2024 年 47 巻 5 号 p. 988-996
Patients with hematological malignancies (HM) often receive tazobactam/piperacillin (TAZ/PIPC) and glycopeptide antibiotics for febrile neutropenia. The effect of concomitant use of TAZ/PIPC on risk of teicoplanin (TEIC)-associated acute kidney injury (AKI) remains unclear. We investigated the impact of concomitant TAZ/PIPC use on TEIC-associated AKI in HM patients and identified the risk factors. In this retrospective, single-center, observational cohort study, 203 patients received TEIC, 176 of whom satisfied the selection criteria and were divided into TEIC cohort (no TAZ/PIPC; n = 118) and TEIC + TAZ/PIPC cohort (n = 58). AKI was defined as serum creatinine increase ≥0.3 mg/dL within 48 h or ≥50% from baseline. Incidence of AKI in TEIC cohort before and after propensity score matching was 9.3 and 5.9%, respectively, and that in TEIC + TAZ/PIPC cohort was 10.3 and 11.8%. AKI incidence and risk were not significantly different between two cohorts before (p = 0.829; odds ratio (OR) 1.122, 95% confidence interval (CI) 0.393–3.202) and after matching (p = 0.244; OR 2.133, 95% CI 0.503–9.043). Logistic regression analysis with factors clinically or mechanistically potentially related to TEIC-associated AKI, including concomitant TAZ/PIPC use, as independent variables identified baseline hemoglobin level as the only significant risk factor for TEIC-associated AKI (p = 0.011; OR 0.484, 95% CI 0.276–0.848). In HM patients treated with TEIC, concomitant TAZ/PIPC use did not increase AKI risk whereas lower hemoglobin levels had higher risk for TEIC-associated AKI development, suggesting the necessity to monitor serum creatinine when using TEIC in patients with anemia.
Tazobactam/piperacillin (TAZ/PIPC), an antibiotic containing a beta-lactamase inhibitor, has a broad antibacterial spectrum against many Gram-positive and Gram-negative aerobic and anaerobic bacteria, including many pathogens producing beta-lactamases.1) It is widely used for treating various infections and frequently used as a first-line drug for febrile neutropenia (FN) treatment. In addition, concomitant use of a glycopeptide antibiotic; vancomycin (VCM) or teicoplanin (TEIC), is recommended for FN treatment when methicillin-resistant Staphylococcus aureus (MRSA) infection is suspected, or response to broad-spectrum antibiotics such as TAZ/PIPC is inadequate. Therefore, patients with hematological malignancies have many opportunities to receive VCM or TEIC in addition to TAZ/PIPC.
VCM has broad-spectrum activities for Gram-positive bacteria such as MRSA, but accumulates in renal tubular epithelial cells and exerts direct toxicity to damage renal tubules, frequently causing acute kidney injury (AKI).2) Histologically, AKI is characterized by acute interstitial nephritis and acute tubular necrosis, and the mechanism of cast nephropathy has also been demonstrated.3) Risk factors for VCM-associated AKI include elevated trough levels, concomitant use of nephrotoxic agents such as aminoglycoside, amphotericin B, furosemide, non-steroidal anti-inflammatory drugs (NSAIDs), contrast media, and vasopressor; patients with impaired renal function, dehydration, and severe illness; use of high-dose VCM; admission to ICU; and treatment period of more than one week.4)
VCM alone has a risk of renal damage, but many reports in recent years have shown that concomitant use of TAZ/PIPC increases that risk. Burgess and Drew5) reported that concomitant use of VCM and TAZ/PIPC increased the incidence of nephrotoxicity (odds ratio 2.48) compared to VCM alone. Oda et al.6) revealed that AKI incidence in the VCM plus TAZ/PIPC group was 44.6%, compared to 17.6% in the group with either VCM or TAZ/PIPC replaced. Systematic reviews have also shown that concomitant use of VCM and TAZ/PIPC increases the frequency of renal damage.7,8) Furthermore, multiple reports suggest that compared with other broad-spectrum antibiotics such as cefepime and meropenem, VCM increases the risk of AKI when concomitantly administered with TAZ/PIPC.9–11)
TEIC has an antibacterial spectrum similar to VCM and is used to treat infections caused by resistant Gram-positive bacteria. Because of the lower incidence of AKI compared to VCM, TEIC can be used as an alternative when renal damage is a concern.12) Therefore, TEIC is sometimes selected when adding a glycopeptide antibiotic to TAZ/PIPC in patients with FN having high risk of renal damage by using VCM concomitantly. A study comparing the efficacy and safety between VCM and TEIC in FN patients showed similar clinical efficacy for both drugs, whereas serum creatinine increase tended to be higher in VCM-treated than in TEIC-treated group.13) Although a study by Shao et al.14) showed no significant difference in AKI risk between TEIC plus TAZ/PIPC combination and VCM plus TAZ/PIPC combination, several recent reports have shown a significantly lower incidence of AKI in TEIC plus TAZ/PIPC than in VCM plus TAZ/PIPC.15,16) Tai et al.17) demonstrated no difference in AKI risk between concomitant use of TEIC with TAZ/PIPC and that of TEIC with other anti-pseudomonal β-lactams. In contrast, a single-center retrospective cohort study of 4202 patients suggested an association of TEIC plus TAZ/PIPC with a higher prevalence of AKI compared with monotherapy, although the overall decline in renal function was small.18) However, there is no report evaluating whether TAZ/PIPC is a risk factor of TEIC-associated AKI.
