Journal of Antimicrobial Chemotherapy Advance Access published February 3, 2015
J Antimicrob Chemother doi:10.1093/jac/dkv014
Risk factors for nephrotoxicity onset associated with polymyxin B therapy Yanina Dubrovskaya1, Nishant Prasad2, Yuman Lee3†, Diana Esaian1, Deborah A. Figueroa4 and Vincent H. Tam5* 1
Department of Pharmacy, New York University Langone Medical Center, New York, NY, USA; 2The Dr. James J. Rahal, Jr. Division of Infectious Diseases, New York Hospital Queens, Flushing, NY, USA; 3Department of Pharmacy, Maimonides Medical Center, Brooklyn, NY, USA; 4Department of Pharmacy, New York Hospital Queens, Flushing, NY, USA; 5Department of Clinical Sciences and Administration, University of Houston College of Pharmacy, Houston, TX, USA
Received 21 November 2014; returned 13 December 2014; revised 30 December 2014; accepted 7 January 2015 Objectives: Polymyxin B is an active agent against many MDR Gram-negative bacteria, but nephrotoxicity is a major hindrance to its widespread use. To guide its optimal use, we determined the risk factors for nephrotoxicity onset associated with polymyxin B. Methods: In a multicentre, retrospective, cohort study, we evaluated adult patients with normal renal function who received ≥72 h of polymyxin B therapy. Pertinent information was retrieved from medical records; patients were followed for up to 30 days after therapy was started. The primary endpoint of this study was the onset of nephrotoxicity. A Cox proportional hazards model was used for analysis. Results: A total of 192 patients (52.1% male, 67.7% Caucasian) were evaluated. The mean+SD age, actual body weight (ABW) and daily dose by ABW were 68.3+17.2 years, 71.5+20.4 kg and 1.5+0.5 mg/kg, respectively. The median duration of therapy was 9.5 days. The overall prevalence rate of nephrotoxicity was 45.8% and the median onset of nephrotoxicity was 9 days. Independent risk factors for the onset of nephrotoxicity included daily dose by ABW (HR ¼1.73; P ¼0.022), concurrent use of vancomycin (HR¼ 1.89; P ¼ 0.005) and contrast media (HR ¼1.79; P ¼ 0.009). Nephrotoxicity was seen earlier in the high-risk group (P ¼ 0.003). Conclusions: Risk factors for nephrotoxicity onset associated with polymyxin B were identified. In conjunction with susceptibility and other pharmacokinetic/pharmacodynamic data, our results can be used to optimize treatment for MDR Gram-negative infections. Keywords: MDR Gram-negative infections, model, nephrotoxic
Introduction Infections due to MDR Gram-negative bacteria are increasing worldwide.1,2 Treatment of these infections presents a clinical challenge and multiple studies reported high mortality rates in patients with carbapenem-resistant Enterobacteriaceae, MDR Pseudomonas aeruginosa and MDR Acinetobacter baumannii.3 – 7 Polymyxin B is one of the two polymyxin antibiotics that remain active against these MDR Gram-negative bacteria. However, nephrotoxicity is a major hindrance to widespread clinical use of the polymyxins and historically other alternatives have been preferentially used. Three recent retrospective cohort studies with a large number of patients reported a potentially better nephrotoxicity profile with polymyxin B compared with colistin.8 – 10 In view of these findings and more consistent pharmacological characteristics, the use of polymyxin B is expected to increase in the near future.11
Identifying ways to minimize nephrotoxicity associated with polymyxin B use is critical to advance clinical practice. Previous studies have evaluated the prevalence rates of nephrotoxicity, described time to the development of acute kidney injury (AKI), examined the impact of different polymyxin B dosing frequencies on the nephrotoxicity and identified independent predictors of AKI.12 – 14 The objective of this study was to investigate the risk factors associated with the onset of nephrotoxicity in patients on polymyxin B therapy.
Methods Study design and clinical sites This multicentre, retrospective, cohort study was conducted from January 2007 to September 2013 in three US teaching hospitals: New York
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
*Corresponding author. Tel: +1-832-842-8316; Fax: +1-832-842-8383; E-mail: [email protected] †Present address: St. Barnabas Hospital, 4422 3rd Ave., Bronx, NY 10457, USA.
Dubrovskaya et al.
University Langone Medical Center (a 700 bed academic medical centre in New York, NY, USA), New York Hospital Queens (a 520 bed universityaffiliated community hospital in Flushing, NY, USA) and Maimonides Medical Center (a 700 bed tertiary teaching hospital in Brooklyn, NY, USA). Institutional Review Board approvals at each study site and the University of Houston were obtained prior to the initiation of this study. In view of the retrospective nature of the study, the need for informed consent was waived.
Table 1. Characteristics of all patients (N¼192) 68.3+17.2
Male, n (%)
100 (52.1)
Caucasian, n (%)
130 (67.7)
Length of hospital stay (days), mean+SD
37.9+30.3
Patient selection criteria
Baseline serum creatinine (mg/dL), mean+SD
0.8+0.3
Adult patients (≥18 years) who were given ≥72 h of intravenous polymyxin B daily for suspected or documented infections were included in this study. Patients with underlying baseline renal insufficiency (baseline serum creatinine .1.5 mg/dL or requiring any form of renal replacement therapy) or fluctuating renal function (increase or decrease in serum creatinine of .50% from baseline 72 h prior to initiation of polymyxin B therapy) were excluded.
