Abstract
Background: Vancomycin and piperacillin-tazobactam are commonly used first guns in the empiric management of critically ill patients. Current studies suggest an increased prevalence of acute kidney injury with concomitant use, however, these studies are few and limited by small sample size. The purpose of this study was to compare the prevalence of nephrotoxicity after treatment with vancomycin alone and concomitant vancomycin and piperacillin-tazobactam treatment at our institution.
Hypothesis: Concomitant vancomycin and piperacillin-tazobactam–treated patients will experience greater prevalence of nephrotoxicity compared with vancomycin-only treated patients.
Methods: This was a retrospective cohort of patients treated with vancomycin for gram-positive or mixed infections in our facility from 2005 to 2009 who were not receiving hemodialysis at the time of admission. Included patients were stratified by treatment with vancomycin, vancomycin/piperacillin-tazobactam, or vancomycin/an alternative gram-negative rod (GNR) antibiotic. p values for categorical variables were computed using χ2 while continuous variables were computed using Kruskal-Wallis. Variables deemed statistically significant (< 0.05) were included in the multivariable, log-binomial regression model. Relative risk (RR) and 95% confidence intervals (CI), and p values were computed using a generalized estimating equation (GEE) approach with robust standard errors (i.e., Huber White “sandwich variance” estimates) to accommodate a correlated data structure corresponding to multiple episodes of infection per individual.
Results: A total of 530 patients with 1,007 episodes of infection, were treated with vancomycin (150 patients/302 episodes of infection), vancomycin/piperacillin-tazobactam (213 patients/372 episodes of infection), or vancomycin/GNR alternative (167 patients/333 episodes of infection). Patient demographics, comorbidities, sites of infection, and organisms of infection were compared among groups. After adjusting for statistically significant variables, neither vancomycin/piperacillin-tazobactam (RR = 1.1, 95% CI = 0.99–1.2; p = 0.073) nor vancomycin/GNR alternative (RR = 1.1, 95% CI = 0.98–1.2; p = 0.097) were found to be associated with an increased risk for nephrotoxicity compared with vancomycin alone.
Conclusion: A difference in nephrotoxicity was not observed between vancomycin and vancomycin/piperacillin-tazobactam–treated patients at our institution. Concomitant use as empiric therapy is appropriate, although larger sample sizes are needed to analyze closely this relation among at-risk subsets of this population.
Broad-spectrum antimicrobial therapy administered within the first hour of sepsis is one of the many recommendations highlighted by the Surviving Sepsis Campaign for the treatment of sepsis [1]. This spectrum of activity should take into consideration patient history, comorbidities, and drug allergies and local pathogens, but also cover the most common pathogens responsible for sepsis in hospitalized patients (i.e., gram-positive and gram-negative bacteria). Given the increase of multi-drug–resistant pathogens (e.g., methicillin-resistant Staphylococcus aureus [MRSA], vancomycin-intermediate S. aureus [VISA], and Pseudomonas aeruginosa), vancomycin and piperacillin-tazobactam are used commonly as empiric treatment agents [1–5].
Previous studies reporting an association between vancomycin and nephrotoxicity have been conflicting [6–20]. By contrast, piperacillin-tazobactam has rarely been associated with nephrotoxicity [21]. More recently, an increased incidence of nephrotoxicity has been associated with the combined use of vancomycin and piperacillin-tazobactam compared with vancomycin alone [21,22]. However, these studies are few and limited by small sample size.
Thus, the purpose of this study was to compare the prevalence of nephrotoxicity after treatment with vancomycin alone, and concomitant vancomycin and piperacillin-tazobactam treatment at our institution. We hypothesized that concomitant vancomycin and piperacillin-tazobactam–treated patients will experience a greater prevalence of nephrotoxicity compared with vancomycin-only treated patients or patients treated with vancomycin and an alternative gram-negative agent.
Patients and Methods
Study design
Institutional Review Board approval was obtained prior to data analysis. Details of the database and methodology have been described previously and are summarized below [20]. This was a retrospective analysis of a prospectively maintained database of all adult surgical patients (i.e., general, abdominal organ transplant, and trauma) admitted to a level I trauma and tertiary care center from 1996 to 2012 and treated for sepsis. We queried the database for all patients treated with vancomycin from December 6, 2005 to December 4, 2009; data collection was complete for this time period. Patients receiving hemodialysis prior to vancomycin initiation were excluded. Included patients were then stratified by treatment with vancomycin alone, vancomycin/piperacillin-tazobactam, or vancomycin/an alternative gram-negative rod (GNR) antibiotic (i.e., ceftriaxone, cefepime, ceftazidime, ticarcillin-clavulanate, imipenem, meropenem, gentamicin, tobramycin, amikacin, aztreonam, ciprofloxacin, or levofloxacin). Demographics and comorbidities, sites of infection, organisms of infection, and outcomes were compared between groups.
Database
Patient data were collected prospectively every other day by electronic medical chart, patient interview, and examination. Unique episodes of infection were identified for each patient, and classified as separate if positive cultures were present more than 72 h apart. Site(s) of infection, antibiotic therapy and duration, and culture-determined organisms were recorded for each episode of infection.
Patients
Gender, age, race (patient-defined), solid organ transplant, trauma, obesity (body mass index [BMI] ≥30), diabetes mellitus, hypertension, hyperlipidemia, cardiovascular disease, peripheral vascular disease, respiratory disease, ventilator dependence, renal insufficiency), initial creatinine, initial estimated glomerular filtration rate (eGFR), hepatic insufficiency, malignant disease, chronic steroid use, prior transfusion during same hospitalization, nosocomial infection, patient location at time of infection, and Acute Physiology and Chronic Health Evaluation II (APACHE II) score were evaluated at time of initial infectious episode. New onset or change in ventilator dependence, renal insufficiency, hyperlipidemia, transfusion, patient location, and APACHE II score were measured at the time of each subsequent infectious episode. Similarly, sites of infection and organisms cultured were measured at time of each infectious episode. Outcomes evaluated included total number of infectious episodes treated per group, total number of antibiotics used per infectious episode and stratified by group, hospital length of stay, duration of vancomycin treatment per infectious episode, maximum vancomycin trough per infectious episode, maximum creatinine during vancomycin treatment per infectious episode, final creatinine after vancomycin treatment per infectious episode, change in creatinine during vancomycin treatment (i.e., maximum creatinine-initial creatinine) and after vancomycin treatment (i.e., final creatinine-initial creatinine), lowest eGFR during vancomycin treatment per infectious episode, final eGFR after vancomycin treatment per infectious episode, nephrotoxicity per infectious episode, new onset hemodialysis during vancomycin treatment, and death. Patients were followed until death or hospital discharge.
