1. RESEARCH REPORTS
Nephrology
Prevalence and Risk Factors for Acute Kidney Injury Associated
with Parenteral Polymyxin B Use
Carlos AC Mendes, José A Cordeiro, and Emmanuel A Burdmann
olymyxins are a group of closely re-
P lated antibiotics consisting of polymy-
xin A, B, C, D, and E, but only polymyxin
BACKGROUND: The main adverse effect of polymyxin B is nephrotoxicity. There
are few data on polymyxin-associated renal injury.
B and E (colistin) are used in clinical prac- OBJECTIVE: To assess the prevalence of and risk factors for acute kidney injury
tice due to the high toxicity of the remain- (AKI) in patients treated with polymyxin B.
ing agents.1 They are cyclic polycationic METHODS: The studied population included 114 patients who received at least 3
decapeptide, containing a heptapeptide consecutive days of intravenous polymyxin B and had baseline serum creatinine
ring, a high percentage of 2,4-diaminobu- (SCr) and at least one further SCr measurement during treatment. AKI was
defined as an SCr increase to 1.8 mg/dL or greater in patients with baseline SCr
tyric acid, and a fatty acid amide bond. less than 1.5 mg/dL, or an increase greater than or equal to 50% in baseline SCr
Polymyxins are amphipathic molecules when it was already greater than or equal to 1.5 mg/dL, or need for dialysis.
that link to the phospholipids and lipo- RESULTS: AKI developed in 22% of the patients. They were older, had a higher
polysaccharides of the bacterial cellular baseline SCr, had a higher frequency of baseline SCr greater than or equal to 1.5
envelope. They competitively dislocate mg/dL, used other nephrotoxic drugs and furosemide more often, and required
calcium and magnesium ions, increasing vasoactive drugs and mechanical ventilation more frequently. Progression to
renal failure was significantly more probable when the bacteria were isolated in
the bacterial cellular membrane perme- the abdomen, catheter, or blood. AKI patients had a higher mortality rate (92% vs
ability and disrupting it, causing bacterial 53%; p < 0.001). Logistic regression identified abnormal baseline SCr (odds ratio
cell death.2,3 [OR] 3.51); need for vasoactive drugs (OR 3.03); and abdomen, blood, or catheter
Polymyxin B is active only against as the infection site (OR 3.82) as independent risk factors for AKI.
gram-negative bacteria, including most CONCLUSIONS: Patients who developed AKI had a strikingly elevated mortality
Pseudomonas aeruginosa strains, Acine- rate. Polymyxin B should be used with extreme caution in patients who have an
abnormal baseline SCr; use vasoactive drugs; or have abdomen, blood, or
tobacter baumannii, Escherichia coli,
catheter as the infection site.
Salmonella, Haemophilus, Shigella, Pas-
KEY WORDS: acute kidney injury, nephrotoxicity, polymyxin B.
teurella, Brucella, Aerobacter aerogenes,
Ann Pharmacother 2009;43:xxxx.
and Bordetella. Other bacteria, such as
Enterobacter spp. and Stenotrophomonas Published Online, 3 Nov 2009, theannals.com, DOI 10.1345/aph.1M277
maltophilia, are less susceptible. Some
bacteria, such as Proteus mirabilis, Ser-
ratia marcescens, Providentia, and Edwardsiella tarda, are ins remains practically unchanged since its introduction into
naturally resistant.4,5 The bactericidal activity of polymyx- clinical practice, with only a few cases of bacterial resis-
tance reported.6,7
In the 1980s, polymyxins were replaced by aminoglyco-
Author information provided at the end of the text. sides as the therapy of choice for severe infections by P.
theannals.com The Annals of Pharmacotherapy I 2009 December, Volume 43 I
Copyright © 2009 by Harvey Whitney Books Company. All rights reserved.