ABW (kg), mean+SD
71.5+20.4
Polymyxin B daily dose based on ABW (mg/kg), mean+SD
1.5+0.5
Polymyxin B daily dose based on IBW (mg/kg), mean+SD
1.8+0.6
Duration of therapy (days), mean+SD
11.7+7.7
Prevalence of nephrotoxicity, n (%) risk injury failure
88 (45.8) 37 (19.3) 23 (12.0) 28 (14.6)
Median nephrotoxicity onset (days)
9.0
Study variables/outcome definition Data collected included demographics (e.g. age, ethnicity and gender), comorbidities (e.g. diabetes mellitus, cystic fibrosis and ICU stay during polymyxin B therapy), pertinent laboratory findings (e.g. serum creatinine, site of positive culture and organisms isolated), polymyxin B therapy regimens (daily dose prior to the onset of nephrotoxicity and duration) and concomitant nephrotoxins (e.g. aminoglycosides, loop diuretics, vasopressors, calcineurin inhibitors, vancomycin, angiotensin-converting enzyme inhibitors and angiotensin receptor blockers) with polymyxin B therapy. For descriptive purposes, the severity of nephrotoxicity was defined by the RIFLE (risk, injury, failure, loss, end-stage kidney disease) criteria based on the highest serum creatinine observed during polymyxin B therapy and compared with baseline.15 The primary outcome of the study was the onset of nephrotoxicity regardless of severity (the first day of therapy on which serum creatinine was ≥1.5× the baseline). Patients were followed for up to 30 days from the first day of polymyxin B therapy (considered as baseline) or until hospital discharge, whichever was earlier.
ABW, actual body weight; IBW, ideal body weight.
49.5%), followed by urine (51 patients; 26.6%) and the bloodstream (22 patients; 11.5%). The most common bacteria isolated were Klebsiella spp. (92 patients; 47.9%) and A. baumannii (53 patients; 27.6%). The overall prevalence rate of nephrotoxicity was 45.8% (88/192) and the median onset of nephrotoxicity was 9 days (IQR ¼ 5– 16 days). Among the 88 patients who developed nephrotoxicity, 3 required haemodialysis. Based on the data recorded on the last day, serum creatinine returned to ,1.5× the baseline in 36 out of 85 (42.4%) non-dialysis patients.
Statistical analysis
Risk factors for the onset of nephrotoxicity
To identify independent risk factors associated with the onset of nephrotoxicity, a Cox proportional hazards model was used to analyse the data. Using this analysis framework, we provided an additional time dimension to the outcome of interest, compared with the more conventional (logistic) regression. Right censoring was used for events not directly observed. Univariate analysis was used to derive HRs of different variables. Variables with P values ,0.2 in the univariate analysis were included in a multivariate analysis and backward selection was used. Variables with P values ≤0.05 were considered significant in the final model. To validate the best-fit model, patients were stratified into above versus below the median predicted hazards. Continuous variables were compared using Student’s t-test or Kruskal – Wallis one-way analysis of variance. Categorical variables were compared using Fisher’s exact test. The time to nephrotoxicity in the stratified cohorts was compared using the Kaplan – Meier survival analysis and log-rank test. All statistical analyses were performed by using SYSTAT version 12 (SYSTAT Software, Chicago, IL, USA).
Risk factors for the onset of nephrotoxicity are listed in Table 2. In the multivariate analysis, only three variables (daily dose by actual body weight, concurrent use of vancomycin and contrast media) were found to be independently associated with the onset of nephrotoxicity, after adjusting for underlying confounders.
Model validation By conditioning the best-fit model using patient-specific risk factors, the patients were stratified into two cohorts based on the predicted hazards. Patients with above the median predicted hazards were classified as high-risk patients and patients with below the median predicted hazards were categorized as low-risk patients. A comparison of the cohorts stratified by the predicted hazards is shown in Table 3. The time to nephrotoxicity between the two stratified cohorts is as shown in Figure 1. As anticipated, the time to nephrotoxicity was significantly earlier in the high-risk group (P ¼ 0.003).
Results Baseline characteristics of patients
Discussion
A total of 192 patients were evaluated. Key demographic and clinical characteristics are summarized in Table 1. The most common site of positive culture was the respiratory tract (95 patients;
With a rising prevalence of multidrug resistance in Gram-negative bacteria, the polymyxins are going to be increasingly used as the treatment of last resort. Over the past decade, several
2 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
Age (years), mean+SD
JAC
Nephrotoxicity onset with polymyxin B
Table 2. Risk factors associated with the onset of nephrotoxicity (overall)
Variable
P value
0.99 1.50 1.12 0.88 1.45 1.43 1.90 1.85 1.33 0.51 1.19 1.60 0.60
0.212 0.073 0.515 0.576 0.085 0.097 0.004 0.006 0.211 0.109 0.021 0.033 0.066
ABW Daily dose (mg/kg ABW) Daily dose (mg/kg IBW) Caucasian ICU stay during therapy Loop diuretics Intravenous contrast media Vancomycin Vasopressors ACE inhibitor/angiotensin receptor blocker Total number of nephrotoxins Positive culture in respiratory tract Positive culture in urinary tract
Multivariate analysis HR
P value
1.73
0.022
1.79 1.89
0.009 0.005
ABW, actual body weight; IBW, ideal body weight; ACE, angiotensin-converting enzyme.