Given the above background, this study compared AKI incidence between TEIC (no TAZ/PIPC) therapy and TEIC plus TAZ/PIPC combination therapy in patients with hematological malignancies and identified the risk factor of TEIC-associated AKI using patient background and clinical laboratory parameters including concomitant use of TAZ/PIPC.
The study design was a retrospective, single-center, observational cohort study. The electronic medical records at the Department of Hematology in Oita University Hospital from April 2014 to November 2020 were reviewed to identify patients who received intravenous TEIC. The exclusion criteria were as follows: age under 18 years; admission to intensive care unit or emergency center during TEIC treatment; no measurement of trough TEIC concentration; addition or discontinuation of TAZ/PIPC during TEIC treatment; switch from VCM; and death during TEIC treatment. The study was conducted in accordance with the ethical standards of our institute and the principles of Helsinki Declaration of 1975, as revised in 2013. The study was started following approval by the Ethics Committee of Oita University Faculty of Medicine (Review Reference Number: 2254).
Data CollectionThe following demographic and laboratory parameters at the initiation of TEIC treatment were collected: sex, age, body weight, height, highest daily body temperature, albumin (ALB), white blood cell count (WBC), hemoglobin (HGB), platelet count (PLT), C-reactive protein (CRP), blood urea nitrogen (BUN), serum creatinine (Scr), and primary disease. Creatinine clearance (Ccr) was calculated with Cockcroft–Gault formula.19) Primary disease was extracted from doctor’s medical records. Primary disease and infection type were extracted from doctor’s medical records. Infection type includes the suspected case. Concomitant drugs other than TAZ/PIPC included aminoglycosides, amphotericin B, loop diuretics, calcineurin inhibitors, and cisplatin, which are listed in the Japanese package insert of TEIC (Targocid®) as having an increased risk for renal impairment; as well as sulfamethoxazole/trimethoprim, acyclovir, NSAIDs, angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers, which possibly cause renal impairment. Concomitant use was defined as when one or more of these drugs were administered for at least three consecutive days during TEIC treatment.20) Topical formulations were not included in concomitant drugs. Serum TEIC concentrations were measured by a latex immunoturbidimetric assay using a TBA-25FR® analyzer (Canon Medical Systems Co., Otawara, Japan). The highest trough concentration after more than three days of TEIC administration was used for data analysis.
Definition of AKIAKI was defined as an increase in Scr level by ≥0.3 mg/dL within 48 h or by ≥50% from baseline, following the Kidney Disease Improving Global Outcomes (KDIGO) criteria.21) Urine volume was not considered in the definition of AKI, because the data was not available. Scr levels were monitored from the beginning of TEIC treatment to the subsequent blood sampling after the end of treatment.
Statistical AnalysisPredictive Analytics Software (PASW) Statistics version 27 (SPSS Inc., Chicago, IL, U.S.A.) was used for statistical analysis. Data are expressed as number (%) for categorical variables and median [interquartile range] for continuous variables. Patients who met the selection criteria were classified into TEIC and TEIC + TAZ/PIPC cohorts as well as AKI and non-AKI cohorts. Data normality was assessed by Shapiro–Wilk test. Categorical variables were analyzed by chi-square test, or by Fisher’s exact test when more than 20% of the cells have expected frequencies <5. A p-value less than 0.05 was considered statistically significant.
One-to-one propensity score matching was implemented to adjust for differences in patient characteristics between TEIC (no TAZ/PIPC) and TEIC + TAZ/PIPC cohorts. A logistic regression model was adopted to estimate propensity scores, and c-index was calculated to assess the goodness of fit. Using the nearest neighbor matching method, one patient in one cohort was matched to one patient in the other cohort based on estimated propensity scores with the caliper width set at 0.2 standard deviations without replacement. The following demographic and laboratory variables potentially associated with AKI were incorporated into the one-to-one propensity score matching: sex, age, ALB, HGB, CRP, BUN, Ccr, infection type (febrile neutropenia and sepsis), and concomitant drugs. After propensity score matching, adjusted p-values and odds ratios were estimated using the same statistical methods as for the data before propensity score matching. The risk factors for the development of TEIC-associated AKI were analyzed by univariate and multivariate logistic regression model. For multivariate logistic regression analysis, independent variables comprising clinically or mechanistically potentially relevant factors with AKI development such as sex, age, highest trough TEIC concentration ALB, HGB, CRP, BUN, infection type (febrile neutropenia and sepsis), and concomitant drugs were inserted into the model by forced entry using Schwarz’s Bayesian information criterion, and variables remained in the model if p < 0.05. Aminoglycosides were excluded from the variable as they were used in only one patient.
