Therapeutic drug monitoring in the ICU: clinical and economical implications

Background: Therapeutic drug monitoring (TDM) is the clinical measurement of drug concentrations in biological samples to ensure target exposure. Critically ill patients are subject to abnormal physiological parameters and organ dysfunction, so antibiotic pharmacokinetics is affected by high and unpredictable variability and standard dosing may fail to achieve target exposure. TDM has been proposed to individualize antibiotic dosing in this population, but implementation remains limited, partly by the cost and logistical requirements of monitoring and partly because evidence of benefit, both clinical and economic, remains inconsistent. Aim of the study: This pilot pharmaco-economic study evaluates whether implementing systematic TDM for antibiotic therapy in the intensive care unit could be beneficial, assessing its potential for cost savings, target attainment, and the number of patients requiring monitoring to inform a clinically relevant decision. Materials and methods: This pilot study had an observational, single-cohort design with an intra-patient comparison, each patient serving as their own control, conducted in the intensive care unit (ICU) of Padova University Hospital. Observed TDM-guided expenditure was compared against the cost of a day-by-day reconstructed reference dose, representing the dose the patient would have received in the absence of TDM, updated with dose modifications based on clinical judgement rather than TDM. Antibiotics were valued at ex-factory price rather than at the actual hospital-specific price. Measured variables were net saving per patient and per episode and number needed to monitor (NNM); target attainment, defined as TDM within clinically defined limits, was compared between first and last sample of each episode. Confidence intervals (CI) were calculated by percentile bootstrap with 10,000 resamples, with smoothed bootstrap used for the median net saving; the Wilcoxon signed-rank test assessed whether the median net saving differed from zero, and the McNemar test was used for the paired comparison of target attainment. Results: Patients were admitted to the ICU between January 17 and July 3, 2026, for a total of 104 patients and 196 antibiotic therapy episodes, defined as a single cycle of treatment with a defined antibiotic for a specific pathogen. TDM was performed when available, for a total of 393 measurements. Net saving for the cohort was € 7,196.30 (95% CI -1,217.01 - 17,069.84); mean saving per patient was € 69.20 (95% CI -10.59 - 162.90); median saving per patient was € - 3.61 (95% CI -6.24 - 8.55). The Wilcoxon signed-rank test found no significant difference from zero (p = 0.508). Savings were concentrated in a minority of patients: of 104, 41 saved while 63 showed a net loss. Ceftazidime/avibactam, piperacillin/tazobactam and meropenem/vaborbactam showed the largest net savings, of € 2,371, € 2,270 and € 2,089; linezolid, by contrast, showed a net loss of € 1,246. Of 175 monitored episodes, 75 (42.9%) led to a dose modification (95% CI 35.4-50.5), resulting in an NNM of 2.3 (95% CI 2.0-2.8). Among the 110 episodes with at least two samples, attainment of the therapeutic range increased from 59/110 (54%) at first sample to 80/110 (73%) at last, with 28 episodes corrected into range against 7 worsened (McNemar p < 0.001). Conclusions: A net saving was observed at cohort level, although concentrated in a subset of episodes involving high-cost agents; there was no significant evidence that TDM produced a net saving per patient different from zero, and the typical patient showed a marginal increase in spending due to the fixed cost of testing. Given the pilot design, it is not possible to establish whether TDM is economically advantageous overall; results suggest its economic value depends on the cost of the agent monitored. TDM prompted a dose modification in 43% of monitored episodes and significantly improved attainment of the therapeutic range.