Table 3. Comparison of stratified cohorts High risk, N ¼96
Low risk, N¼96
Age (years), mean+SD
66.4+18.1
70.1+16.2
0.137
Male, n (%)
55 (57.3)
45 (46.9)
0.153
Caucasian, n (%)
59 (61.5)
71 (74.0)
0.053
Length of hospital stay (days), mean+SD
40.3+32.7
35.4+27.7
0.264
Baseline serum creatinine (mg/dL), mean+SD
0.8+0.3
0.8+0.3
0.127
ABW (kg), mean+SD
67.3+18.7
75.7+21.3
0.004
Polymyxin B daily dose based on ABW (mg/kg), mean+SD
1.7+0.5
1.3+0.4
,0.001
Polymyxin B daily dose based on IBW (mg/kg), mean+SD
1.9+0.6
1.7+0.5
0.019
Duration of therapy (days), mean+SD
11.9+7.5
11.5+7.9
0.681
Intravenous contrast media, n (%)
42 (43.8)
5 (5.2)
,0.001
Vancomycin, n (%)
77 (80.2)
28 (29.2)
,0.001
Prevalence of nephrotoxicity, n (%) risk injury failure
55 (57.3) 22 (22.9) 15 (15.6) 18 (18.8)
33 (34.4) 15 (15.6) 8 (8.3) 10 (10.4)
0.002 0.272 0.181 0.151
Variable
P value
ABW, actual body weight; IBW, ideal body weight.
pivotal studies have provided critical information on the optimal dosing regimen of polymyxin B. Since the area under the concentration – time curve (AUC)/MIC best describes the bactericidal activity of polymyxin B, using higher doses should be associated with better outcomes (when susceptibility remains unchanged).16 Consistent with our expectation, Elias et al.14 reported that a daily dosage of ≥2 million IU of polymyxin B (equivalent to the activity of 200 mg of purified polymyxin B1 component) was associated with lower in-hospital mortality. However, a daily polymyxin B
dose of ≥2 million IU was found to significantly increase the probability of AKI.10,14 A study by Kubin et al.13 also reported that therapy was discontinued in 14% of patients given polymyxin B due to nephrotoxicity. These findings suggest that nephrotoxicity is a dose-limiting adverse effect of polymyxin B. The overall prevalence rate of nephrotoxicity in our study was 45.8% (88/192), with the majority of the patients qualifying for the risk category (n ¼ 37), followed by the failure category (n ¼ 28). The rate of nephrotoxicity associated with polymyxin B
3 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
Univariate analysis HR
Dubrovskaya et al.
Proportion of patients not reaching endpoint 1.0
0.8
0.6
0.4
0.0
0
5
10
15 20 Time (days)
25
30
35
Figure 1. Comparison of nephrotoxicity onset. P ¼0.003. The broken line represents low-risk patients and the continuous line represents high-risk patients. Relative hazards function ¼ 0.547 (daily dose in mg/kg actual body weight) +0.639 (vancomycin) +0.585 (contrast). Vancomycin ¼ 1 if given concurrently, otherwise ¼ 0. Contrast ¼ 1 if given concurrently, otherwise¼0.
assessed using RIFLE criteria by four other studies ranged from 23.1% to 60%.8,9,13,17 The severity of nephrotoxicity was reported in two of these studies, where the majority of the patients qualified for the injury or failure categories.8,9 Although the observed rates of nephrotoxicity were high in these studies, kidney injury was reversible in the majority of the patients. Understanding the temporal relationship for the development of kidney injury associated with polymyxin B therapy is vital to minimize nephrotoxicity and improve clinical outcomes. We have previously examined and reported the risk factors for nephrotoxicity associated with polymyxin B.8 In the present study, we focused on the onset of nephrotoxicity. Currently there are limited data on the onset of nephrotoxicity. In an animal model, Abdelraouf et al.12 reported that polymyxin B-induced nephrotoxicity manifested as acute tubular necrosis and extensive injury of the proximal tubules was observed. Polymyxin B dosing frequency had an impact on the onset of nephrotoxicity. The lesions were more severe and higher drug concentrations were achieved in the kidney in the six hourly dosing group. With the same daily dose, the once-daily dosing group experienced a more gradual onset of nephrotoxicity, which could be attributed to the lower kidney tissue drug concentrations.12,18 Kubin et al.13 showed that earlier onset and increased risk of nephrotoxicity was observed in patients with BMI ≥25 kg/m2 including overweight, obese and volume-overloaded patients. Two other studies reported that onset of nephrotoxicity occurred earlier in patients who received colistin than in patients who were given polymyxin B.8,9 In our patients with normal baseline renal function, the median duration of polymyxin B treatment was 9.5 days
4 of 5
Acknowledgements This study was presented in part at the Fifty-fourth Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, 2014 (Abstract A-031).
Funding No grant, specific funding or any other financial support was received for this study.
