Studies on the epidemiology of common adverse cutaneous drug reactions have rarely been reported, since they can only be successfully conducted in clinics of internal medicine employing consultant dermatologists and having a comprehensive or intensive system of monitoring. Between 1974 and 1993, the adverse skin reactions occurring in divisions of general internal medicine of three different hospitals were monitored by a computerized comprehensive system. The "drug-monitoring patient" was defined as the recipient of at least one drug during hospitalization. The relationship of the skin reactions to drug causality in these patients had to be either definite (proven by re-exposure) or probable (drug relation greater than that of nondrug causality). The skin reactions were classified into four diagnostic groups. Maculopapular exanthema, urticaria, and vasculitis were the three main groups. The fourth group comprised cases of nonhomogeneous but clinically well-defined special exanthema. For selected drugs and years of observation, special emphasis was placed on the study of time patterns (reaction time, exposure time). A total of 1317 definite or probable drug-induced skin reactions occurred during the hospitalization of 48,005 consecutively admitted "drug-monitoring patients": 1201 cases of maculopapular exanthema, 78 cases of urticaria, 18 cases of cutaneous vasculitis, and 20 cases of special exanthema (five of erythema multiforme minor, six of fixed eruption, one of photosensitivity reaction, and eight of acneiform eruption). The main drugs involved did not differ for the three main types of skin reactions, penicillins ranking in the first place, followed by sulfonamides--most often combined with trimethoprim--and in the third place nonsteroidal anti-inflammatory drugs. The reaction time (time from last drug exposure to first skin manifestation) for urticaria showed a relevant proportion of the acute type (within 1 h) and most of the subacute type (1-24 h). For maculopapular exanthema, the subacute or, rarely, the latent type (1-8 days, exceptionally more than 8 days) predominated. For aminopenicillins, the rate of occurrence of skin reactions increased with increasing exposure time up to 12 days, and then markedly diminished. Possibly due to the tendency to withdraw suspected drugs even in the case of minor (e.g., maculopapular) skin reactions, no severe events such as erythema multiforme major/Stevens-Johnson syndrome or toxic epidermal necrolysis occurred.