Of 203 patients who received TEIC, 176 satisfied the inclusion criteria and were divided into those who received TEIC but not TAZ/PIPC (TEIC cohort, n = 118) and those who received concomitant TEIC and TAZ/PIPC (TEIC + TAZ/PIPC cohort, n = 58). The background and clinical laboratory parameters of all patients analyzed are summarized in Table 1. The male:female ratio was 54 : 46; median age was 57.0 [45.0–66.0] years; and median body weight was 55.7 [48.5–63.3] kg. The median highest trough TEIC concentration was 20.7 [16.7–24.7] µg/mL, and median Ccr before TEIC initiation was 84.3 [59.7–120.2] mL/min. The most frequent primary disease was acute myeloid leukemia (34.1%), followed by lymphoma (20.5%). The most frequent infection type was febrile neutropenia (64.2%), followed by fever during neutropenia without fulfilling diagnostic for febrile neutropenia (17.0%). NSAIDs (53.4%) were the most frequently used concomitant drugs with AKI risk.
Parameter | Value |
---|---|
Patients; n | 176 |
Sex (male/female); n (%) | 95 (54.0)/81 (46.0) |
Age (years) | 57.0 [45.0–66.0] |
Body weight (kg) | 55.7 [48.5–63.3] |
Height (cm) | 161.2 [155.8–168.5] |
Highest body temperature (°C) | 38.2 [37.8–38.7] |
Highest trough teicoplanine concentration (µg/mL) | 20.7 [16.7–24.7] |
Albumin | 3.03 [2.65–3.32] |
White blood cell count (×103/mm3) | 0.64 [0.15–1.84] |
Hemoglobin (g/dL) | 8.1 [7.6–9.0] |
Platelet count (×103/µL) | 29.0 [15.8–49.0] |
C-reactive protein (mg/dL) | 4.61 [2.06–8.77] |
Blood urea nitrogen (mg/dL) | 14.8 [11.0–21.7] |
Serum creatinine (mg/dL) | 0.68 [0.53–0.91] |
Creatinine clearance (mL/min) | 84.3 [59.7–120.2] |
Primary disease; n (%) | |
Acute myeloid leukemia | 60 (34.1) |
Acute lymphocytic leukemia | 20 (11.4) |
Adult T-cell leukemia | 4 (2.3) |
Multiple myeloma | 13 (7.4) |
Lymphoma | 36 (20.5) |
Myelodysplastic syndromes | 13 (7.4) |
Others | 30 (17.0) |
Infection type; n (%) | |
Febrile neutropenia | 113 (64.2) |
Pneumonia | 8 (4.5) |
Sepsis | 18 (10.2) |
Fever during neutropenia without fulfilling diagnostic for febrile neutropenia | 30 (17.0) |
Others | 7 (4.0) |
Concomitant drug; n (%) | |
Aminoglycoside | 1 (0.6) |
Sulfamethoxazole/trimethoprim | 70 (39.8) |
Amphotericin B | 7 (4.0) |
Acyclovir | 87 (49.4) |
Loop diuretic | 58 (33.0) |
Calcineurin inhibitor | 56 (31.8) |
Non-steroidal anti-inflammatory drugs | 94 (53.4) |
Angiotensin-converting enzyme inhibitors or angiotensin receptor blocker | 11 (6.3) |
Tazobactam/piperacillin | 58 (33.0) |
Data are expressed as number (%) for categorical variables, and median [interquartile range] for continuous variables.
Fifty-one patients in each cohort were matched by propensity scores. Tables 2, 3 present univariate analyses comparing patient characteristics and clinical laboratory parameters between TEIC and TEIC + TAZ/PIPC cohorts before and after propensity score matching, respectively. The highest trough TEIC concentration was significantly higher in TEIC cohort than in TEIC + TAZ/PIPC cohort before matching (p = 0.045), while the difference was no longer significant after matching (p = 0.216).