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
0.2
(IQR ¼ 7 – 14 days) and the median onset of nephrotoxicity (regardless of severity) was 9 days (IQR¼ 5 –16 days). These findings are consistent with several other studies that examined the onset of nephrotoxicity in patients who received polymyxin B. In a retrospective study by Kubin et al.13 (n¼ 73), the median duration of polymyxin B therapy was 11 days and the median time to the first sign of AKI was 4 days (range 1 – 14 days). In comparison, Akajagbor et al.9 (n ¼ 67) reported a mean time to peak serum creatinine of 7.3 days with a mean treatment duration of 12.5 days. Similarly, Esaian et al.17 (n ¼ 80) reported a mean time to peak serum creatinine of 9 days. Collectively, these findings suggested that kidney injury typically occurred after 1 week of therapy and the average time to peak serum creatinine was 7 –9 days. Independent risk factors for the onset of polymyxin B-associated nephrotoxicity identified in our study included daily dose by actual body weight, concurrent use of vancomycin and contrast media. Taking one step further to validate the best-fit model derived, we stratified the patients by the predicted hazards and compared the time to nephrotoxicity between the two patient strata. Time to nephrotoxicity was significantly earlier in the highrisk group (P ¼ 0.003). The finding that concomitant vancomycin might play a role in augmenting the nephrotoxicity associated with polymyxin B was also suggested by two other studies.10,13 Additional independent predictors of polymyxin B-associated AKI described by other studies include age, the need for vasoactive drugs, duration of therapy and BMI ≥25 kg/m2.8,9,13,19,20 There are several limitations to our study. We had a small sample size; therefore, we might not have been able to examine the contribution of all possible risk factors. Also, since it was a multicentre study, there could be variation among the participating hospitals in patient population, the standard of care, specific polymyxin B products used, dosing regimens and infusion time utilized. We did not include patients with underlying renal insufficiency; therefore, the results should not be extrapolated to these patients. In addition, we did not evaluate illness acuteness comprehensively (other than tracking patient stay in ICUs and the use of vasoactive agents). Finally, we did not evaluate fluid balance status and urinary output of the patients, which might have an effect on the development of kidney injury. In summary, our study identified risk factors for the onset of nephrotoxicity associated with polymyxin B therapy. In conjunction with susceptibility and other pharmacokinetic/pharmacodynamic data, our results can be used to optimize polymyxin B treatment for MDR Gram-negative infections. Further investigations are ongoing to minimize nephrotoxicity associated with polymyxin B.
Nephrotoxicity onset with polymyxin B
Transparency declarations None to declare.
References
JAC 10 Tuon FF, Rigatto MH, Lopes CK et al. Risk factors for acute kidney injury in patients treated with polymyxin B or colistin methanesulfonate sodium. Int J Antimicrob Agents 2014; 43: 349–52. 11 Nation RL, Velkov T, Li J. Colistin and polymyxin B: peas in a pod, or chalk and cheese? Clin Infect Dis 2014; 59: 88 –94.
1 Nordmann P, Naas T, Poirel L. Global spread of carbapenemaseproducing Enterobacteriaceae. Emerg Infect Dis 2011; 17: 1791 –8.
12 Abdelraouf K, Braggs KH, Yin T et al. Characterization of polymyxin B-induced nephrotoxicity: implications for dosing regimen design. Antimicrob Agents Chemother 2012; 56: 4625– 9.
2 Gupta N, Limbago BM, Patel JB et al. Carbapenem-resistant Enterobacteriaceae: epidemiology and prevention. Clin Infect Dis 2011; 53: 60– 7.
13 Kubin CJ, Ellman TM, Phadke V et al. Incidence and predictors of acute kidney injury associated with intravenous polymyxin B therapy. J Infect 2012; 65: 80– 7.
3 Tam VH, Rogers CA, Chang KT et al. Impact of multidrug-resistant Pseudomonas aeruginosa bacteremia on patient outcomes. Antimicrob Agents Chemother 2010; 54: 3717 –22.
14 Elias LS, Konzen D, Krebs JM et al. The impact of polymyxin B dosage on in-hospital mortality of patients treated with this antibiotic. J Antimicrob Chemother 2010; 65: 2231– 7.
5 Kwa AL, Low JG, Lee E et al. The impact of multidrug resistance on the outcomes of critically ill patients with Gram-negative bacterial pneumonia. Diagn Microbiol Infect Dis 2007; 58: 99–104. 6 Hirsch EB, Tam VH. Impact of multidrug-resistant Pseudomonas aeruginosa infection on patient outcomes. Expert Rev Pharmacoecon Outcomes Res 2010; 10: 441– 51. 7 Patel G, Huprikar S, Factor SH et al. Outcomes of carbapenem-resistant Klebsiella pneumoniae infection and the impact of antimicrobial and adjunctive therapies. Infect Control Hosp Epidemiol 2008; 29: 1099 – 106. 8 Phe K, Lee Y, McDaneld PM et al. In vitro assessment and multicenter cohort study of comparative nephrotoxicity rates associated with colistimethate versus polymyxin B therapy. Antimicrob Agents Chemother 2014; 58: 2740– 6. 9 Akajagbor DS, Wilson SL, Shere-Wolfe KD et al. Higher incidence of acute kidney injury with intravenous colistimethate sodium compared with polymyxin B in critically ill patients at a tertiary care medical center. Clin Infect Dis 2013; 57: 1300– 3.
15 Bellomo R, Ronco C, Kellum JA et al. Acute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8: R204 – 12. 16 Tam VH, Schilling AN, Vo G et al. Pharmacodynamics of polymyxin B against Pseudomonas aeruginosa. Antimicrob Agents Chemother 2005; 49: 3624– 30. 17 Esaian D, Dubrovskaya Y, Phillips M et al. Effectiveness and tolerability of a polymyxin B dosing protocol. Ann Pharmacother 2012; 46: 455–6. 18 Abdelraouf K, He J, Ledesma KR et al. Pharmacokinetics and renal disposition of polymyxin B in an animal model. Antimicrob Agents Chemother 2012; 56: 5724– 7. 19 Ouderkirk JP, Nord JA, Turett GS et al. Polymyxin B nephrotoxicity and efficacy against nosocomial infections caused by multiresistant gram-negative bacteria. Antimicrob Agents Chemother 2003; 47: 2659 – 62. 20 Mendes CA, Cordeiro JA, Burdmann EA. Prevalence and risk factors for acute kidney injury associated with parenteral polymyxin B use. Ann Pharmacother 2009; 43: 1948– 55.
5 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
4 Abbo A, Carmeli Y, Navon-Venezia S et al. Impact of multi-drug-resistant Acinetobacter baumannii on clinical outcomes. Eur J Clin Microbiol Infect Dis 2007; 26: 793–800.
