Purpose Amikacin requires pharmacodynamic targets of peak serum concentration ( C max ) of 8–10 times the minimal inhibitory concentration, corresponding to a target C max of 60–80 mg/L for the less susceptible bacteria. Even with new dosing regimens of 25 mg/kg, 30 % of patients do not meet the pharmacodynamic target. We aimed to identify predictive factors for insufficient C max in a population of critically ill patients. Methods Prospective observational monocentric study of patients admitted to a general ICU and requiring a loading dose of amikacin. Amikacin was administered intravenously at the dose of 25 mg/kg of total body weight. Independent determinants of C max  < 60 mg/L were identified by mixed model multivariate analysis. Results Over a 1-year period, 181 episodes in 146 patients (SAPS 2 = 51 41–68) were included. At inclusion, the SOFA score was 8 6–12, 119 (66 %) episodes required vasopressors, 150 (83 %) mechanical ventilation, and 81 (45 %) renal replacement therapy. The amikacin C max was 69 54.9–84.4 mg/L. Overall, 60 (33 %) episodes had a C max  < 60 mg/L. The risk of C max  < 60 mg/L associated with BMI < 25 kg/m 2 varied across quarters of inclusion. Independent risk factors for C max  < 60 mg/L were a BMI < 25 kg/m 2 over the first quarter (odds ratio (OR) 15.95, 95 % confidence interval (CI) 3.68–69.20, p  < 0.001) and positive 24-h fluid balance (OR per 250-mL increment 1.06, 95 % CI 1.01–1.11, p  = 0.018). Conclusions Despite an amikacin dose of 25 mg/kg of total body weight, 33 % of patients still had an amikacin C max  < 60 mg/L. Positive 24-h fluid balance was identified as a predictive factor of C max  < 60 mg/L. When total body weight is used, low BMI tended to be associated with amikacin underdosing. These results suggest the need for higher doses in patients with a positive 24-h fluid balance in order to reach adequate therapeutic targets.