AbstractIn principle, an aseptic package should protect against any type of microbe penetration. It means that each microhole must be less than about 0.5μm in diameter. In practice this critical limit may be much higher, owing to various factors, and consequently a controlling system does not necessarily have to have such a sensitivity.Controlling systems based on gas leakage detection have been suggested. In this report the basic validity range of such systems, including either mass spectrometer detection or an infra‐red spectrometer, is analysed and various crucial factors are clarified as a basis for technical development.The actual control principle assumes that transmission of gases through the package originates only from microholes and cracks and that gas permeation through the packaging material is negligible.However, in practice the gas permeation may be considerable. For example, a package having a polyethylene (LDPE) barrier layer of 50μm in thickness has a gas permeation comparable to what is coming out through a pinhole of 10μm in diameter, and accordingly only much bigger microholes can be determined, with reference to calculations presented. If the barrier is improved by a Saran® layer (PVDC) of 25μm the gas permeation is still disturbing and only holes bigger than about 5μm in diameter may be detected. If instead the barrier is improved by an aluminium layer, gas permeation is negligible in most cases, and the basic principle of the gas leak detection systems is valid.Infra‐red spectrometer based systems theoretically need at least a test time of 0.5 min and in practice much more to detect a microhole of 10μm in diameter‐to collect a sufficient quantity of trace gas. Mass spectrometer based systems are in principle not testing time limited assuming a rapid enough vacuum pump can be arranged and that moisture interferences do not arise from the packaging material or from the packaged contents.Further, some aspects on the necessity of introduction a 100% control system,vis‐a‐ visa statistical control system, are given in terms of producer's and consumer's risks. About 100 packagings/h must be tested at ordinary production speed (1000–3000 packagingslh−1) to check accepted quality levels of 0.1% defectives. At higher quality levels, i.e. less than 0.1% defectives, the sample size needed increases rapidly towards 100% of