Shaking or stirring in a miniaturized device is often applied in the development of protein pharmaceuticals, serving as a test of stability under physical stress conditions also occurring in the real process. The overall purpose of these "stress tests" is to accelerate protein aggregation that could otherwise take place at a much slower rate, thereby enhancing experimental throughput to speed up determination of critical process parameters of stability. Results are useful to determine critical parameters for formulation development. However, shaking differs from stirring in the forces applied on proteins in solution and therefore, there is the possibility that characteristics of protein stability interrogated with each method may not be the same. We have performed a detailed, time-resolved analysis of aggregation of the human growth hormone (hGH), exposed to stirring and shaking in a well-defined and comparable mini-reactor set-up. We show that aggregation of hGH, which in its native form is a monomer, is not the result of protein deamidation, a chemical protein modification that occurs in hGH under the incubation conditions used. We also show that both under stirring and shaking, hGH appears to precipitate into mainly spherical particles in the presence of only small amounts of soluble protein oligomers. The precipitate appears to contain protein that has a substantial amount of hydrophobic surface exposed. An air liquid interface, constantly renewed by stirring or shaking, was identified to be mainly responsible for aggregation of hGH under each of the stress conditions applied. In hGH, therefore, controlled shaking and stirring lead to identical conclusions about the overall protein stability and the main critical parameter driving the aggregation, thus informing stabilization strategies and formulation development. (c) 2012 Elsevier Ltd. All rights reserved.