Setting
As previously stated, the University of Virginia Health System is a level I trauma and tertiary care center. Our pharmacy department doses ancomycin based on total body weight (15 mg/kg every 12 h for normal renal function) and estimated renal function. However, the total initial empiric dose is capped at 1,750–2,000 mg per dose. A trough is measured prior to the fourth dose and the regimen is adjusted to achieve a trough of 10–20 mcg/mL (10–15 mcg/mL for less severe infections and 15–20 mcg/mL for MRSA, osteomyelitis, endocarditis, meningitis, sepsis, etc.).
Our pharmacy department doses piperacillin-tazobactam based on renal function and severity of infection. For normal renal function (creatinine clearance >40 mL/min), standard dosing is 3.375 g intravenously every 6 h infused over 30 min for uncomplicated infections or 4.5 g intravenously every 6 h to 8 h (maximum: 18 g daily) for complicated infections (i.e., nosocomial pneumonia, or intra-abdominal or surgical site infection with P. aeruginosa). For patients with creatinine clearances from 20–40 mL/min, dosing is 2.25 g every 6 h (3.375 g every 6 h for nosocomial pneumonia). For patients with creatinine clearances <20 mL/min, dosing is 2.25 g every 8 h (2.25 g every 6 h for nosocomial pneumonia).
Definitions
Infection criteria were defined according to the U.S. Centers for Disease Control and Prevention (CDC) [23]. Solid organ transplant was defined as kidney, liver, pancreas, heart, lung, kidney/pancreas, liver/pancreas, and small bowel. Pulmonary disease was defined as active treatment for lung disease prior to hospital admission. Renal insufficiency was defined as a serum creatinine ≥2.0 mg/dL at time of infection episode but without need for dialysis. Patient location at time of infection episode was defined as home, hospital ward, intensive care unit (ICU), or other. Initial creatinine and eGFR were recorded immediately prior to vancomycin initiation per infectious episode. Other comorbidities were defined by chart documentation, medication history, and/or patient examination. Vancomycin trough levels were defined as the greatest trough level documented in the patient's chart during the treatment period. Nephrotoxicity was defined as a 0.5 mg/dL increase in serum creatinine, a 50% increase in baseline serum creatinine, or a 50% decrease in baseline creatinine clearance [9,10,13,24]. End points of antibiotic therapy included culture-proven resistance or insensitivity, adequate treatment duration as defined by unit guidelines and specific to site of infection (i.e., lung [8–14 d], abdomen [5–7 d], vascular catheter [5 d], urine [5–7 d], skin and skin structure [until resolution of surrounding cellulitis], and blood stream [10–14 d]), and/or death. Uncomplicated blood stream infections involving S. aureus and a treatable focus were treated for 14 d after culture-proven clearance. Complicated blood stream infections involving S. aureus (e.g., infected endovascular grafts, endocardial vegetations) were treated for 4 wks to 6 wks after culture-proven clearance. Mortality was defined as any cause of death after infection diagnosis while hospitalized.
Statistical analysis
Demographics and comorbidities, infection sites, pathogenic organisms, and outcomes were compared between patients treated with vancomycin alone, vancomycin/piperacillin-tazobactam, or vancomycin/alternative GNR antibiotic. p values for categorical variables were computed using χ2 whereas continuous variables were computed using Kruskal-Wallis. Additionally, relative risk (RR) and 95% confidence intervals (CI), and p values were computed predicting nephrotoxicity using a generalized estimating equation (GEE) approach with robust standard errors (i.e., Huber White “sandwich variance” estimates) to accommodate a correlated data structure corresponding to multiple episodes of infection per individual [25,26]. Variables deemed statistically significant for either comparison between groups or nephrotoxicity association, among the demographics and comorbidities, infection sites, and infection-related organisms, were included in the multivariable, log-binomial regression model. Analysis was performed using SAS Version 9.3 (SAS Institute, Cary, NC) programming software. Statistical significance was defined as p < 0.05.
Results
A total of 530 patients with 1,007 episodes of infection were treated with vancomycin (150 patients/302 episodes of infection), vancomycin/piperacillin-tazobactam (213 patients/372 episodes of infection), or vancomycin/alternative GNR antibiotic (167 patients/333 episodes of infection). On average, each patient treated with vancomycin experienced 1.8 ± 1.2 episodes of infection, each patient treated with vancomycin/piperacillin-tazobactam experienced 1.7 ± 1.2 episodes of infection, and each patient treated with vancomycin/alternative GNR antibiotic experienced 1.9 ± 1.2 episodes of infection (p = 0.058). Additionally, each patient treated with vancomycin was treated on average with 3.9 ± 1.6 antibiotics per infectious episode, each patient treated with vancomycin/piperacillin-tazobactam was treated with 3.2 ± 1.1 antibiotics per infectious episode, and each patient treated with vancomycin/alternative GNR antibiotic was treated with 3.8 ± 1.4 antibiotics per infectious episode (p < 0.0001).
Demographics and comorbidities stratified by treatment arm (i.e., vancomycin, vancomycin/piperacillin-tazobactam, or vancomycin/alternative GNR antibiotic) are listed in Table 1. Additionally, each variable's association with nephrotoxicity has been provided. Females, age, trauma, cardiovascular disease, ventilator dependence, renal insufficiency, initial creatinine, prior blood transfusion during the same hospitalization, nosocomial infection, patient location at time of infection, and APACHE II score were different between groups. Additionally, APACHE II score was associated with nephrotoxicity.
Table 1.