Parameter | TEIC | TEIC + TAZ/PIPC | p-Value |
---|---|---|---|
Patients; n | 118 | 58 | |
Sex (male/female); n (%) | 65 (55.1)/53 (44.9) | 30 (51.7)/28 (48.3) | 0.674 |
Age (years) | 57.0 [45.0–66.0] | 56.0 [46.0–64.8] | 0.760 |
Body weight (kg) | 56.4 [48.5–63.7] | 53.7 [48.4–62.4] | 0.446 |
Height (cm) | 161.9 [156.4–168.4] | 160.9 [154.1–168.5] | 0.582 |
Highest body temperature (°C) | 38.3 [37.9–38.6] | 38.2 [37.7–38.8] | 0.627 |
TEIC administration period | 9.0 [6.0–14.0] | 8.0 [5.0–12.0] | 0.069 |
Highest trough teicoplanine concentration (µg/mL) | 20.7 [18.1–25.1] | 19.9 [15.7–22.3] | 0.045 |
Albumin | 2.98 [2.64–3.27] | 3.11 [2.73–3.47] | 0.112 |
White blood cell count (×103/mm3) | 0.67 [0.20–1.82] | 0.61 [0.13–1.85] | 0.363 |
Hemoglobin (g/dL) | 8.2 [7.7–8.9] | 8.0 [7.6–9.1] | 0.935 |
Platelet count (×103/µL) | 29.0 [15.0–45.8] | 28.5 [16.0–51.3] | 0.552 |
C-reactive protein (mg/dL) | 5.18 [1.82–9.76] | 4.45 [2.50–7.43] | 0.761 |
Blood urea nitrogen (mg/dL) | 15.3 [11.1–23.7] | 13.6 [10.3–17.6] | 0.052 |
Serum creatinine (mg/dL) | 0.68 [0.55–0.91] | 0.70 [0.51–0.87] | 0.695 |
Creatinine clearance (mL/min) | 86.9 [58.3–118.0] | 79.1 [60.3–120.4] | 0.885 |
Infection type; n (%) | |||
Febrile neutropenia | 76 (64.4) | 37 (63.8) | 0.936 |
Pneumonia | 6 (5.1) | 2 (3.4) | 0.475 |
Sepsis | 11 (9.3) | 7 (12.1) | 0.572 |
Fever during neutropenia without fulfilling diagnostic for febrile neutropenia | 22 (18.6) | 8 (13.8) | 0.421 |
Others | 3 (2.5) | 4 (6.9) | 0.163 |
Concomitant drug; n (%) | |||
Aminoglycoside | 1 (0.8) | 0 (0.0) | 0.670 |
Sulfamethoxazole/trimethoprim | 44 (37.3) | 26 (44.8) | 0.337 |
Amphotericin B | 6 (5.1) | 1 (1.7) | 0.265 |
Acyclovir | 59 (50.0) | 28 (48.3) | 0.830 |
Loop diuretic | 44 (37.3) | 14 (24.1) | 0.081 |
Calcineurin inhibitor | 35 (29.7) | 21 (36.2) | 0.381 |
Non-steroidal anti-inflammatory drugs | 62 (52.5) | 32 (55.2) | 0.742 |
Angiotensin-converting enzyme inhibitors or angiotensin receptor blocker | 11 (9.3) | 5 (8.6) | 0.879 |
Data are expressed as number (%) for categorical variables, and median [interquartile range] for continuous variables. Categorical variables were analyzed by chi-square test or Fisher’s exact test. Continuous variables were analyzed by Mann–Whitney U test. TEIC, teicoplanin; TAZ/PIPC, tazobactam/piperacillin.
Parameter | TEIC | TEIC + TAZ/PIPC | p-Value | Std diff |
---|---|---|---|---|
Patients; n | 51 | 51 | ||
*Sex (male/female); n (%) | 27 (52.9)/24 (47.1) | 26 (51.0)/25 (49.0) | 0.843 | 0.039 |
*Age (years) | 53.0 [40.0–66.0] | 57.0 [46.0–64.5] | 0.715 | 0.003 |
Body weight (kg) | 54.2 [47.5–62.5] | 52.3 [47.9–62.8] | 0.875 | — |
Height (cm) | 162.5 [156.4–168.3] | 161.0 [154.0–168.5] | 0.585 | — |
Highest body temperature (°C) | 38.3 [38.0–38.6] | 38.1 [37.7–38.8] | 0.620 | — |
TEIC administration period | 9.0 [6.0–14.0] | 7.0 [4.0–12.0] | 0.063 | — |
Highest trough teicoplanine concentration (µg/mL) | 21.2 [18.6–24.5] | 20.1 [15.7–22.5] | 0.216 | — |
*Albumin (g/dL) | 2.98 [2.70–3.38] | 3.11 [2.65–3.46] | 0.763 | 0.010 |
White blood cell count (×103/mm3) | 0.70 [0.20–1.81] | 0.63 [0.13–1.73] | 0.154 | — |
*Hemoglobin (g/dL) | 8.0 [7.5–8.6] | 7.9 [7.6–9.1] | 0.896 | 0.020 |
Platelet count (×103/µL) | 32.0 [17.0–45.5] | 29.0 [16.5–49.0] | 0.875 | — |
*C-reactive protein (mg/dL) | 5.31 [2.26–9.98] | 4.39 [2.54–7.38] | 0.720 | 0.054 |
