Context. We investigated the plasma and magnetic field characteristics of the upstream regions of interplanetary coronal mass ejections (ICMEs) and their evolution as function of distance to the Sun in the inner heliosphere. Results are related both to the development of interplanetary shocks, sheath regions, and compressed solar wind plasma ahead of the magnetic ejecta (ME). Aims. From a sample of 45 ICMEs observed by Helios 1/2 and the Parker Solar Probe, we aim to identify four main density structures; namely shock, sheath, leading edge, and ME itself. We compared characteristic parameters (proton particle density, plasma-beta, temperature, magnetic field strength, proton bulk speed, and duration) to the upstream solar wind in order to investigate the interrelation between the different density structures. Methods. For the statistical investigation, we used plasma and magnetic field measurements from 40 well-observed Helios 1/2 events from 1974–1981. Helios data cover the distance range from 0.3–1 au. For comparison, we added a sample of five ICMEs observed with the Parker Solar Probe from 2019–2021 over the distance range of 0.32–0.75 au. Results. It is found that the sheath structure consists of compressed plasma as a consequence of the turbulent solar wind material following the shock and lies ahead of a region of compressed ambient solar wind. The region of compressed solar wind plasma is typically found directly in front of the magnetic driver and seems to match the bright leading edge commonly observed in remote sensing observations of CMEs. From the statistically derived density evolution over distance, we find the CME sheath becomes denser than the ambient solar wind at about 0.06 au. From 0.09–0.28 au, the sheath structure density starts to dominate over the density within the ME. The ME density seems to fall below the ambient solar wind density over 0.45–1.18 au. Besides the well-known expansion of the ME, the sheath size shows a weak positive correlation with distance, while the leading edge seems not to expand with distance from the Sun. We further find a moderate anti-correlation between sheath density and local solar wind plasma speed upstream of the ICME shock. An empirical relation is derived connecting the ambient solar wind speed with sheath and leading edge density. We provide constraints to these results in this paper. Conclusions. The average starting distance for actual sheath formation could be as close as 0.06 au. The early strong ME expansion quickly ceases with distance from the Sun and might lead to a dominance in the sheath density between 0.09 and 0.28 au. The leading edge can be understood as a separate structure of compressed ambient solar wind directly ahead of the ME and is likely the bright leading edge of CMEs often seen in coronagraph images. The results allow for better interpretation of ICME evolution and possibly the observed mass increase due to enlargement of the sheath material. The empirical relation between sheath and leading edge density and ambient solar wind speed can be used for more detailed modeling of ICME evolution in the inner heliosphere.