Demographics and Comorbiditiesa Stratified by Vancomycin versus Vanccomycin/Gram-Negative Rod Antibiotic (Excluding Piperacillin-Tazobactam) versus Vancomycin/Piperacillin-Tazobactam
| Demographics/comorbidities | Vanc n (%) | Vanc/GNR n (%) | Vanc/P-T n (%) | pb | RR (95% CI); pc(nephrotoxicity) |
|---|---|---|---|---|---|
| Number of patients | 150 (28) | 167 (32) | 213 (40) | — | — |
| Number of infectious episodes | 302 (30) | 333 (33) | 372 (37) | — | — |
| Gender | |||||
| Female | 93 (31) | 147 (44) | 138 (37) | 0.0024 | 0.98 (0.93–1.03); 0.36 |
| Age (years) | |||||
| Mean ± SD | 58 ± 15 | 55± 17 | 53 ± 15 | 0.0005 | |
| Median (IQR) | 58 (22) | 53 (22) | 53 (20) | ||
| Q1 (≤45) | 60 (20) | 101 (30) | 98 (26) | 0.0008 | 1.0 ref |
| Q2 (46–55) | 72 (24) | 79 (24) | 108 (29) | 1.03 (0.95–1.1); 0.52 | |
| Q3 (56–67) | 76 (25) | 73 (22) | 99 (27) | 1.04 (0.96–1.1); 0.34 | |
| Q4 (>67) | 94 (31) | 80 (24) | 67 (18) | 1.01 (0.93–1.1); 0.83 | |
| Race | |||||
| White | 265 (88) | 289 (87) | 316 (85) | — | 1.0 ref |
| Black | 35 (12) | 36 (11) | 50 (13) | 0.95 (0.85–1.1); 0.36 | |
| Other | 2 (1) | 2 (1) | 2 (1) | 1.2 (0.92–1.6); 0.18 | |
| Hispanic | 0 (0) | 6 (2) | 4 (1) | 1.1 (0.91–1.4); 0.27 | |
| Transplantd | |||||
| Yes | 58 (19) | 67 (20) | 64 (17) | 0.60 | 1.1 (0.98–1.1); 0.13 |
| Trauma | |||||
| Yes | 31 (12) | 64 (20) | 35 (10) | 0.0010 | 0.94 (0.85–1.03); 0.20 |
| Obesity | |||||
| BMI ≥30 | 118 (39) | 129 (39) | 142 (38) | 0.97 | 0.99 (0.93–1.04); 0.62 |
| Diabetes mellitus | |||||
| Yes | 68 (23) | 74 (22) | 90 (24) | 0.80 | 1.04 (0.93–1.2); 0.49 |
| Hypertension | |||||
| Yes | 114 (38) | 129 (39) | 145 (39) | 0.94 | 1.1 (0.99–1.1); 0.13 |
| Hyperlipidemia | |||||
| Yes | 48 (16) | 46 (14) | 50 (13) | 0.63 | 1.03 (0.94–1.1); 0.55 |
| Cardiovascular disease | |||||
| Yes | 101 (33) | 75 (23) | 78 (21) | 0.0004 | 1.001 (0.94–1.1); 0.98 |
| PVD | |||||
| Yes | 14 (5) | 20 (6) | 12 (3) | 0.21 | 1.1 (0.91–1.2); 0.51 |
| Pulmonary disease | |||||
| Yes | 41 (14) | 45 (14) | 61 (16) | 0.46 | 0.98 (0.90–1.1); 0.72 |
| Ventilator dependence | |||||
| Yes | 86 (28) | 105 (32) | 54 (15) | <0.0001 | 0.95 (0.86–1.05); 0.32 |
| Renal insufficiency | |||||
| Yes | 34 (11) | 18 (5) | 17 (5) | 0.0013 | 0.96 (0.82–1.1); 0.61 |
| Initial creatinine | |||||
| Mean ± SD | 1.3 ± 0.90 | 1.3 ± 0.95 | 1.3 ± 1.02 | 0.74 | |
| Median (IQR) | 0.95 (0.70) | 1.0 (0.70) | 1.0 (0.60) | ||
| (≤1.0) | 173 (57) | 174 (52) | 199 (53) | 0.0013 | 1.0 ref |
| (1.1–1.5) | 54 (18) | 77 (23) | 93 (25) | 0.99 (0.89–1.1); 0.78 | |
| (1.6–2.0) | 19 (6) | 46 (14) | 35 (9) | 0.998 (0.94–1.2); 0.98 | |
| (>2.0) | 56 (19) | 36 (11) | 45 (12) | 1.02 (0.89–1.2); 0.79 | |
| Initial eGFR | |||||
| (>60) | 190 (63) | 195 (59) | 239 (64) | 0.0503 | 1.0 ref |
| (31–60) | 72 (24) | 90 (27) | 87 (23) | 0.97 (0.88–1.1); 0.61 | |
| (15–30) | 37 (12) | 35 (11) | 29 (8) | 1.1 (0.94–1.2); 0.30 | |
| (<15) | 3 (1) | 13 (4) | 17 (5) | 0.93 (0.72–1.2); 0.59 | |
| Hepatic insufficiency | |||||
| Yes | 18 (6) | 13 (4) | 24 (6) | 0.30 | 1.1 (0.88–1.3); 0.44 |
| Malignant disease | |||||
| Yes | 43 (14) | 45 (14) | 36 (10) | 0.14 | 0.93 (0.80–1.1); 0.37 |
| Long-term steroid use | |||||
| Yes | 71 (24) | 83 (25) | 72 (19) | 0.18 | 1.1 (0.99–1.2); 0.092 |
| Prior transfusion | |||||
| Yes | 137 (45) | 148 (44) | 105 (28) | <0.0001 | 0.99 (0.94–1.1); 0.84 |
| Nosocomial infection | |||||
| Yes | 228 (76) | 252 (76) | 217 (58) | <0.0001 | 1.01 (0.92–1.1); 0.91 |
| Patient location | |||||
| Home | 72 (24) | 81 (24) | 155 (42) | <0.0001 | 1.0 ref |
| Hospital ward | 91 (30) | 78 (23) | 107 (29) | 1.003 (0.90–1.1); 0.96 | |
| ICU | 116 (38) | 143 (43) | 92 (25) | 0.999 (0.91–1.1); 0.99 | |
| Other | 23 (8) | 31 (9) | 18 (5) | 1.04 (0.92–1.2); 0.54 | |
| APACHE II Score | |||||
| Mean ± SD | 17 ± 8 | 18 ± 8 | 15 ± 7 | <0.0001 | |
| Median (IQR) | 17 (12) | 18 (10) | 15 (10) | ||
| Q1 (≤11) | 77 (26) | 67 (20) | 138 (37) | <0.0001 | 1.0 ref |
| Q2 (12–16) | 66 (22) | 74 (22) | 88 (24) | 0.96 (0.85–1.1); 0.50 | |
| Q3 (17–21) | 71 (24) | 92 (28) | 86 (23) | 0.96 (0.86–1.1); 0.50 | |
| Q4 (>21) | 88 (29) | 100 (30) | 60 (16) | 1.1 (1.006–1.2); 0.039 | |
Variables analyzed per infectious episode.
Chi square and Kruskal-Wallis testing were used to compare demographics and comorbidities between groups.
To accommodate for a correlated data structure corresponding to multiple episodes of infection per individual, the analysis of nephrotoxicity among episodes of infection was computed using a generalized estimating equation (GEE) approach with robust standard errors (i.e., Huber-White “sandwich variance” estimates).
Transplants included: 24 kidney, 58 liver, two pancreas, nine kidney/pancreas, two liver/kidney, and one small bowel.