*Blood urea nitrogen (mg/dL) | 16.5 [11.4–21.7] | 13.6 [10.3–17.8] | 0.098 | 0.287 |
Serum creatinine (mg/dL) | 0.68 [0.55–1.02] | 0.66 [0.49–0.85] | 0.518 | — |
*Creatinine clearance (mL/min) | 90.4 [58.8–111.0] | 84.7 [60.4–123.4] | 0.792 | 0.154 |
Infection type; n (%) | ||||
*Febrile neutropenia | 33 (64.7) | 35 (68.6) | 0.674 | 0.083 |
Pneumonia | 5 (9.8) | 2 (3.9) | 0.218 | — |
*Sepsis | 3 (5.9) | 3 (5.9) | 0.661 | 0.000 |
Fever during neutropenia without fulfilling diagnostic for febrile neutropenia | 8 (15.7) | 10 (19.6) | 0.603 | — |
Others | 2 (3.9) | 1 (2.0) | 0.500 | — |
Concomitant drug; n (%) | ||||
*Aminoglycoside | 0 (0.0) | 0 (0.0) | — | — |
*Sulfamethoxazole/trimethoprim | 24 (47.1) | 22 (43.1) | 0.691 | 0.079 |
*Amphotericin B | 1 (2.0) | 1 (2.0) | 0.752 | 0.000 |
*Acyclovir | 23 (45.1) | 25 (49.0) | 0.692 | 0.079 |
*Loop diuretic | 18 (35.3) | 13 (25.5) | 0.282 | 0.214 |
*Calcineurin inhibitor | 24 (47.1) | 16 (31.4) | 0.105 | 0.326 |
*Non-steroidal anti-inflammatory drugs | 29 (56.9) | 28 (54.9) | 0.842 | 0.039 |
*Angiotensin-converting enzyme inhibitors or angiotensin receptor blocker | 4 (7.8) | 4 (7.8) | 0.642 | 0.000 |
Data are expressed as number (%) for categorical variables, and median [interquartile range] for continuous variables. Categorical variables were analyzed by chi-square test or Fisher’s exact test. Continuous variables were analyzed by Mann–Whitney U test. * indicates factors used for propensity score matching. TEIC, teicoplanin; TAZ/PIPC, tazobactam/piperacillin; Std diff, standardized difference.
Table 4 compares the incidence of AKI between TEIC and TEIC + TAZ/PIPC cohorts before and after propensity score matching. Eleven (9.3%) of 115 patients in TEIC cohort and 6 (10.3%) of 58 in TEIC + TAZ/PIPC cohort developed AKI, with no significant difference in incidence of AKI between the two cohorts (p = 0.829). Similarly, no significant difference between cohorts was found after matching (p = 0.244).
Before propensity score matching | After propensity score matching | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
TEIC | TEIC + TAZ/PIPC | p-Value | Odds ratio | 95%CI | TEIC | TEIC + TAZ/PIPC | p-Value | Odds ratio | 95%CI | |
Acute kidney injury (+) | 11 (9.3) | 6 (10.3) | 0.829 | 1.122 | 0.393–3.202 | 3 (5.9) | 6 (11.8) | 0.244 | 2.133 | 0.503–9.043 |
Acute kidney injury (−) | 107 (90.7) | 52 (89.7) | 48 (94.1) | 45 (88.2) |
Data are expressed as number (%). p-values, odds ratio, and 95% CI were obtained from chi-square test of TEIC vs. TEIC + TAZ/PIPC cohorts before and after propensity score matching. TEIC, teicoplanin; TAZ/PIPC, tazobactam/piperacillin; CI, confidence interval.
Of all patients, 17 (9.7%) satisfied the definition of AKI. In univariate analysis, the highest trough TEIC concentration was significantly higher and HGB level was significantly lower in AKI cohort than in non-AKI cohort (highest trough TEIC concentration: p = 0.016; HGB level: p = 0.004) (Table 5). In addition, AKI cohort used loop diuretics concomitantly with TEIC more frequently than non-AKI cohort (p = 0.017). A logistic regression analysis using 16 factors potentially related to TEIC-associated AKI as independent variables identified HGB level prior to TEIC initiation as the only significant risk factor for the development of AKI when using TEIC (p = 0.011; OR 0.484, 95% CI 0.276–0.848) (Table 6).
