Four unembalmed human cadavers were used in eight direct-forearm-airbag-interaction static deployments to assess the relative aggressivity of two different airbag modules. Instrumentation of the forearm bones included triaxial accelerometry, crack detection gages, and film targets. The forearm-fracture predictors, peak and average distal forearm speed (PDFS and ADFS), were evaluated and compared to the incidence of transverse, oblique, and wedge fractures of the radius and ulna. Internal-airbag pressure and axial column loads were also measured.The results of this study support the use of PDFS or ADFS for the prediction of airbag-induced upper-extremity fractures. The results also suggest that there is no direct relationship between internal-airbag pressure and forearm fracture. The less-aggressive system (LAS) examined in this study produced half the number of forearm fracture as the more-aggressive system (MAS), yet exhibited a more aggressive internal-pressure performance. Both the peak-internal pressure and the initial-inflation rate of the LAS were higher than for the MAS, but the PDFS, ADFS, and axial column loads of the LAS were lower. This inverse relationship between internal-airbag pressure and airbag aggressivity prompted an investigation of the LAS and MAS design characteristics.The closed-module design of the LAS, coupled with longer, thicker tear seams, results in higher peak-internal pressures and greater rates of pressure increase when compared to the MAS. Therefore, more inflator energy must be used to achieve bag egress from the LAS module, making less energy available to be imparted to a forearm. The reduced and more distributed mass and size of the LAS doors may assist in the reduction of focused energy transfer to a forearm and the less-aggressive-tank-test characteristics compared to the MAS inflator. A combination of these factors causes a reduction in the level of fracture predictors, such as PDFS and ADFS, when using the LAS, and a reduction in the incidence and severity of forearm fractures.