Vanc = vancomycin; GNR = gram-negative rod; P-T = piperacillin-tazobactam; RR = relative risk; CI = confidence interval; SD = standard deviation; IQR = interquartile range; Q1 = first quartile.; Q2 = second quartile; Q3 = third quartile; Q4 = fourth quartile; BMI = body mass index; PVD = peripheral vascular disease; eGFR = estimated glomerular filtration rate; ICU = intensive care unit; APACHE II = Acute Physiology and Chronic Health Evaluation II.
Sites of infection stratified by treatment arm are listed in Table 2. Additionally, each variable's association with nephrotoxicity has been provided. Infections related to the peritoneum, colon, skin/soft tissue, wound, and blood stream were different between groups.
Table 2.
Sites of Infectiona Stratified by Vancomycin versus Vancomycin/Gram-Negative Rod Antibiotic (Excluding Piperacillin-Tazobactam) versus Vancomycin/Piperacillin-Tazobactam
| Sites | Vanc n (%) | Vanc/GNR n (%) | Vanc/P-T n (%) | pb | RR (95% CI); pc(nephrotoxicity) |
|---|---|---|---|---|---|
| Number of patients | 150 (28) | 167 (32) | 213 (40) | — | — |
| Number of infectious episodes | 302 (30) | 333 (33) | 372 (37) | — | — |
| CNS | |||||
| Yes | 0 (0) | 1 (0) | 0 (0) | — | — |
| Peritoneum | |||||
| Yes | 53 (18) | 78 (23) | 107 (29) | 0.0030 | 0.996 (0.95–1.04); 0.87 |
| Upper GI | |||||
| Yes | 1 (0) | 1 (0) | 2 (1) | 0.86 | 0.93 (0.77–1.1); 0.42 |
| Colon | |||||
| Yes | 31 (10) | 12 (4) | 6 (2) | <0.0001 | 1.1 (0.998–1.2); 0.058 |
| Lung | |||||
| Yes | 76 (25) | 80 (24) | 68 (18) | 0.065 | 0.997 (0.96–1.04); 0.89 |
| Pleura | |||||
| Yes | 2 (1) | 4 (1) | 3 (1) | 0.75 | 0.90 (0.73–1.1); 0.30 |
| Skin/soft tissue | |||||
| Yes | 16 (5) | 19 (6) | 36 (10) | 0.044 | 0.96 (0.89–1.03); 0.25 |
| Wound | |||||
| Yes | 18 (6) | 38 (11) | 54 (15) | 0.0018 | 0.9997 (0.95–1.1); 0.99 |
| Line | |||||
| Yes | 11 (4) | 10 (3) | 11 (3) | 0.86 | 1.03 (0.95–1.1); 0.47 |
| Blood | |||||
| Yes | 59 (20) | 52 (16) | 46 (12) | 0.039 | 0.98 (0.94–1.009); 0.16 |
| Urine | |||||
| Yes | 34 (11) | 36 (11) | 37 (10) | 0.85 | 1.02 (0.99–1.1); 0.23 |
Variables analyzed per infectious episode.
Chi square was used to compare sites of infection between groups.
To accommodate for a correlated data structure corresponding to multiple episodes of infection per individual, the analysis of nephrotoxicity among episodes of infection was computed using a generalized estimating equation (GEE) approach with robust standard errors (i.e., Huber-White “sandwich variance” estimates).
Vanc = vancomycin; GNR = gram-negative rod; P-T = piperacillin-tazobactam; RR = relative risk; CI = confidence interval; CNS = central nervous system; GI = gastrointestinal.
Culture-identified organisms stratified by treatment arm are listed in Table 3. Additionally, each variable's association with nephrotoxicity has been provided. Gram-negative bacteria in general, Serratia spp., P. aeruginosa, and Enterobacter cloacae were significantly different between groups. Additionally, P. aeruginosa and vancomycin-resistant Enterococcus were associated with nephrotoxicity.
Table 3.
Organismsa Stratified by Vancomycin versus Vancomycin/Gram-Negative Rod Antibiotic (Excluding Piperacillin-Tazobactam) versus Vancomycin/Piperacillin-Tazobactam
| Organism | Vanc n (%) | Vanc/GNR n (%) | Vanc/P-T n (%) | pb | RR (95% CI); pc(nephrotoxicity) |
|---|---|---|---|---|---|
| Number of patients | 150 (28) | 167 (32) | 213 (40) | — | — |
| Number of infectious episodes | 302 (30) | 333 (33) | 372 (37) | — | — |
| Fungal | 46 (15) | 51 (15) | 53 (14) | 0.91 | 0.98 (0.92–1.05); 0.56 |
| Candida albicans | 15 (5) | 21 (6) | 25 (7) | 0.62 | 0.94 (0.86–1.03); 0.19 |
| Candida glabrata | 18 (6) | 14 (4) | 20 (5) | 0.59 | 1.04 (0.95–1.1); 0.42 |
| Gram-negative bacteria | 142 (47) | 117 (35) | 104 (28) | <0.0001 | 0.99 (0.94–1.05); 0.81 |
| Escherichia coli | 32 (11) | 33 (10) | 33 (9) | 0.75 | 0.94 (0.88–1.01); 0.076 |
| Klebsiella pneumonia | 26 (9) | 18 (5) | 22 (6) | 0.22 | 1.01 (0.90–1.1); 0.92 |
| Serratia spp. | 10 (3) | 2 (1) | 3 (1) | 0.0074 | 1.02 (0.72–1.4); 0.92 |
| Pseudomonas aeruginosa | 34 (11) | 11 (3) | 10 (3) | <0.0001 | 0.92 (0.86–0.98); 0.0060 |
| Enterobacter cloacae | 21 (7) | 12 (4) | 8 (2) | 0.0063 | 1.03 (0.96–1.1); 0.40 |
| Gram-positive bacteria | 112 (37) | 123 (37) | 158 (42) | 0.23 | 0.97 (0.93–1.02); 0.22 |
| MSSA | 19 (6) | 22 (7) | 20 (5) | 0.78 | 0.98 (0.93–1.02); 0.33 |
| MRSA | 23 (8) | 28 (8) | 28 (8) | 0.90 | 1.003 (0.92–1.1); 0.95 |
| CNS | 2 (1) | 1 (0) | 2 (1) | 0.80 | 1.1 (0.88–1.5); 0.32 |
| Enterococcus faecalis | 11 (4) | 23 (7) | 20 (5) | 0.19 | 0.9994 (0.93–1.1); 0.99 |
| Enterococcus faecium | 6 (2) | 5 (2) | 5 (1) | 0.79 | 1.1 (0.93–1.3); 0.29 |
| VRE | 12 (4) | 16 (5) | 28 (8) | 0.10 | 0.90 (0.82–0.98); 0.015 |
| Streptococcus spp. | 15 (5) | 13 (4) | 28 (8) | 0.096 | 1.1 (0.95–1.2); 0.30 |
| Anaerobic bacteria | 38 (13) | 35 (11) | 31 (8) | 0.20 | 1.03 (0.97–1.1); 0.33 |
Variables analyzed per infectious episode.