J Antimicrob Chemother doi:10.1093/jac/dkv014
Risk factors for nephrotoxicity onset associated with polymyxin B therapy Yanina Dubrovskaya1, Nishant Prasad2, Yuman Lee3†, Diana Esaian1, Deborah A. Figueroa4 and Vincent H. Tam5* 1
Department of Pharmacy, New York University Langone Medical Center, New York, NY, USA; 2The Dr. James J. Rahal, Jr. Division of Infectious Diseases, New York Hospital Queens, Flushing, NY, USA; 3Department of Pharmacy, Maimonides Medical Center, Brooklyn, NY, USA; 4Department of Pharmacy, New York Hospital Queens, Flushing, NY, USA; 5Department of Clinical Sciences and Administration, University of Houston College of Pharmacy, Houston, TX, USA
Received 21 November 2014; returned 13 December 2014; revised 30 December 2014; accepted 7 January 2015 Objectives: Polymyxin B is an active agent against many MDR Gram-negative bacteria, but nephrotoxicity is a major hindrance to its widespread use. To guide its optimal use, we determined the risk factors for nephrotoxicity onset associated with polymyxin B. Methods: In a multicentre, retrospective, cohort study, we evaluated adult patients with normal renal function who received ≥72 h of polymyxin B therapy. Pertinent information was retrieved from medical records; patients were followed for up to 30 days after therapy was started. The primary endpoint of this study was the onset of nephrotoxicity. A Cox proportional hazards model was used for analysis. Results: A total of 192 patients (52.1% male, 67.7% Caucasian) were evaluated. The mean+SD age, actual body weight (ABW) and daily dose by ABW were 68.3+17.2 years, 71.5+20.4 kg and 1.5+0.5 mg/kg, respectively. The median duration of therapy was 9.5 days. The overall prevalence rate of nephrotoxicity was 45.8% and the median onset of nephrotoxicity was 9 days. Independent risk factors for the onset of nephrotoxicity included daily dose by ABW (HR ¼1.73; P ¼0.022), concurrent use of vancomycin (HR¼ 1.89; P ¼ 0.005) and contrast media (HR ¼1.79; P ¼ 0.009). Nephrotoxicity was seen earlier in the high-risk group (P ¼ 0.003). Conclusions: Risk factors for nephrotoxicity onset associated with polymyxin B were identified. In conjunction with susceptibility and other pharmacokinetic/pharmacodynamic data, our results can be used to optimize treatment for MDR Gram-negative infections. Keywords: MDR Gram-negative infections, model, nephrotoxic
Introduction Infections due to MDR Gram-negative bacteria are increasing worldwide.1,2 Treatment of these infections presents a clinical challenge and multiple studies reported high mortality rates in patients with carbapenem-resistant Enterobacteriaceae, MDR Pseudomonas aeruginosa and MDR Acinetobacter baumannii.3 – 7 Polymyxin B is one of the two polymyxin antibiotics that remain active against these MDR Gram-negative bacteria. However, nephrotoxicity is a major hindrance to widespread clinical use of the polymyxins and historically other alternatives have been preferentially used. Three recent retrospective cohort studies with a large number of patients reported a potentially better nephrotoxicity profile with polymyxin B compared with colistin.8 – 10 In view of these findings and more consistent pharmacological characteristics, the use of polymyxin B is expected to increase in the near future.11
Identifying ways to minimize nephrotoxicity associated with polymyxin B use is critical to advance clinical practice. Previous studies have evaluated the prevalence rates of nephrotoxicity, described time to the development of acute kidney injury (AKI), examined the impact of different polymyxin B dosing frequencies on the nephrotoxicity and identified independent predictors of AKI.12 – 14 The objective of this study was to investigate the risk factors associated with the onset of nephrotoxicity in patients on polymyxin B therapy.
Methods Study design and clinical sites This multicentre, retrospective, cohort study was conducted from January 2007 to September 2013 in three US teaching hospitals: New York
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
1 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
*Corresponding author. Tel: +1-832-842-8316; Fax: +1-832-842-8383; E-mail: [email protected] †Present address: St. Barnabas Hospital, 4422 3rd Ave., Bronx, NY 10457, USA.
Dubrovskaya et al.
University Langone Medical Center (a 700 bed academic medical centre in New York, NY, USA), New York Hospital Queens (a 520 bed universityaffiliated community hospital in Flushing, NY, USA) and Maimonides Medical Center (a 700 bed tertiary teaching hospital in Brooklyn, NY, USA). Institutional Review Board approvals at each study site and the University of Houston were obtained prior to the initiation of this study. In view of the retrospective nature of the study, the need for informed consent was waived.
Table 1. Characteristics of all patients (N¼192) 68.3+17.2
Male, n (%)
100 (52.1)
Caucasian, n (%)
130 (67.7)
Length of hospital stay (days), mean+SD
37.9+30.3
Patient selection criteria
Baseline serum creatinine (mg/dL), mean+SD
0.8+0.3
Adult patients (≥18 years) who were given ≥72 h of intravenous polymyxin B daily for suspected or documented infections were included in this study. Patients with underlying baseline renal insufficiency (baseline serum creatinine .1.5 mg/dL or requiring any form of renal replacement therapy) or fluctuating renal function (increase or decrease in serum creatinine of .50% from baseline 72 h prior to initiation of polymyxin B therapy) were excluded.
