To analyze the clinical and genetic characteristics of patients with pyridoxine dependent epilepsy (PDE), and build a method to detect and analyze the concentration of urinary pipecolic acid in PDE patients receiving pyridoxine treatment. Twelve patients (8 were male, 4 were female) were diagnosed as PDE in Peking University First Hospital between April 2012 and September 2015. The clinical manifestations, diagnosis and treatment process, video-electroencephalogram, magnetic resonance imaging were retrospectively analyzed. ALDH7A1 gene was detected using Sanger sequencing or targeted next-generation sequencing. The concentration of urinary pipecolic acid in PDE patients was detected with gas chromatography-mass spectrometry (GC-MS), as well as in some non-PDE children served as normal control. All controls, 58 cases totally, were neonates born in our hospital or children came to our hospital for reasons such as syncope (without disturbing pipecolic acid metabolism) from November 2015 to January 2016. Of them, 25 were ≤6 months old(14 were male, 11 were female), 33 were >6 months old(14 were male, 19 were female). The Student's t-test or Mann-Whitney U test was used for comparing the pipecolic acid between the two groups. Correlation analysis was conducted using Pearson or Spearman test. Of the 12 patients, seven of them were abnormal at birth. The age of epilepsy onset was from 5 h to 5 months, within 10 d in 8 patients. After a diagnostic delay time of 15 d to 20 months, seizures in all patients were controlled by pyridoxine monotherapy, at a dose of higher than 10.0 mg/(kg·d) in 10 patients, and 8.5 and 2.5 mg/(kg·d) in the other 2 patients respectively. The range of maintenance dose was 2.5-20.0 mg/(kg·d) during the follow up. Interictal electroencephalogram showed nonspecific abnormality in 10, normal in 2. Brain magnetic resonance imaging showed nonspecific abnormality in 7, normal in 5. ALDH7A1 mutations were found in all patients, including 15 different mutation sites, four of which were never reported before. Splicing mutation IVS11+ 1G>A was carried in 6 patients, with a frequency of 25% (6/24). At the last follow-up, eleven patients were in various degree of psychomotor development delay, including the 4 patients with severe delay in whom birth abnormalities presented, and no significant delay was found in one patient. The concentration of urinary pipecolic acid in control: age ≤6 months, median 8.47 (0.46-35.33) mmol/mol creatinine; age >6 months, median 0.66 (0.12-3.52) mmol/mol creatinine. The concentration of urinary pipecolic acid was different between two groups of control (Z=-5.464, P<0.01). Twelve patients were all older than 6 months when they were tested, and the concentration was only mildly elevated in one patient, the range of 12 patients was 0.14-4.08 mmol/mol creatinine. The concentration was not significantly different between the control with age >6 months and our PDE patients (Z=-0.655, P>0.05). There were no significant correlations between the concentration of urinary pipecolic acid and the initial dose or maintenance dose at last follow-up of pyridoxine (r=0.418 and 0.166, P=0.176 and 0.607). Seizures start in early infancy in most PDE patients. The splicing mutation IVS11+ 1G>A is supposed to be a probable"hotspot"mutation with a high frequency in Chinese PDE patients. Most patients have different levels of psychomotor development delay after seizures are controlled, and the patients with birth abnormalities may have worse outcomes. No relationships between the diagnostic delay time and the development outcome are found. The concentration of urinary pipecolic acid can return to normal during treatment with pyridoxine.