The paper analyses experimental data exhibiting heat transfer deterioration phenomena at supercritical pressure resulting in quite different trends of wall temperature. The two modes of deteriorated heat transfer considered here are both due to the laminarisation of flow, owing to lighter fluid appearing at the wall as a consequence of heating. However, in one of the two modes, mainly occurring in the liquid-like region at sufficiently large bulk subcooling with respect to the pseudocritical threshold, after a local deterioration heat transfer is restored almost at normal levels, while in the other, at lower pseudo-subcooling and involving at a larger degree also the gas-like region, a jump of wall temperature to a stable deteriorated condition occurs. In the latter case, the wall temperature finally decreases only when the bulk fluid approaches the pseudocritical temperature, thus causing a smaller density difference between bulk and wall, whose increase was at the root of buoyancy effects leading to deterioration.The two heat transfer deterioration modes are here investigated considering the measured trends of wall temperature in two relevant data sets exhibiting the two behaviours and by the use of a CFD model for predicting detailed data. It is shown that while the first mode of deterioration can be interpreted as an entry-length problem, related to strong buoyancy effects, the second mode is complicated by an excursion to a different manifold of what can be defined as a pseudo-boiling curve, occurring when the imposed heat flux cannot be sustained by normal or enhanced heat transfer. In the latter case, an interesting analogy can be considered with phenomena related to the boiling crisis occurring at subcritical pressures, though the causes of such similar behaviours are obviously quite different. The CFD model, validated by comparison with experimental data of wall temperature, is used as a tool to generalise the observed trends and to draw conclusions.The work is carried out in the frame of studies for assessing and possibly improving heat transfer correlations for supercritical pressure fluids, for the use in the future licensing of supercritical water reactors (SCWRs), in the aim to gain better understanding and to suggest possible modelling strategies. Conclusions are proposed in this regard.