Parameter | Non-AKI | AKI | p-Value |
---|---|---|---|
Patients; n | 159 | 17 | |
Sex (male/female); n (%) | 88 (55.3)/71 (44.7) | 7 (41.2)/10 (58.8) | 0.265 |
Age (years) | 57.0 [45.0–66.0] | 55.0 [45.0–58.0] | 0.341 |
Body weight (kg) | 56.0 [48.3–63.8] | 55.2 [51.6–60.4] | 0.822 |
Height (cm) | 161.5 [156.2–168.5] | 160.8 [152.5–165.5] | 0.236 |
Highest body temperature (°C) | 38.2 [37.8–38.7] | 38.3 [37.5–38.7] | 0.930 |
TEIC administration period | 8.0 [5.0–13.0] | 9.0 [5.0–20.0] | 0.284 |
Highest trough teicoplanine concentration (µg/mL) | 20.5 [16.4–23.5] | 24.9 [20.7–27.1] | 0.016 |
Albumin (g/dL) | 3.04 [2.67–3.33] | 2.89 [2.56–3.21] | 0.410 |
White blood cell count (×103/mm3) | 0.58 [0.15–1.80] | 0.79 [0.28–2.37] | 0.336 |
Hemoglobin (g/dL) | 8.3 [7.7–9.1] | 7.8 [7.2–8.0] | 0.004 |
Platelet count (×103/µL) | 30.0 [16.0–49.0] | 16.0 [15.0–40.0] | 0.104 |
C-reactive protein (mg/dL) | 4.72 [2.16–8.87] | 2.50 [1.30–7.66] | 0.454 |
Blood urea nitrogen (mg/dL) | 14.5 [11.0–21.7] | 17.0 [11.1–20.9] | 0.335 |
Serum creatinine (mg/dL) | 0.69 [0.54–0.92] | 0.64 [0.52–0.77] | 0.531 |
Creatinine clearance (mL/min) | 83.9 [59.3–118.6] | 103.4 [75.4–122.9] | 0.510 |
Infection type; n (%) | |||
Febrile neutropenia | 101 (63.5) | 12 (70.6) | 0.564 |
Pneumonia | 8 (5.0) | 0 (0.0) | 0.436 |
Sepsis | 17 (10.7) | 1 (5.9) | 0.459 |
Fever during neutropenia without fulfilling diagnostic for febrile neutropenia | 27 (17.0) | 3 (17.6) | 0.583 |
Others | 6 (3.8) | 1 (5.9) | 0.515 |
Concomitant drug; n (%) | |||
Tazobactam/piperacillin | 52 (32.7) | 6 (35.3) | 0.829 |
Aminoglycoside | 0 (0.0) | 1 (5.9) | 0.097 |
Sulfamethoxazole/trimethoprim | 63 (39.6) | 7 (41.2) | 0.901 |
Amphotericin B | 5 (3.1) | 2 (11.8) | 0.138 |
Acyclovir | 77 (48.4) | 10 (58.8) | 0.415 |
Loop diuretic | 48 (30.2) | 10 (58.8) | 0.017 |
Calcineurin inhibitor | 49 (30.8) | 7 (41.2) | 0.156 |
Non-steroidal anti-inflammatory agent | 84 (52.8) | 10 (58.8) | 0.638 |
Angiotensin-converting enzyme inhibitors or angiotensin receptor blocker | 16 (10.1) | 0 (0.0) | 0.182 |
Data are expressed as number (%) for categorical variables, and median [interquartile range] for continuous variables. Categorical variables were analyzed by chi-square test or Fisher’s exact test. Continuous variables were analyzed by Mann–Whitney U test.
Covariate (multivariate analysis) | Coefficient | p-Value | Odds ratio | |
---|---|---|---|---|
Estimate | 95% CI | |||
Sex | −0.721 | 0.279 | 0.486 | 0.132 to 1.792 |
Age | −0.014 | 0.543 | 0.986 | 0.944 to 1.031 |
Highest trough teicoplanin concentration | 0.07 | 0.139 | 1.073 | 0.978 to 1.177 |
Albumin | −0.455 | 0.536 | 0.635 | 0.15 to 2.682 |
Hemoglobin | −0.727 | 0.011 | 0.484 | 0.276 to 0.848 |
C-reactive protein | −0.039 | 0.466 | 0.961 | 0.865 to 1.069 |
Blood urea nitrogen | 0.04 | 0.111 | 1.041 | 0.991 to 1.094 |
Febrile neutropenia | 0.138 | 0.859 | 1.147 | 0.253 to 5.204 |
Sepsis | 1.384 | 0.295 | 3.991 | 0.299 to 53.185 |
Tazobactam/piperacillin | −0.703 | 0.284 | 0.495 | 0.137 to 1.793 |
Sulfamethoxazole/trimethoprim | 0.361 | 0.577 | 1.435 | 0.402 to 5.12 |
Amphotericin B | −1.774 | 0.117 | 0.170 | 0.019 to 1.555 |
Acyclovir | −0.003 | 0.996 | 0.997 | 0.270 to 3.681 |
Loop diuretic | −1.221 | 0.055 | 0.295 | 0.085 to 1.026 |
Calcineurin inhibitor | −0.276 | 0.652 | 0.759 | 0.229 to 2.517 |
Non-steroidal anti-inflammatory drugs | −0.592 | 0.354 | 0.553 | 0.158 to 1.936 |
Angiotensin-converting enzyme inhibitorsor angiotensin receptor blocker | 19.987 | 0.997 | <0.001 | — |
CI, confidence interval.
