Construction and use of a library of bona fide heparins employing 1H NMR and multivariate analysis.

In contrast to most pharmaceutical agents, the major anticoagulant agent, heparin, lacks a uniquely defined chemical structure. It is an inherently structurally varying, poly-disperse polymer, rendering quality control problematic. Structural modifications introduced during manufacture and the presence of possible contaminants are dangers. The comparison of any heparin test sample to a library of bona fide, but intrinsically variable heparins, is limited fundamentally to measuring the degree of similarity between them. A general approach for constructing a suitable illustrative heparin library employing (1)H NMR and multivariate analysis, is proposed and examples shown. Heparin samples contaminated with non-N-acetylated, persulfated, non-glycosaminoglycan carbohydrates (10%) were detected, some of which would not be easily detected under current regulatory guidelines. The ability to identify contaminated heparin is fundamentally dependent on the contents of the library, which should contain all the 'normal' variability within 'heparin'. Oversulfated chondroitin sulfate can be detected (∼5%) (using components 1 and 2, but is detectable at 3% using additional components, e.g. 1 and 5) analysing the full spectrum and at <1% if only the N-acetyl region is used. Signals arising from oxidation during the manufacturing processes can also be found. The design, properties and limitations of this approach are discussed. Assembly of such a library of market heparin can be termed the 'Current Statistical Definition of Heparin' (CSDH) and will form the basis of future quantitative methods.