Patch testing may help to assess the culpability of a drug in an adverse reaction. Our aim was to study patch testing in severe cutaneous ad verse drug reactions [ADRs] (Stevens-Johnson syndromeitoxic epidermal necrolysis (SJS/TEN). acute genera exanthematous pustulosis (AGEP), and other cutaneous ADRs). 59 patients with cutaneous ADRs were included: 22 had SJS/TEN. 14 AGEP, and 23 other cutaneous ADRs. Patients were patch tested with the suspect drug and with H standard series of drugs. 2 patients among the 22 SJSTEN cases had a relevant positive test. 7 patients among the 14 AGEP cases had a relevant positive test. 6 patients among the 23 other cutaneous ADRs had a relevant positive test. Our results suggest that patch testing has a weak sensitivity in SJS'TEN and is not appropriate in these diseases. Patch testing seems more adapted to other cutaneous ADRs, such as AC it: P. in which the proportion of positive patch tests was significantly higher (P<0.02). Nevertheless, the difference of sensitivity of patch testing in SJS TEN, AGEP or other cutaneous ADRs could be linked not only to the clinical type of eruption, but also lo the different spectrum of culprit drugs in each type of eruption.