TAZ/PIPC combined with glycopeptide antibiotics are often used in patients with hematological malignancies who develop severe infections. Many previous reports have revealed significant increases in AKI incidence with concomitant use of VCM and TAZ/PIPC.5–11) Hence, TEIC that has lower AKI risk than VCM is often selected when used with TAZ/PIPC in patients with hematological malignancies. However, to our knowledge, no reports have evaluated whether TAZ/PIPC is a risk factor for AKI associated with TEIC. This study retrospectively explored the impact of concomitant TAZ/PIPC use on TEIC-associated AKI in patients with hematological malignancies and identified potential risk factors. The following results were obtained: (1) there were no significant differences in AKI incidence and odds ratio between TEIC and TEIC + TAZ/PIPC cohorts, even after propensity score matching; (2) univariate and multivariate analyses identified baseline HGB level as a significant risk factor for TEIC-associated AKI, but not concomitant use of TAZ/PIPC.
No difference in AKI incidence between TEIC and TEIC + TAZ/PIPC cohorts was found even after propensity score matching. This result is consistent with a previous report.17) Although the underlying pathogenesis of VCM-related AKI is not fully understood, experimental studies support the involvement of mitochondrial dysfunction, proinflammatory oxidative stress, and renal tubular cell apoptosis.22) On the contrary, there are no reports of in vitro and in vivo experimental studies on the mechanism of TEIC-associated AKI, although it cannot be ruled out that TEIC causes AKI by similar mechanisms as VCM. Beta-lactam antibiotics including TAZ/PIPC are well-known to cause drug-associated acute interstitial nephritis.23) No data from previous experimental models, human renal biopsies, and biomarker studies demonstrate a synergistic increase in acute tubular injury by combining TAZ/PIPC with VCM compared to VCM alone.24) The mechanisms of renal damage caused by VCM and TAZ/PIPC are different, and co-administration may increase the risk additively. A systematic review and meta-analysis demonstrated less frequent nephrotoxicity with TEIC than with VCM (relative risk 0.44; 95% CI 0.32 to 0.61).25) Hence, the risk of developing TEIC-associated AKI may not increase by combining TAZ/PIPC with TEIC, unlike with VCM. On the other hand, both TAZ and PIPC are known substrates of OAT1 and OAT3,26) and vancomycin partially downregulates OAT1 and OAT3 mRNA levels and protein expression.27) These drugs could diminish tubular transporters causing impaired creatinine secretion and elevated Scr, which resembles AKI, and is referred to as pseudo-toxic drug interaction. However, whether exposure of renal tubules to TEIC downregulates OAT1 and OAT3 expression remains unclear. While the possibility of pseudo-toxic drug interaction between TEIC and TAZ/PIPC is still controversial, our results suggest that the interaction would be unlikely because AKI risk associated with TEIC did not increase when TAZ/PIPC was used concomitantly.
This retrospective study recruited patients with hematological malignancies, whose renal function is likely to augment. This may lead to a pseudo-decrease in the incidence of AKI. However, the incidence of AKI in TEIC and TEIC + TAZ/PIPC cohorts was, respectively, 9.3 and 10.3% before propensity score matching, and 5.9 and 11.8% after matching, which is in line with previous literature.13,15–17,21) In addition, 17 of a total of 176 patients receiving TEIC developed AKI, which is comparable with the incidence in a previous systematic review and meta-analysis by Svetitsky et al. (94 of 914 patients).25) Univariate analysis found that the highest trough TEIC concentration in AKI cohort was significantly higher than that in non-AKI cohort. Clinical practice guidelines for therapeutic drug monitoring of TEIC revised by the Japanese Society of Chemotherapy and the Japanese Society of Therapeutic Drug Monitoring recommend maintaining the trough concentration within 15–30 µg/mL for uncomplicated MRSA infections and within 20–40 µg/mL for complicated or severe MRSA infections.28) Maintaining these ranges is suggested to improve treatment success without increasing the risk of adverse effects including AKI. The median highest trough TEIC concentration in AKI cohort was 24.9 µg/mL, and no patient exceeded 40 µg/mL. In addition, logistic regression analysis did not identify the highest trough TEIC concentration as a risk factor for AKI. These findings suggest a low impact of TEIC trough concentration on AKI development.
