner in which the calculations are performed minimizes experimental errors that arise from differences in spectral signal-to-noise, heparin (1)J(CH) through bond coupling constants or relaxation times.