2. CAC Mendes et al.
aeruginosa but were still sometimes used topically or as an use of angiotensin-converting enzyme (ACE) inhibitors or
alternative in resistance to these agents.8 In the 1990s, the angiotensin receptor blockers; use of diuretics (furosemide,
emergence of bacterial strains resistant to almost all antimi- hydrochlorothiazide); use of ranitidine; total daily dose of
crobial agents (β-lactams, aminoglycosides, quinolones)6,9,10 polymyxin B; presence of oliguria (<400 mL urine out-
renewed interest in polymyxins.1,11 put/day) during the administration of the antibiotic; hospi-
The major adverse effects of polymyxins are nephrotox- talization site (clinical or surgical intensive care units
icity and neurotoxicity. In animal studies, polymyxin B in- [ICUs], wards); hypotension (systolic blood pressure <90
duced acute tubular necrosis, albuminuria, urine output de- mm Hg); use of vasoactive drugs (dopamine, noradrena-
crease, and increase of serum urea and creatinine levels.12 line); use of dobutamine; baseline SCr, highest SCr during
Clinical studies on polymyxin B are scarce, and the preva- treatment, and SCr after therapy; need and number of dial-
lence of and risk factors for its nephrotoxicity have not ysis sessions; presence of hyperbilirubinemia, hypo-
been clearly established. kalemia (potassium <3.0 mEq/L), or rhabdomyolysis; cre-
The objective of this study was to assess the prevalence atine kinase greater than 2000 mg/dL; mortality after ther-
of and risk factors for acute kidney injury (AKI) associated apy discontinuation; and hospital mortality or discharge.
with the use of intravenous polymyxin B sulfate in patients
hospitalized at a tertiary university hospital. DEFINITION OF ACUTE KIDNEY INJURY
AKI was identified in patients with baseline SCr lower
Methods than 1.5 mg/dL, when SCr levels increased to 1.8 mg/dL
This study was carried out at a tertiary university hospi- or greater. In patients with baseline SCr greater than or
tal with 700 beds, Hospital de Base, São José do Rio Preto equal to 1.5 mg/dL, AKI was identified when SCr levels
Medical School, São José do Rio Preto, Brazil. It was ap- increased to more than 50% above baseline or there was
proved by the local Ethics Committee (São José do Rio need for dialysis.13 Patients with baseline SCr greater than
Preto Medical School), which waived the need for in- 4 mg/dL, patients with prerenal AKI (normalization of cre-
formed consent due to its retrospective nature. atinine levels within 48 hours after volume replacement or
Patients using polymyxin B (Bedford Laboratories, adjustment of cardiac output), those with obstructive renal
Bedford, OH) were identified by the antibiotic dispensing failure (diagnosed by the clinical picture or by kidney and
urinary tract ultrasound), and patients on dialysis were ex-
list of the hospital for 5 consecutive years (2000 –2004).
cluded.
The indication for the antibiotic was determined by the
patient’s attending physician and by a physician of the
Hospital Infection Control Committee. Patients using Statistical Analysis
polymyxin B sulfate for at least 3 consecutive days, with Data are presented as mean ± SD, median (in the case
serum creatinine (SCr) measurement 15 days or less prior of asymmetric data), or percent (%). The groups with
to treatment start and at least another creatinine measure- AKI and no AKI were compared by univariate analysis
ment during treatment with polymyxin, were included. using the χ2 test or Fisher’s exact test or Student’s 2-
Patients with baseline creatinine levels greater than 4.0 tailed nonpaired t-test, according to the kind and fre-
mg/dL or patients undergoing dialysis at the beginning of quency of the studied variable. The risk factors for the
antibiotic therapy were excluded. In patients undergoing development of AKI were detected by binary logistic re-
more than one course of therapy with this antibiotic, only gression. The initial model included all of the variables
the first was considered for the purposes of this study. All associated with AKI in the univariate analysis: age;
of the patients received the antibiotic by 24-hour continu- baseline SCr greater than or equal to 1.5 mg/dL and less
ous intravenous infusion diluted in NaCl 0.9%, as recom- than or equal to 4.0 mg/dL; hypotension; use of dopa-
mended at that time by the Hospital Infection Control mine, vancomycin, amphotericin B, furosemide, raniti-
Committee. dine, ACE inhibitors, amikacin, and dobutamine; dia-
The following hospital chart data were analyzed: sex; betes; oliguria; mechanical ventilation; and infection site
