AbstractWe investigated the effectiveness of monitoring serum carnitine levels in hemodialysis patients receiving L‐carnitine supplementation. One‐hundred forty‐five hemodialysis patients were divided into three groups. Group 1 consisted of patients (n = 30) who had been receiving supplementation before this study and then discontinued at the beginning. The remaining patients were divided into Group 2 (n = 13) and Group 3 (n = 102) based on their baseline free carnitine (FC) level, <20 or ≥ 20 μmol/L. Group 2 was started on supplementation, and Groups 1 and 3 were observed without any intervention for the first 6 months. FC was measured every 6 months in all three groups up to 18 months. All patients in whom FC was less than 20 μmol/L at 6 and 12 months were prescribed supplementation. After the first 6 months, the mean ± SD of FC changed from 262.5 ± 87.5 μmol/L at baseline to 70.8 ± 33.6 μmol/L (P < 0.001) in Group 1, from 17.4 ± 1.9 to 193.9 ± 43.3 μmol/L (P < 0.001) in Group 2, and from 49.2 ± 24.6 to 44.2 ± 19.8 μmol/L (P < 0.05) in Group 3. The acyl/free carnitine changed from 0.62 to 0.59 in Group 1 (P = 0.287), from 0.76 to 0.66 in Group 2 (P = 0.054) and from 0.57 to 0.60 in Group 3 (P < 0.05). Of the 145 patients, 126 continued follow‐up for the full 18 months. FC remained in the normal range (36–74 μmol/L) within the 95% CI. FC was considered a strong predictor of carnitine deficiency after 6 months (AUC: 0.9146, cut‐off value: 33.8 μmol/L). FC monitoring is essential for appropriate carnitine supplementation in hemodialysis patients.