ABW (kg), mean+SD
71.5+20.4
Polymyxin B daily dose based on ABW (mg/kg), mean+SD
1.5+0.5
Polymyxin B daily dose based on IBW (mg/kg), mean+SD
1.8+0.6
Duration of therapy (days), mean+SD
11.7+7.7
Prevalence of nephrotoxicity, n (%) risk injury failure
88 (45.8) 37 (19.3) 23 (12.0) 28 (14.6)
Median nephrotoxicity onset (days)
9.0
Study variables/outcome definition Data collected included demographics (e.g. age, ethnicity and gender), comorbidities (e.g. diabetes mellitus, cystic fibrosis and ICU stay during polymyxin B therapy), pertinent laboratory findings (e.g. serum creatinine, site of positive culture and organisms isolated), polymyxin B therapy regimens (daily dose prior to the onset of nephrotoxicity and duration) and concomitant nephrotoxins (e.g. aminoglycosides, loop diuretics, vasopressors, calcineurin inhibitors, vancomycin, angiotensin-converting enzyme inhibitors and angiotensin receptor blockers) with polymyxin B therapy. For descriptive purposes, the severity of nephrotoxicity was defined by the RIFLE (risk, injury, failure, loss, end-stage kidney disease) criteria based on the highest serum creatinine observed during polymyxin B therapy and compared with baseline.15 The primary outcome of the study was the onset of nephrotoxicity regardless of severity (the first day of therapy on which serum creatinine was ≥1.5× the baseline). Patients were followed for up to 30 days from the first day of polymyxin B therapy (considered as baseline) or until hospital discharge, whichever was earlier.
ABW, actual body weight; IBW, ideal body weight.
49.5%), followed by urine (51 patients; 26.6%) and the bloodstream (22 patients; 11.5%). The most common bacteria isolated were Klebsiella spp. (92 patients; 47.9%) and A. baumannii (53 patients; 27.6%). The overall prevalence rate of nephrotoxicity was 45.8% (88/192) and the median onset of nephrotoxicity was 9 days (IQR ¼ 5– 16 days). Among the 88 patients who developed nephrotoxicity, 3 required haemodialysis. Based on the data recorded on the last day, serum creatinine returned to ,1.5× the baseline in 36 out of 85 (42.4%) non-dialysis patients.
Statistical analysis
Risk factors for the onset of nephrotoxicity
To identify independent risk factors associated with the onset of nephrotoxicity, a Cox proportional hazards model was used to analyse the data. Using this analysis framework, we provided an additional time dimension to the outcome of interest, compared with the more conventional (logistic) regression. Right censoring was used for events not directly observed. Univariate analysis was used to derive HRs of different variables. Variables with P values ,0.2 in the univariate analysis were included in a multivariate analysis and backward selection was used. Variables with P values ≤0.05 were considered significant in the final model. To validate the best-fit model, patients were stratified into above versus below the median predicted hazards. Continuous variables were compared using Student’s t-test or Kruskal – Wallis one-way analysis of variance. Categorical variables were compared using Fisher’s exact test. The time to nephrotoxicity in the stratified cohorts was compared using the Kaplan – Meier survival analysis and log-rank test. All statistical analyses were performed by using SYSTAT version 12 (SYSTAT Software, Chicago, IL, USA).
Risk factors for the onset of nephrotoxicity are listed in Table 2. In the multivariate analysis, only three variables (daily dose by actual body weight, concurrent use of vancomycin and contrast media) were found to be independently associated with the onset of nephrotoxicity, after adjusting for underlying confounders.
Model validation By conditioning the best-fit model using patient-specific risk factors, the patients were stratified into two cohorts based on the predicted hazards. Patients with above the median predicted hazards were classified as high-risk patients and patients with below the median predicted hazards were categorized as low-risk patients. A comparison of the cohorts stratified by the predicted hazards is shown in Table 3. The time to nephrotoxicity between the two stratified cohorts is as shown in Figure 1. As anticipated, the time to nephrotoxicity was significantly earlier in the high-risk group (P ¼ 0.003).
Results Baseline characteristics of patients
Discussion
A total of 192 patients were evaluated. Key demographic and clinical characteristics are summarized in Table 1. The most common site of positive culture was the respiratory tract (95 patients;
With a rising prevalence of multidrug resistance in Gram-negative bacteria, the polymyxins are going to be increasingly used as the treatment of last resort. Over the past decade, several
2 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
Age (years), mean+SD
JAC
Nephrotoxicity onset with polymyxin B
Table 2. Risk factors associated with the onset of nephrotoxicity (overall)
Variable
P value
0.99 1.50 1.12 0.88 1.45 1.43 1.90 1.85 1.33 0.51 1.19 1.60 0.60
0.212 0.073 0.515 0.576 0.085 0.097 0.004 0.006 0.211 0.109 0.021 0.033 0.066
ABW Daily dose (mg/kg ABW) Daily dose (mg/kg IBW) Caucasian ICU stay during therapy Loop diuretics Intravenous contrast media Vancomycin Vasopressors ACE inhibitor/angiotensin receptor blocker Total number of nephrotoxins Positive culture in respiratory tract Positive culture in urinary tract
Multivariate analysis HR
P value
1.73
0.022
1.79 1.89
0.009 0.005
ABW, actual body weight; IBW, ideal body weight; ACE, angiotensin-converting enzyme.
