An automated solid-phase extraction procedure combined with the gas chromatography-mass spectrometry (GC-MS) methodology, without derivatization, has been developed for the determination of ketamine (K), norketamine (NK), and dehydronorketamine (DHNK) in urine. The analytical approach is simple and rapid, yet reliable, achieving good linearity (r(2)>0.999 over the concentration range of 30 to 1000 ng/mL), sensitivity (limits of quantification = 15, 10, and 20 ng/mL for K, NK, and DHNK, respectively), accuracy (90-104%), and precision (RSD<8.1%) for all analytes. Two hundred and six urine specimens collected from suspected drug users were analyzed by this protocol and also screened by Neogen ELISA method to evaluate the efficiency as well as the compatibility of these two methods. Neogen ELISA showed high efficiency (98.1%), high sensitivity (90.9%), high specificity (98.9%), low false-positive rate (1.1%), and moderate false-negative rate (9.1%), adopting 10 ng/mL K as the cutoff. Neogen ELISA screening followed by GC-MS analysis appeared to be a good screening-confirmation test scheme for the analysis of K in urine. Twenty of the 22 positive urine specimens contained all three analytes simultaneously, with DHNK showing the highest and K the lowest concentrations.