Characteristics of study participants

We included 35 patients in the study. Table 1 shows the characteristics of patients between initial daily dosing of GM. Among the study participants, 22 patients initially received once-daily GM dosing, 13 received multiple-daily GM dosing (one patient received twice-daily GM dosing, and 12 received GM dosing thrice daily). At the start of GM treatment, there were no significant differences in age, sex, weight, BMI, renal function, daily GM dosage, and concomitant medications between patients with once-daily GM dosing and those with multiple-daily GM dosing (Table 1). GM trough concentrations were measured at a median of 4 days after GM administration in the once-daily and multiple-daily dosing groups. The site of infection and causative organisms are shown in Supplemental Digital Content (see Table S1). In addition, 29 of the 35 patients had their peak concentrations measured at the time of the initial blood concentration assessment. The median peak concentration for the once-daily group (n = 19) was 12.3 (interquartile range: 9.2, 16.2) µg/mL, while that for the multiple-dose group (n = 10) was 4.7 (3.5, 6.3) µg/mL.

Table 1 Comparison of demographic and clinical characteristics of study participants between once-daily dosing and multiple daily dosing groupsInfluence of daily dosing frequency on trough concentrations of gentamicin

The median of initial GM trough concentrations was 0.45 (interquartile range; 0.33, 0.88) µg/mL and 1.10 (0.90, 1.60) µg/mL in the group of GM once-daily dosing and the group of multiple-daily dosing, respectively. The multiple daily dosing group had significantly higher trough concentrations (P = 0.016; Fig. 1A). The frequency of patients with an initial GM trough concentration lower than 1.0 µg/mL was 77% (17/22) in the once-daily group and 31% (4/13) in the multiple-daily group (P = 0.012). At our institute, dose optimization is performed using TDM after checking the initial GM trough concentrations. The minimum trough concentration within 2 weeks of GM initiation was 0.45 (0.33, 0.78) µg/mL in the once-daily group and 0.90 (0.60, 1.00) µg/mL in the multiple daily groups, which was still significantly higher in the multiple-daily dosing group (P = 0.042, Fig. 1B). However, there were no significant differences in the frequency of minimum trough concentrations lower than 1.0 µg/mL between the once-daily and multiple-daily groups [82% (18/22) vs. 62% (8/13), P = 0.243].

Fig. 1

Effect of daily dosing frequency on gentamicin trough concentration. (A) Comparison of the initial trough concentrations of serum gentamicin between the once-daily dosing and multiple-daily dosing groups. (B) Comparison of the minimum trough concentration

Association with the incidence of acute kidney injury

Nine patients (26%) developed AKI after GM treatment initiation. There were no significant differences in age, sex, weight, BMI, renal function at the start of GM treatment, days to initial GM trough concentration measurement, infectious diseases, or concomitant medications between the patients with and without AKI (Table 2). In addition, we evaluated the association between the daily GM dosing frequency and AKI. However, the incidence rate of AKI was not significantly different between participants with once-daily GM dosing and those with multiple-daily GM dosing (Table 2). The initial GM trough concentrations were slightly higher in patients with AKI than in those without AKI; however, this difference was not statistically significant [median (interquartile range):0.90 (0.70, 1.60) µg/mL vs. 0.65 (0.40, 1.08) µg/mL, P = 0.215, Fig. 2A]. In contrast, the minimum trough concentration within 2 weeks of GM initiation was significantly higher in patients with AKI than in patients without AKI [0.80 (0.70, 1.60) µg/mL vs. 0.50 (0.40, 0.88) µg/mL, P = 0.037, Fig. 2B]. The site of infection, causative organism, and concomitant antimicrobials are shown in the Supplementary Digital Content (see Table S2).

Table 2 Comparison of study participants with and without acute kidney injury (AKI)Fig. 2

Associations between gentamicin trough concentration and the incidence of acute kidney injury. (A) Comparison of the initial serum gentamicin trough concentrations between patients with and without acute kidney injury. (B) Comparison of the minimum trough concentration

Association of trough concentrations of gentamicin on the change of renal function

Because there was no significant influence of daily dosing frequency on the incidence rate of AKI, we evaluated the association between GM trough concentrations and changes in serum creatinine levels. The relative change in serum creatinine within seven days of GM administration was significantly correlated with the initial GM trough concentrations (r = 0.338, P = 0.047, Fig. 3A) and minimum GM trough concentration (r = 0.495, P = 0.003, Fig. 3B). We confirmed these relationships after adjusting for the influence of concomitant drugs that may affect renal function, using multivariate linear regression analysis. The minimum GM trough concentrations [coefficient (β) = 0.28 (95% CI; 0.10, 0.47), P = 0.004], but not initial trough concentration [β = 0.15 (− 0.01, 0.30), P = 0.065] maintain statistically significant relationship to the relative change in serum creatinine. Finally, we evaluated the association between minimum GM trough concentration and the incidence of AKI. In logistic regression analysis, the odds ratio for AKI was 9.2 (95% CI; 1.3 to 65.5, P = 0.026) per 1.0 µg/mL increase in the minimum GM trough concentrations. The ROC curve analysis revealed that the minimum GM trough concentrations could predict the incidence of AKI with an AUC of 0.73 (95% CI; 0.53, 0.94), a sensitivity of 67%, and a specificity of 69%. The best cut-off value of minimum GM trough concentrations was 0.75 µg/mL to predict the incidence of AKI (Fig. 4).

Fig. 3

Association of gentamicin trough concentrations with changes in renal function. (A) Correlation between the initial trough concentrations of serum gentamicin and fold change in serum creatinine within seven days after gentamicin administration. (B) Correlation between the minimum trough concentration and the fold-change in serum creatinine levels

Fig. 4

Receiver operating characteristic (ROC) curve for predicting patients with acute kidney injury using minimum trough concentration