Table 3. Comparison of stratified cohorts High risk, N ¼96
Low risk, N¼96
Age (years), mean+SD
66.4+18.1
70.1+16.2
0.137
Male, n (%)
55 (57.3)
45 (46.9)
0.153
Caucasian, n (%)
59 (61.5)
71 (74.0)
0.053
Length of hospital stay (days), mean+SD
40.3+32.7
35.4+27.7
0.264
Baseline serum creatinine (mg/dL), mean+SD
0.8+0.3
0.8+0.3
0.127
ABW (kg), mean+SD
67.3+18.7
75.7+21.3
0.004
Polymyxin B daily dose based on ABW (mg/kg), mean+SD
1.7+0.5
1.3+0.4
,0.001
Polymyxin B daily dose based on IBW (mg/kg), mean+SD
1.9+0.6
1.7+0.5
0.019
Duration of therapy (days), mean+SD
11.9+7.5
11.5+7.9
0.681
Intravenous contrast media, n (%)
42 (43.8)
5 (5.2)
,0.001
Vancomycin, n (%)
77 (80.2)
28 (29.2)
,0.001
Prevalence of nephrotoxicity, n (%) risk injury failure
55 (57.3) 22 (22.9) 15 (15.6) 18 (18.8)
33 (34.4) 15 (15.6) 8 (8.3) 10 (10.4)
0.002 0.272 0.181 0.151
Variable
P value
ABW, actual body weight; IBW, ideal body weight.
pivotal studies have provided critical information on the optimal dosing regimen of polymyxin B. Since the area under the concentration – time curve (AUC)/MIC best describes the bactericidal activity of polymyxin B, using higher doses should be associated with better outcomes (when susceptibility remains unchanged).16 Consistent with our expectation, Elias et al.14 reported that a daily dosage of ≥2 million IU of polymyxin B (equivalent to the activity of 200 mg of purified polymyxin B1 component) was associated with lower in-hospital mortality. However, a daily polymyxin B
dose of ≥2 million IU was found to significantly increase the probability of AKI.10,14 A study by Kubin et al.13 also reported that therapy was discontinued in 14% of patients given polymyxin B due to nephrotoxicity. These findings suggest that nephrotoxicity is a dose-limiting adverse effect of polymyxin B. The overall prevalence rate of nephrotoxicity in our study was 45.8% (88/192), with the majority of the patients qualifying for the risk category (n ¼ 37), followed by the failure category (n ¼ 28). The rate of nephrotoxicity associated with polymyxin B
3 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
Univariate analysis HR
Dubrovskaya et al.
Proportion of patients not reaching endpoint 1.0
0.8
0.6
0.4
0.0
0
5
10
15 20 Time (days)
25
30
35
Figure 1. Comparison of nephrotoxicity onset. P ¼0.003. The broken line represents low-risk patients and the continuous line represents high-risk patients. Relative hazards function ¼ 0.547 (daily dose in mg/kg actual body weight) +0.639 (vancomycin) +0.585 (contrast). Vancomycin ¼ 1 if given concurrently, otherwise ¼ 0. Contrast ¼ 1 if given concurrently, otherwise¼0.
assessed using RIFLE criteria by four other studies ranged from 23.1% to 60%.8,9,13,17 The severity of nephrotoxicity was reported in two of these studies, where the majority of the patients qualified for the injury or failure categories.8,9 Although the observed rates of nephrotoxicity were high in these studies, kidney injury was reversible in the majority of the patients. Understanding the temporal relationship for the development of kidney injury associated with polymyxin B therapy is vital to minimize nephrotoxicity and improve clinical outcomes. We have previously examined and reported the risk factors for nephrotoxicity associated with polymyxin B.8 In the present study, we focused on the onset of nephrotoxicity. Currently there are limited data on the onset of nephrotoxicity. In an animal model, Abdelraouf et al.12 reported that polymyxin B-induced nephrotoxicity manifested as acute tubular necrosis and extensive injury of the proximal tubules was observed. Polymyxin B dosing frequency had an impact on the onset of nephrotoxicity. The lesions were more severe and higher drug concentrations were achieved in the kidney in the six hourly dosing group. With the same daily dose, the once-daily dosing group experienced a more gradual onset of nephrotoxicity, which could be attributed to the lower kidney tissue drug concentrations.12,18 Kubin et al.13 showed that earlier onset and increased risk of nephrotoxicity was observed in patients with BMI ≥25 kg/m2 including overweight, obese and volume-overloaded patients. Two other studies reported that onset of nephrotoxicity occurred earlier in patients who received colistin than in patients who were given polymyxin B.8,9 In our patients with normal baseline renal function, the median duration of polymyxin B treatment was 9.5 days
4 of 5
Acknowledgements This study was presented in part at the Fifty-fourth Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, 2014 (Abstract A-031).
Funding No grant, specific funding or any other financial support was received for this study.
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
0.2
(IQR ¼ 7 – 14 days) and the median onset of nephrotoxicity (regardless of severity) was 9 days (IQR¼ 5 –16 days). These findings are consistent with several other studies that examined the onset of nephrotoxicity in patients who received polymyxin B. In a retrospective study by Kubin et al.13 (n¼ 73), the median duration of polymyxin B therapy was 11 days and the median time to the first sign of AKI was 4 days (range 1 – 14 days). In comparison, Akajagbor et al.9 (n ¼ 67) reported a mean time to peak serum creatinine of 7.3 days with a mean treatment duration of 12.5 days. Similarly, Esaian et al.17 (n ¼ 80) reported a mean time to peak serum creatinine of 9 days. Collectively, these findings suggested that kidney injury typically occurred after 1 week of therapy and the average time to peak serum creatinine was 7 –9 days. Independent risk factors for the onset of polymyxin B-associated nephrotoxicity identified in our study included daily dose by actual body weight, concurrent use of vancomycin and contrast media. Taking one step further to validate the best-fit model derived, we stratified the patients by the predicted hazards and compared the time to nephrotoxicity between the two patient strata. Time to nephrotoxicity was significantly earlier in the highrisk group (P ¼ 0.003). The finding that concomitant vancomycin might play a role in augmenting the nephrotoxicity associated with polymyxin B was also suggested by two other studies.10,13 Additional independent predictors of polymyxin B-associated AKI described by other studies include age, the need for vasoactive drugs, duration of therapy and BMI ≥25 kg/m2.8,9,13,19,20 There are several limitations to our study. We had a small sample size; therefore, we might not have been able to examine the contribution of all possible risk factors. Also, since it was a multicentre study, there could be variation among the participating hospitals in patient population, the standard of care, specific polymyxin B products used, dosing regimens and infusion time utilized. We did not include patients with underlying renal insufficiency; therefore, the results should not be extrapolated to these patients. In addition, we did not evaluate illness acuteness comprehensively (other than tracking patient stay in ICUs and the use of vasoactive agents). Finally, we did not evaluate fluid balance status and urinary output of the patients, which might have an effect on the development of kidney injury. In summary, our study identified risk factors for the onset of nephrotoxicity associated with polymyxin B therapy. In conjunction with susceptibility and other pharmacokinetic/pharmacodynamic data, our results can be used to optimize polymyxin B treatment for MDR Gram-negative infections. Further investigations are ongoing to minimize nephrotoxicity associated with polymyxin B.