Chi square was used to compare organisms between groups.
To accommodate for a correlated data structure corresponding to multiple episodes of infection per individual, the analysis of nephrotoxicity among episodes of infection was computed using a generalized estimating equation (GEE) approach with robust standard errors (i.e. Huber-White “sandwich variance” estimates).
Vanc = vancomycin; GNR = gram-negative rod; P-T = piperacillin-tazobactam; RR = relative risk; CI = confidence interval; MSSA = methicillin-sensitive Staphylococcus aureus; MRSA = methicillin-resistant Staphylococcus aureus; CNS = coagulase negative staphylococcus; VRE = vancomycin-resistant enterococcus.
Patient outcomes stratified by treatment arm are listed in Table 4. Patients were treated with vancomycin for a similar duration across all groups. Although differences regarding eGFR were present during treatment with vancomycin across all groups, nephrotoxicity and new onset hemodialysis did not differ. Interestingly, hospital length of stay and incidence of death was longer and greater, respectively, within groups treated with vancomycin alone and vancomycin/alternative GNR antibiotic compared with the group treated with vancomycin/piperacillin-tazobactam.
Table 4.
Patient Outcomes after Vancomycin Treatment Stratified by Vancomycin versus Vancomycin/Gram-Negative Rod Antibiotic (Excluding Piperacillin-Tazobactam) versus Vancomycin/Piperacillin-Tazobactam
| Outcome | Vanc n (%) | Vanc/GNR n (%) | Vanc/P-T n (%) | pa |
|---|---|---|---|---|
| Number of patients | 150 (28) | 167 (32) | 213 (40) | — |
| Number of infectious episodes | 302 (30) | 333 (33) | 372 (37) | |
| HLOS (days)b | ||||
| Mean ± SD | 25 ± 27 | 27 ± 35 | 21 ± 26 | 0.0015 |
| Median (IQR) | 15 (28) | 18 (22) | 12 (18) | |
| Duration of vancomycin (days)c | ||||
| Mean ± SD | 8.9 ± 8.8 | 9.3 ± 9.8 | 8.3 ± 8.04 | 0.23 |
| Median (IQR) | 6 (6) | 7 (6) | 6 (5) | |
| Vancomycin troughb | ||||
| Mean ± SD | 15 ± 8.6 | 16 ± 11 | 17 ± 12 | 0.49 |
| Median (IQR) | 15 (13) | 13 (12) | 15 (15) | |
| (>20) | 22 (15) | 28 (18) | 49 (24) | 0.0085 |
| (16–20) | 16 (11) | 12 (8) | 13 (6) | |
| (11–15) | 8 (6) | 23 (15) | 12 (6) | |
| (≤10) | 97 (68) | 91 (59) | 129 (64) | |
| Maximum creatinine during Txc | ||||
| Mean ± SD | 1.6 ± 1.2 | 1.5 ± 1.1 | 1.6 ± 1.2 | 0.98 |
| Median (IQR) | 1.2 (1.3) | 1.2 (0.9) | 1.2 (0.9) | |
| Final creatinine after Txc | ||||
| Mean ± SD | 1.2 ± 0.83 | 1.2 ± 1.04 | 1.2 ± 0.77 | 0.048 |
| Median (IQR) | 1.0 (0.70) | 0.90 (0.60) | 1.0 (0.50) | |
| Change in creatinine (maximum-initial)c | ||||
| Mean ± SD | 0.34 ± 0.60 | 0.27 ± 0.52 | 0.29 ± 0.52 | 0.57 |
| Median (IQR) | 0.10 (0.40) | 0.10 (0.30) | 0.10 (0.30) | |
| Change in creatinine (final-initial)c | ||||
| Mean ± SD | −0.075 ± 0.58 | −0.072 ± 0.58 | −0.10 ± 0.76 | 0.055 |
| Median (IQR) | 0.0 (0.30) | −0.10 (0.40) | 0.0 (0.30) | |
| Lowest eGFR during Txc | ||||
| (>60) | 147 (49) | 160 (48) | 193 (52) | 0.012 |
| (31–60) | 84 (28) | 92 (28) | 116 (31) | |
| (15–30) | 60 (20) | 60 (18) | 38 (10) | |
| (<15) | 11 (4) | 21 (6) | 25 (7) | |
| Final eGFR after Txc | ||||
| (>60) | 204 (68) | 216 (65) | 249 (67) | 0.0076 |
| (31–60) | 66 (22) | 66 (20) | 93 (25) | |
| (15–30) | 28 (9) | 32 (10) | 22 (6) | |
| (<15) | 4 (1) | 19 (6) | 8 (2) | |
| Nephrotoxicityc | ||||
| Yes | 66 (22) | 68 (20) | 76 (21) | 0.89 |
| New-onset hemodialysisb | ||||
| Yes | 13 (9) | 19 (11) | 14 (7) | 0.25 |
| Deathb | ||||
| Yes | 21 (14) | 23 (14) | 14 (7) | 0.030 |
Chi square and Kruskal-Wallis testing were used to compare patient outcomes between groups.
Variables analyzed per patient.
Variables analyzed per infectious episode.
Vanc = vancomycin; GNR = gram-negative rod; P-T = piperacillin-tazobactam; HLOS = hospital length of stay; SD = standard deviation; IQR = interquartile range; Tx = treatment; eGFR = estimated glomerular filtration rate.
Variables found to be different between groups and/or those found to be associated with nephrotoxicity (i.e., gender, age, trauma, cardiovascular disease, ventilator dependence, initial creatinine, prior transfusion, nosocomial infection, APACHE II score, peritoneal infections, colon infections, skin infections, surgical site infections, blood stream infections, Seratia spp. infections, P. aeruginosa infections, E. cloacae infections, and vancomycin-resistant enterococcus) were included in the multivariable, log-binomial regression model listed in Table 5. APACHE II score >21 (RR = 1.1, 95% CI = 1.01–1.3; p = 0.032) was found to be associated independently with increased nephrotoxic risk. Neither patients treated with vancomycin and piperacillin-tazobactam (RR = 1.1, 95% CI = 0.99–1.2; p = 0.073) nor patients treated with vancomycin and alternative GNR antibiotic (RR = 1.1, 95% CI = 0.98–1.2; p = 0.097) experienced an increase in nephrotoxic risk compared with patients treated with vancomycin alone.
