Abstract Introduction and aims Lower limb prosthesis has proven to be an extremely successful intervention for amputees. However, to achieve adequate function, the device places significant friction and load on the skin, which leads to substantial skin breakdown. Current rehabilitation practices require skin to be ‘trained’ to perform a load-bearing role. Despite this, there are very high injury statistics for skin on the residual limb in the literature, we hypothesize that rehabilitation practices to adapt amputee skin are limited. Methods To interrogate this, we used in vivo Raman spectroscopy to evaluate the lower limbs of 13 unilateral lower limb amputees. This enabled us to assess how key constituents of the stratum corneum (SC) such as natural moisturizing factor (NMF), water and lipid content vary with depth. We also analysed sebum content to quantify cytokines present and performed mechanical characterization of the lower limb to investigate stiffness and friction. Results We found that the Raman spectra of the amputated limb show reduced NMF, lipid content and water content at the top of the SC when compared with the intact limb (P < 0.05). In addition, we found that amputee skin contains higher levels of inflammatory cytokines compared with control skin. We also recorded that skin on the residuum is four times softer than the intact limb, while friction is much higher at 0.5 on the residuum skin, compared with 0.3 on the intact limb. Conclusions Here, we show the adaptations of amputee skin in response to rehabilitation to be minimal. Contrary to the dogma that rehabilitation trains skin to bear load, we found that skin barrier function in amputee skin is impaired, as indicated by altered water and lipid profile in addition to higher baseline inflammation.