The unique magnetic properties of superparamagnetic iron oxide nanoparticles (SPIONs) have led to their increasing use in drug delivery and imaging applications. Some polymer-coated SPIONs, however, share with many other nanoparticles the potential of causing hypersensitivity reactions (HSRs) known as complement (C) activation-related pseudoallergy (CARPA). In order to explore the roles of iron core composition and particle surface coating in SPION-induced CARPA, we measured C activation by 6 different SPIONs in a human serum that is known to react to nanoparticles (NPs) with strong C activation. Remarkably, only the carboxymethyldextran-coated (ferucarbotran, Resosvist®) and dextran-coated (ferumoxtran-10, Sinerem®) SPIONs caused significant C activation, while the citric acid, phosphatidylcholine, starch and chitosan-coated SPIONs had no such effect. Focusing on Resovist and Sinerem, we found Sinerem to be a stronger activator of C than Resovist, although the individual variation in 15 different human sera was substantial. Further analysis of C activation by Sinerem indicated biphasic dose dependence and significant production of C split product Bb but not C4d, attesting to alternative pathway C activation only at low doses. Consistent with the strong C activation by Sinerem and previous reports of HSRs in man, injection of Sinerem in a pig led to dose-dependent CARPA, while Resovist was reaction-free. Using nanoparticle tracking analysis, it was further determined that Sinerem, more than Resovist, displayed multimodal size distribution and significant fraction of aggregates - factors which are known to promote C activation and CARPA. Taken together, our findings offer physicochemical insight into how key compositional factors and nanoparticle size distribution affect SPION-induced CARPA, a knowledge that could lead to the development of SPIONs with improved safety profiles.