Table 5.
Multivariable Analysis Predicting Nephrotoxicitya
| Variables | RR (95% CI); pb |
|---|---|
| Vanc | 1.0 ref |
| Vanc/GNR | 1.1 (0.98–1.2); 0.097 |
| Vanc/P-T | 1.1 (0.99–1.2); 0.073 |
| Gender | |
| Male | 1.0 ref |
| Female | 0.97 (0.92–1.04); 0.41 |
| Age | |
| Q1 (≤45) | 1.0 ref |
| Q2 (46–55) | 1.02 (0.91–1.1); 0.75 |
| Q3 (56–67) | 1.02 (0.93–1.1); 0.66 |
| Q4 (>67) | 0.98 (0.89–1.1); 0.67 |
| Trauma | |
| No | 1.0 ref |
| Yes | 0.94 (0.80–1.1); 0.48 |
| Cardiovascular disease | |
| No | 1.0 ref |
| Yes | 0.99 (0.92–1.1); 0.71 |
| Ventilator dependence | |
| No | 1.0 ref |
| Yes | 0.90 (0.78–1.03); 0.12 |
| Initial creatinine | |
| (≤1.0) | 1.0 ref |
| (1.1–1.5) | 0.93 (0.83–1.04); 0.202 |
| (1.6–2.0) | 0.91 (0.76–1.1); 0.31 |
| (>2.0) | 0.92 (0.79–1.1); 0.31 |
| Prior transfusion | |
| No | 1.0 ref |
| Yes | 1.04 (0.96–1.1); 0.31 |
| Nosocomial infection | |
| No | 1.0 ref |
| Yes | 1.003 (0.91–1.1); 0.94 |
| APACHE II score | |
| Q1 (≤11) | 1.0 ref |
| Q2 (12–16) | 0.97 (0.86–1.1); 0.603 |
| Q3 (17–21) | 0.98 (0.86–1.1); 0.76 |
| Q4 (>21) | 1.2 (1.01–1.3); 0.032 |
| Peritoneum | |
| No | 1.0 ref |
| Yes | 0.96 (0.91–1.02); 0.20 |
| Colon | |
| No | 1.0 ref |
| Yes | 1.1 (1.001–1.2); 0.047 |
| Skin | |
| No | 1.0 ref |
| Yes | 0.93 (0.86–1.001); 0.053 |
| Wound | |
| No | 1.0 ref |
| Yes | 0.98 (0.93–1.03); 0.39 |
| Blood | |
| No | 1.0 ref |
| Yes | 0.97 (0.93–1.008); 0.11 |
| Serratia spp. | |
| No | 1.0 ref |
| Yes | 1.1 (0.80–1.6); 0.49 |
| Pseudomonas aeruginosa | |
| No | 1.0 ref |
| Yes | 0.96 (0.90–1.02); 0.19 |
| Enterobacter cloacae | |
| No | 1.0 ref |
| Yes | 1.06 (0.96–1.2); 0.27 |
| VRE | |
| No | 1.0 ref |
| Yes | 0.87 (0.80–0.96); 0.0033 |
Variables analyzed per infectious episode.
To accommodate for a correlated data structure corresponding to multiple episodes of infection per individual, the analysis of nephrotoxicity among episodes of infection was computed using a generalized estimating equation (GEE) approach with robust standard errors (i.e., Huber-White “sandwich variance” estimates).
RR = relative risk; CI = confidence interval; Vanc = vancomycin; GNR = gram-negative rod; P-T = piperacillin-tazobactam; Q1 = first quartile; Q2 = second quartile; Q3 = third quartile; Q4 = fourth quartile; APACHE II = Acute Physiology and Chronic Health Evaluation II; VRE = vancomycin-resistant enterococcus.
Discussion
We present one of the largest cohorts evaluating concomitant vancomycin/piperacillin-tazobactam therapy and nephrotoxicity published to date [21,22,27]. Our study observed a 21% incidence of nephrotoxicity among all infectious episodes of adult surgical patients admitted from December 6, 2005 to December 4, 2009 and treated for sepsis at our institution using either vancomycin alone (22%), vancomycin/alternative GNR antibiotic (20%), or vancomycin/piperacillin-tazobactam (21%). Additionally, we observed a 12% incidence of new-onset hemodialysis; vancomycin alone (9%), vancomycin/alternative GNR antibiotic (11%), or vancomycin/piperacillin-tazobactam (7%). Using multivariable analysis, treatment with vancomycin/alternative GNR antibiotic and vancomycin/piperacillin-tazobactam were not observed to be associated with nephrotoxicity compared with treatment with vancomycin alone.
Vancomycin is a bactericidal glycopeptide [28]. Its mechanism of action involves the inhibition of peptidoglycan and cell wall growth of gram-positive organisms. Elimination is primarily by the kidneys and vancomycin-associated nephrotoxicity is theorized to occur at the proximal renal tubule as a result of oxidative stress and free radical generation [9,13]. To minimize the incidence of vancomycin-associated nephrotoxicity, current practice guidelines recommend treatment dosages of 15–20 mg/kg every eight to 12 h in patients with normal renal function [29]. This serves to maximize the frequency of troughs between 10–20 mg/L. A previous study of ours comparing incidence of nephrotoxicity between septic, critically ill patients treated with either vancomycin (298 patients/571 episodes of infection) or linezolid (247 patients/475 episodes of infection) observed a similar nephrotoxic profile when treated within the 10–20 mg/L range [20].
Piperacillin-tazobactam is a bactericidal, β-lactam combined with a β-lactamase inhibitor [30]. Its mechanism of action involves the inhibition of cell wall growth of gram-postitive and gram-negative organisms. Bacteria may develop resistance to β-lactams by synthesizing β-lactamase, which serves to decouple the β-lactam ring structure. However, the combined use of a β-lactamase inhibitor competitively binds the β-lactamase, thus preventing this decoupling action. Elimination is primarily by the kidneys and piperacillin-tazobactam-associated nephrotoxicity has rarely been cited in the literature (<1%) [21,31,32]. Case reports of this are largely based on temporal dosing, and a definitive association has not been delineated. Additionally, all of these reports involve patients with oncologic diagnoses who were experienced dehydration, critical illness, sepsis, and/or other nephrotoxic exposure. Current practice guidelines recommend treatment dosages of 3.375 g every 6 h, 4.5 g every eight hours, or 4.5 g every 6 h infused over a 30-min duration [30].