Nephrotoxicity onset with polymyxin B
Transparency declarations None to declare.
References
JAC 10 Tuon FF, Rigatto MH, Lopes CK et al. Risk factors for acute kidney injury in patients treated with polymyxin B or colistin methanesulfonate sodium. Int J Antimicrob Agents 2014; 43: 349–52. 11 Nation RL, Velkov T, Li J. Colistin and polymyxin B: peas in a pod, or chalk and cheese? Clin Infect Dis 2014; 59: 88 –94.
1 Nordmann P, Naas T, Poirel L. Global spread of carbapenemaseproducing Enterobacteriaceae. Emerg Infect Dis 2011; 17: 1791 –8.
12 Abdelraouf K, Braggs KH, Yin T et al. Characterization of polymyxin B-induced nephrotoxicity: implications for dosing regimen design. Antimicrob Agents Chemother 2012; 56: 4625– 9.
2 Gupta N, Limbago BM, Patel JB et al. Carbapenem-resistant Enterobacteriaceae: epidemiology and prevention. Clin Infect Dis 2011; 53: 60– 7.
13 Kubin CJ, Ellman TM, Phadke V et al. Incidence and predictors of acute kidney injury associated with intravenous polymyxin B therapy. J Infect 2012; 65: 80– 7.
3 Tam VH, Rogers CA, Chang KT et al. Impact of multidrug-resistant Pseudomonas aeruginosa bacteremia on patient outcomes. Antimicrob Agents Chemother 2010; 54: 3717 –22.
14 Elias LS, Konzen D, Krebs JM et al. The impact of polymyxin B dosage on in-hospital mortality of patients treated with this antibiotic. J Antimicrob Chemother 2010; 65: 2231– 7.
5 Kwa AL, Low JG, Lee E et al. The impact of multidrug resistance on the outcomes of critically ill patients with Gram-negative bacterial pneumonia. Diagn Microbiol Infect Dis 2007; 58: 99–104. 6 Hirsch EB, Tam VH. Impact of multidrug-resistant Pseudomonas aeruginosa infection on patient outcomes. Expert Rev Pharmacoecon Outcomes Res 2010; 10: 441– 51. 7 Patel G, Huprikar S, Factor SH et al. Outcomes of carbapenem-resistant Klebsiella pneumoniae infection and the impact of antimicrobial and adjunctive therapies. Infect Control Hosp Epidemiol 2008; 29: 1099 – 106. 8 Phe K, Lee Y, McDaneld PM et al. In vitro assessment and multicenter cohort study of comparative nephrotoxicity rates associated with colistimethate versus polymyxin B therapy. Antimicrob Agents Chemother 2014; 58: 2740– 6. 9 Akajagbor DS, Wilson SL, Shere-Wolfe KD et al. Higher incidence of acute kidney injury with intravenous colistimethate sodium compared with polymyxin B in critically ill patients at a tertiary care medical center. Clin Infect Dis 2013; 57: 1300– 3.
15 Bellomo R, Ronco C, Kellum JA et al. Acute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8: R204 – 12. 16 Tam VH, Schilling AN, Vo G et al. Pharmacodynamics of polymyxin B against Pseudomonas aeruginosa. Antimicrob Agents Chemother 2005; 49: 3624– 30. 17 Esaian D, Dubrovskaya Y, Phillips M et al. Effectiveness and tolerability of a polymyxin B dosing protocol. Ann Pharmacother 2012; 46: 455–6. 18 Abdelraouf K, He J, Ledesma KR et al. Pharmacokinetics and renal disposition of polymyxin B in an animal model. Antimicrob Agents Chemother 2012; 56: 5724– 7. 19 Ouderkirk JP, Nord JA, Turett GS et al. Polymyxin B nephrotoxicity and efficacy against nosocomial infections caused by multiresistant gram-negative bacteria. Antimicrob Agents Chemother 2003; 47: 2659 – 62. 20 Mendes CA, Cordeiro JA, Burdmann EA. Prevalence and risk factors for acute kidney injury associated with parenteral polymyxin B use. Ann Pharmacother 2009; 43: 1948– 55.
5 of 5
Downloaded from http://jac.oxfordjournals.org/ at Selçuk Ünivesitesi on February 6, 2015
4 Abbo A, Carmeli Y, Navon-Venezia S et al. Impact of multi-drug-resistant Acinetobacter baumannii on clinical outcomes. Eur J Clin Microbiol Infect Dis 2007; 26: 793–800.