Univariate analysis showed that patients receiving loop diuretics were more likely to develop TEIC-associated AKI than those without loop diuretics (p = 0.017). Loop diuretics suppress sodium reabsorption by inhibiting the Na-K-2Cl cotransporter expressed in the ascending limb of the loop of Henle. Loop diuretics are the most commonly prescribed fluid management agents with indication for water and sodium retention due to heart disease, hypertension, and kidney diseases including AKI.29) However, many studies suggested an increased risk for AKI development with loop diuretic use. Levi et al.30) reported that furosemide use was associated with increased risk of AKI development in critically ill patients with sepsis or septic shock. A retrospective case-control study of 487372 individuals by Lapi et al.31) demonstrated an increased risk for AKI development by concomitant use of diuretics with NSAIDs and angiotensin receptor blockers/angiotensin-converting enzyme inhibitors. Another case-control study suggested synergistically amplified nephrotoxicity associated with combined use of NSAIDs with renin–angiotensin system inhibitors and/or diuretics.32) The underlying biological mechanism for loop diuretic-associated AKI remains unclear. However, loop diuretics can improve filling pressure and relieve water load in the thick ascending limbs of the loop of Henle by suppressing sodium reabsorption, which could cause volume depletion and possible renal hypoperfusion.33) In addition, loop diuretics could bind to antigens in the kidney or act as antigens, thereby depositing in the interstitium to induce immune reactions causing acute interstitial nephritis.34) A recent systematic review and meta-analysis identified loop diuretics as a risk factor for VCM-associated acute kidney injury.35) Loop diuretics reduce volume distribution of VCM by lowering the extracellular volume and deteriorate glomerular filtration rate by increasing urine volume.36) This could lead to increased VCM exposure and nephrotoxicity. Since TEIC is also primarily excreted by glomerular filtration, exposure to TEIC may increase when used with loop diuretics. However, this possibility would be unlikely because TEIC has lower urinary excretion rate and higher threshold for nephrotoxicity than VCM. In fact, loop diuretics were not identified as a risk factor for TEIC-associated AKI in multivariate logistic regression analysis.
Univariate and multivariate analysis identified only HGB level prior to TEIC initiation as a significant risk factor for AKI (p = 0.011). Many prospective and retrospective studies have demonstrated that low HGB is a significant risk factor for AKI associated with sepsis,37) cardiovascular surgery,38) and medications that possibly cause renal impairment.39–41) Sreenivasan et al.39) reported that severity of anemia defined by hemoglobin level was a risk factor for contrast-induced AKI following coronary angiography. In addition, Chou et al.41) identified low albumin or hemoglobin levels as relevant risk factors for parenteral aminoglycosides-associated AKI in a large study of 43259 individuals without a history of chronic kidney disease. Low HGB level causes oxygen depletion as a result of reduced oxygen-carrying capacity in the blood42,43) and decreased hemoglobin level in overt bleeding may reduce renal perfusion pressure.44) As a result, insufficient hemoglobin cannot deliver sufficient oxygen to the kidney tissue, leading to subsequent kidney damage. As aforementioned, the molecular mechanism of TEIC-associated AKI has not been elucidated, but assuming that it is similar to VCM, renal tubular apoptosis and mitochondrial dysfunction are likely to be involved. Therefore, low baseline HGB level was identified as a risk factor presumably reflecting an additive effect of AKI pathogenesis derived from different mechanisms.
This retrospective, single-center, observational cohort study has some limitations. First, the sample sizes in both cohorts were small, with only 51 patients in each cohort after propensity score matching. Many factors were potentially associated with AKI relative to the sample size, resulting in a standardized difference of several covariates after matching >0.1, suggesting inadequate variable balancing. Second, subject selection bias may have occurred when physicians chose TEIC rather than VCM for the treatment regimen. Concrete evidence of increased VCM-associated AKI risk with concomitant use of TAZ/PIPC has been established. Hence, physicians may avoid using VCM in patients receiving TAZ/PIPC, who had a high risk of AKI such as those with reduced kidney function. In other words, the TEIC + TAZ/PIPC cohort possibly had more patients with AKI risk at baseline. This study included only patients with hematological malignancies, and confounding factors were adjusted for patient background and clinical laboratory parameters using propensity score. However, since this propensity score analysis cannot include all the relevant factors in the model, some confounding factors would remain. Third, urine volume was not included in the definition of AKI in this study, despite using AKI criteria according to KDIGO, because these data were not available in many patients. This study could have missed patients with AKI who would have been included if urine volume were considered. Fourth, the severity of the infection is associated with the risk of developing AKI, but this study excluded patients admitted to the emergency center or intensive care unit during TEIC administration. Therefore, the Glasgow Coma Scale, arterial blood gases, and the severity of organ dysfunction were not measured in some patients, and the severity of infection could not be evaluated by APACHE II and SOFA scores in all patients. Finally, serum cystatin C is a better indicator than serum creatinine for the early diagnosis of AKI and may, therefore, provide a more accurate assessment of AKI associated with TEIC. However, it was not possible to obtain sufficient serum cystatin C levels from electronic medical records to assess AKI in many patients.
In conclusion, a propensity score matching analysis found no significant increase in incidence of TEIC-associated AKI by concomitant use of TAZ/PIPC. Univariate and multivariate regression analyses identified baseline HGB level as a risk factor for developing TEIC-associated AKI, but not concomitant TAZ/PIPC use. Given these results, use of TAZ/PIPC in combination did not increase the risk of developing TEIC-associated AKI in patients with hematologic malignancies. On the other hand, lower HGB level had a higher risk of developing TEIC-associated AKI, suggesting the necessity to carefully monitor Scr levels when using TEIC in patients with anemia.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
The authors declare no conflict of interest.