As the incidence of multi-drug–resistant pathogens continues to increase, patients are managed commonly using combination therapy. One example of this is the treatment of MRSA and VISA using combined vancomycin/piperacillin-tazobactam. Dilworth et al. [27] conducted an in vitro study to evaluate whether vancomycin/piperacillin-tazobactam were synergistic against MRSA and VISA. The authors observed an increased antimicrobial activity using sub-inhibitory concentrations of vancomycin when combination therapy was used. However, this synergistic benefit may not be without cost.
Recent studies have begun to evaluate the incidence of nephrotoxicity associated with combination therapy compared with vancomycin treatment alone as this may impact empiric antimicrobial therapy; however, these are few and the results are conflicting [21,22,33]. Burgess et al. [21] conducted a single-center, retrospective cohort study comparing incidence of nephrotoxicity among patients treated with vancomycin for 48 h (n = 99, 8.1% nephrotoxicity incidence) and patients treated with vancomycin/piperacillin-tazobactam for 48 h (n = 92, 16.3% nephrotoxicity incidence) [21]. Using multivariable analysis, the authors observed a significantly greater association between combination therapy and nephrotoxicity (odds ratio = 2.48; p = 0.032) compared with monotherapy with vancomycin. Similarly, Gomes et al. [22] conducted a retrospective matched cohort study comparing nephrotoxicity incidence among patients treated for 48 h with vancomyin/piperacillin-tazobactam (n = 112, 34.8% nephrotoxicity incidence) to vancomycin/cefepime (n = 112, 12.5% nephrotoxicity incidence). Using multivariable analysis, the authors observed that combination therapy with vancomycin/piperacillin-tazobactam had a significantly increased association with nephrotoxicity compared with treatment with vancomycin/cefepime (odds ratio = 5.67, 95% CI = 1.66–19.33). By contrast, Moenster et al. [33] conducted a retrospective cohort study comparing incidence of nephrotoxicity among patients with diabetes mellitus with osteomyelitis treated for 72 h with either vancomycin/piperacillin-tazobactam (n = 109, 29.3% nephrotoxicity incidence) or vancomycin/cefepime (n = 30, 13.3% nephrotoxicity incidence). Using multivariable analysis, the authors did not observe an increased association between nephrotoxicity and vancomycin/piperacillin-tazobactam therapy compared with vancomycin/cefepime (odds ratio = 3.45, 95% CI = 0.96–12.4; p = 0.057). Rather, greater weight and vancomycin troughs were associated with a significantly increased risk of nephrotoxicity.
One explanation for a possible association between combination therapy and nephrotoxicity may be related to drug elimination inhibition. Jensen et al. [34] conducted a randomized control trial evaluating kidney failure in critically ill patients treated with broad-spectrum antibiotics. A multiple effects model comparing the eGFR of all patients showed that patients with existing renal dysfunction experienced delayed renal recovery when treated with piperacillin-tazobactam compared with other commonly administered antibiotics (i.e., meropenem, cefuroxime, ciprofloxacin, and vancomycin).
Interestingly, when piperacillin-tazobactam was discontinued, renal dysfunction improved more rapidly than patients treated with the other antibiotic protocols. This delay in recovery could negatively impact the elimination of vancomycin, thus resulting in greater than normal trough levels further compounding the acute kidney injury.
Strengths and Weaknesses
Our study is strengthened by its large sample size and use of multivariable analysis, however, it may be limited by its retrospective design. This may have contributed to both recall and selection bias. Whereas we did adjust for potential confounders, unforeseen confounders exist and may have impacted our results. For example, baseline creatinine and eGFR values prior to the onset of critical illness, use of nephrotoxic contrast dyes, or medications other than those mentioned previously, vancomycin-loading doses, vasopressor use, or fluid status was not captured within our database. Additionally, we did not evaluate the role of capping vancomycin doses on nephrotoxicity, or the impact of cumulative infectious exposure on nephrotoxicity. Patients were followed until death or hospital discharge, and thus, we were unable to define acute kidney injury using Risk Injury Failure Loss End (RIFLE) stage kidney disease criteria for each patient because of inadequate follow-up. Finally, external validity may be limited given that this was a single-center study of surgical patients with sepsis where a large percentage of infections are treated primarily by source control procedures verses antimicrobial agents alone.
Conclusion
No difference in nephrotoxicity was observed between vancomycin and vancomycin/piperacillin-tazobactam–treated patients at our institution. Concomitant use as empiric therapy is appropriate, although larger sample sizes are needed to analyze this relationship among at-risk subsets of this population.
Author Disclosure Statement
The authors declare no conflicts of interest.
References
- 1.Dellinger RP, Levy MM, Rhodes A, et al. Surviving sepsis campaign: International guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013;41:580–637 [DOI] [PubMed] [Google Scholar]
- 2.Marie JP, Marjanovic Z, Vekhoff A, et al. Piperacillin/tazobactam plus tobramycin versus ceftazidime plus tobramycin as empiric therapy for fever in severely neutropenic patients. Supportive Care Cancer 1999;7:89–94 [DOI] [PubMed] [Google Scholar]
- 3.Bassetti M, Righi E. Multidrug-resistant bacteria: What is the threat? Hematology 2013;2013:428–432 [DOI] [PubMed] [Google Scholar]
- 4.Vazquez-Guillamet C, Scolari M, Zilberberg MD, et al. Using the number needed to treat to assess appropriate antimicrobial therapy as a determinant of outcome in severe sepsis and septic shock. Critical Care Med 2014;42:2342–2349 [DOI] [PubMed] [Google Scholar]
- 5.Rodriguez-Bano J. Selection of empiric therapy in patients with catheter-related infections. Clin Microbiol Infect 2002;8:275–281 [DOI] [PubMed] [Google Scholar]
- 6.Williams ME, Harman C, Scheld M, et al. A controlled study of ticarcillin plus clavulanic acid versus piperacillin as empiric therapy for fever in the immunocompromised host. Am J Med 1985;79:67–72 [DOI] [PubMed] [Google Scholar]
