•Changes of building airtightness of 61 houses are evaluated on different time scales.•Airtightness of houses may deteriorate mainly during the first two years.•2-Storey houses with wood frame roofs present a higher risk of airtightness deterioration.•The increase in the number of leakages is not correlated with changes in the airtightness.•The airtightness of the houses tested is not affected by seasonal variations. The main objectives of this paper are to quantify and qualify changes in building airtightness on different time scales, and to identify factors that may explain the variations observed. A special measurement protocol, based on ISO 9972 but with some additional requirements, was defined after a detailed literature review. For the mid-term campaign, 30 new detached houses were measured once per year over a 3-year period. It was observed that air permeability increases slightly during the first year (a mean increase of 18%), and then stabilizes during the second and third years. Five of them were measured twice per year but did not show any significant seasonal variation. For the long-term campaign, 31 detached houses built during the last 10 years were remeasured once. Air permeability shows a similar increase to that of the mid-term campaign after 3–10 years (a mean increase of 20%). The results show an overall increase in the number of leakages detected for all houses, but this increase is not always correlated with the change in air permeability. Results pointed to three factors that could explain the deterioration of airtightness: the number of levels, the type of roof and the type of building material and air-barrier.