High-resolution IRAS maps are used to search for the presence of stellar-wind bow-shocks around high-mass X-ray binaries (HMXBs). Their high space velocities, recently confirmed with Hipparcos observations, combined with their strong stellar winds should result in the formation of wind bow-shocks. Except for the already known bow-shock around Vela X-1 (Kaper et al. 1997), we do not find convincing evidence for a bow-shock around any of the other HMXBs. Also in the case of (supposedly single) OB-runaway stars, only a minority appears to be associated with a bow-shock (Van Buren et al. 1995). We investigate why wind bow-shocks are not detected for the majority of these OB-runaway systems: is this due to the IRAS sensitivity, the system's space velocity, the stellar-wind properties, or the height above the galactic plane? It turns out that none of these suggested causes can explain the low detection rate (~40%). We propose that the conditions of the interstellar medium mainly determine whether a wind bow-shock is formed or not. In hot, tenuous media (like inside galactic superbubbles) the sound speed is high (~100 km/s), such that many runaways move at subsonic velocity through a low-density medium, thus preventing the formation of an observable bow-shock. Superbubbles are expected (and observed) around OB associations, where the OB-runaway stars were once born. Turning the argument around, we use the absence (or presence) of wind bow-shocks around OB runaways to probe the physical conditions of the interstellar medium in the solar neighbourhood.