- 7.Wang LS, Liu CY, Wang FD, et al. Chromatographically purified vancomycin: Therapy of serious infections caused by Staphylococcus aureus and other gram-positive bacteria. Clin Ther 1988;10:574–584 [PubMed] [Google Scholar]
- 8.Goren MP, Baker DK, Jr, Shenep JL. Vancomycin does not enhance amikacin-induced tubular nephrotoxicity in children. Pediatr Infect Dis J 1989;8:278–282 [PubMed] [Google Scholar]
- 9.Wong-Beringer A, Joo J, Tse E, Beringer P. Vancomycin-associated nephrotoxicity: A critical appraisal of risk with high-dose therapy. Int J Antimicrob Agents 2011;37:95–101 [DOI] [PubMed] [Google Scholar]
- 10.Hazlewood KA, Brouse SD, Pitcher WD, Hall RG. Vancomycin-associated nephrotoxicity: Grave concern or death by character assassination? Am J Med 2010;123:182.e1–7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Menichetti F, Martino P, Bucaneve G, et al. Effects of teicoplanin and those of vancomycin in initial empirical antibiotic regimen for febrile, neutropenic patients with hematologic malignancies. Gimema Infection Program. Antimicrob Agents Chemother 1994;38:2041–2046 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pletz MW, Burkhardt O, Welte T. Nosocomial methicillin-resistant Staphylococcus aureus (MRSA) pneumonia: Linezolid or vancomycin? Comparison of pharmacology and clinical efficacy. Eur J Med Res 2010;15:507–513 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Elyasi S, Khalili H, Dashti-Khavidaki S, Mohammadpour A. Vancomycin-induced nephrotoxicity: Mechanism, incidence, risk factors and special populations. A literature review. Eur J Clin Pharmacol 2012;68:1243–1255 [DOI] [PubMed] [Google Scholar]
- 14.Bosso JA, Nappi J, Rudisill C, et al. Relationship between vancomycin trough concentrations and nephrotoxicity: A prospective multicenter trial. Antimicrob Agents Chemother 2011;55:5475–5479 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kureishi A, Jewesson PJ, Rubinger M, et al. Double-blind comparison of teicoplanin versus vancomycin in febrile neutropenic patients receiving concomitant tobramycin and piperacillin: Effect on cyclosporin A-associated nephrotoxicity. Antimicrob Agents Chemother 1991;35:2246–2252 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lodise TP, Lomaestro B, Graves J, Drusano GL. Larger vancomycin doses (at least four grams per day) are associated with an increased incidence of nephrotoxicity. Antimicrobial Agents Chemother 2008;52:1330–1336 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cohen E, Dadashev A, Drucker M, et al. Once-daily versus twice-daily intravenous administration of vancomycin for infections in hospitalized patients. J Antimicrob Chemother 2002;49:155–160 [DOI] [PubMed] [Google Scholar]
- 18.Cano EL, Haque NZ, Welch VL, et al. Incidence of nephrotoxicity and association with vancomycin use in intensive care unit patients with pneumonia: Retrospective analysis of the IMPACT-HAP Database. Clin Ther 2012;34:149–157 [DOI] [PubMed] [Google Scholar]
- 19.Wunderink RG, Niederman MS, Kollef MH, et al. Linezolid in methicillin-resistant Staphylococcus aureus nosocomial pneumonia: A randomized, controlled study. Clin Infect Dis 2012;54:621–629 [DOI] [PubMed] [Google Scholar]
- 20.Davies SW, Guidry CA, Petroze RT, et al. Vancomycin and nephrotoxicity: just another myth? J Trauma Acute Care Surg 2013;75:830–835 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Burgess LD, Drew RH. Comparison of the incidence of vancomycin-induced nephrotoxicity in hospitalized patients with and without concomitant piperacillin-tazobactam. Pharmacotherapy 2014;34:670–676 [DOI] [PubMed] [Google Scholar]
- 22.Gomes DM, Smotherman C, Birch A, et al. Comparison of acute kidney injury during treatment with vancomycin in combination with piperacillin-tazobactam or cefepime. Pharmacotherapy 2014;34:662–669 [DOI] [PubMed] [Google Scholar]
- 23.Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309–332 [DOI] [PubMed] [Google Scholar]
- 24.Jeffres MN, Isakow W, Doherty JA, et al. A retrospective analysis of possible renal toxicity associated with vancomycin in patients with health care-associated methicillin-resistant Staphylococcus aureus pneumonia. Clin Ther 2007;29:1107–1115 [DOI] [PubMed] [Google Scholar]
- 25.Pan W. Akaike's information criterion in generalized estimating equations. Biometrics 2001;57:120–125 [DOI] [PubMed] [Google Scholar]
- 26.Huber PJ. (ed). The Behavior of Maximum Likelihood Estimates Under Nonstandard Conditions. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, CA: University of California Press, 1967 [Google Scholar]
- 27.Dilworth TJ, Sliwinski J, Ryan K, et al. Evaluation of vancomycin in combination with piperacillin-tazobactam or oxacillin against clinical methicillin-resistant Staphylococcus aureus isolates and vancomycin-intermediate S. aureus isolates in vitro. Antimicrob Agents Chemother 2014;58:1028–1033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Vandecasteele SJ, De Vriese AS, Tacconelli E. The pharmacokinetics and pharmacodynamics of vancomycin in clinical practice: Evidence and uncertainties. J Antimicrob Chemother 2013;68:743–748 [DOI] [PubMed] [Google Scholar]
- 29.Martin JH, Norris R, Barras M, et al. Therapeutic monitoring of vancomycin in adult patients: A consensus review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society Of Infectious Diseases Pharmacists. Clin Biochem Rev 2010;31:21–24 [PMC free article] [PubMed] [Google Scholar]
- 30.Mah GT, Mabasa VH, Chow I, Ensom MH. Evaluating outcomes associated with alternative dosing strategies for piperacillin/tazobactam: A qualitative systematic review. Ann Pharmacother 2012;46:265–275 [DOI] [PubMed] [Google Scholar]
- 31.Pratt JA, Stricherz MK, Verghese PS, Burke MJ. Suspected piperacillin-tazobactam induced nephrotoxicity in the pediatric oncology population. Pediatr Blood Cancer 2014;61:366–368 [DOI] [PubMed] [Google Scholar]
- 32.Pill MW, O'Neill CV, Chapman MM, Singh AK. Suspected acute interstitial nephritis induced by piperacillin-tazobactam. Pharmacotherapy 1997;17:166–169 [PubMed] [Google Scholar]
- 33.Moenster RP, Linneman TW, Finnegan PM, et al. Acute renal failure associated with vancomycin and beta-lactams for the treatment of osteomyelitis in diabetics: Piperacillin-tazobactam as compared with cefepime. Clin Microbiol Infect 2014;20:O384–389 [DOI] [PubMed] [Google Scholar]
- 34.Jensen JU, Hein L, Lundgren B, et al. Kidney failure related to broad-spectrum antibiotics in critically ill patients: Secondary end point results from a 1200 patient randomised trial. BMJ Open 2012;2:e000635. [DOI] [PMC free article] [PubMed] [Google Scholar]