age; race; estimated weight; clinical or surgical pathology; (blood, catheter, abdomen). Dopamine and nora-
use of mechanical ventilation; infection site; bacteria re- drenaline use was considered to be use of vasoactive
sponsible for the infection requiring the use of polymyxin drugs, and the use of vancomycin, amphotericin B,
B; hospitalization time; hospitalization time to positive cul- amikacin, iodinated contrast, tacrolimus, cyclosporine,
ture; use of nephrotoxic drugs simultaneously with and NSAIDs was considered to be a single variable (use
polymyxin B (vancomycin, aminoglycosides, ampho- of nephrotoxic drugs). Several models were tested by an
tericin B, cephalothin, iodinated contrast, cyclosporine, intentional method (nonautomated), and the most ade-
tacrolimus, nonsteroidal antiinflammatory drugs [NSAIDs]); quate, according to clinical criteria (ie, including vari-
I The Annals of Pharmacotherapy I 2009 December, Volume 43 theannals.com

3. Acute Kidney Injury Associated with Parenteral Polymyxin B
ables considered more clinically relevant for AKI devel- ment with the study antibiotic ranged from 3 days (mini-
opment), was selected. mum for assessment of adverse effects) to 43 days, with a
Dependence analysis (DEPANA) was used for the in- median of 11.5 days. The daily mean dose of antibiotic
fection site variable to assess the probability of progression was 96.7 mg/day (median, 100 mg), ranging from 50 mg
to AKI according to the infection site. DEPANA is similar (500,000 U) in patients with abnormal renal function to
to correspondence analysis,14 which could not be used in this 150 mg (1,500,000 U) in patients with normal renal func-
case because AKI has only 2 categories. DEPANA uses tion. Mean baseline SCr was 1.04 ± 0.74 mg/dL (0.3–3.7
Hellinger’s distance instead of Pearson’s χ2 divergent for ana- mg/dL), and the mean highest creatinine level during ther-
lyzing association in contingency tables. It reveals and estab- apy was 1.59 ± 1.21 mg/dL (0.4–6.4 mg/dL). These data
lishes the hierarchy of latent variables (factors) that explains are summarized in Table 1. Of these 114 patients, 73
the way the categories of crossed variables associate with (64%) simultaneously received other nephrotoxic drugs
each other. Each factor synergistically combines AKI and no such as vancomycin (n = 55), amphotericin B (n = 23), and
AKI with the categories of infection focuses and explains a amikacin (n = 29), and 10 received other nephrotoxic
portion of total association (here named total dependence). drugs (iodinated contrast, 5; NSAIDs, 2; NSAID and gan-
Factor 1, which explains 95.5% of total dependence, per- ciclovir, 1; ganciclovir alone, 1; tacrolimus, 1; cyclo-
ceives the difference between AKI and no AKI. sporine, 1). Sixty patients (53%) received polymyxin B
The Minitab software program, version 12.22 (Minitab and furosemide simultaneously and 36 patients (32%) re-
Inc., State College, PA, www.minitab.com) was used. ceived ACE inhibitors and polymyxin B simultaneously.
Mortality at the end of the treatment was 32.5% (37 pa-
tients), and hospital mortality was 61% (70 patients).
Results The most frequently found bacteria were P. aeruginosa
GENERAL CHARACTERISTICS OF THE STUDY (83%), followed by A. baumannii (11%). One patient was
POPULATION infected by Klebsiella pneumoniae, one patient by Cit-
robacter spp., and another by P. aeruginosa and A. bau-
The hospital charts of 168 patients, who received a total mannii; in 3 other patients, the antibiotic was administered
of 178 courses of polymyxin B therapy in 5 years, were empirically.
studied. One hundred fourteen patients met the inclusion The major bacteria isolation sites were: urinary (42%),
criteria, and 54 patients were excluded according to the pulmonary (27%), blood (13%), abdominal (8%), soft tis-
study criteria (Figure 1). sue and bone (5%), and central venous catheter (2%); in
Most (60%) of the patients were male, the large majority 3%, the infection site was not identified (empirical use of
(92%) were white, and median age was 59 years (range antibiotic).
14 –86 y). Most of the studied population consisted of
severely ill patients hospitalized in the ICU and in semi-in-
tensive care units (70% and 8%, respectively) and had
clinical diseases not related to surgical conditions (60%). Table 1. Description of Study Population
Mean hospitalization time was 66.5 days (12– 428 days). Pts., Mean Minimum Maximum
The median duration of culture positivity was 26 days. Variable n ± SD Median Value Value
There was a large variation in the time to detect infection Age, y 114 56 ± 18 59 14 86
with multiresistant bacteria after initial hospitalization, Hospitalization, 114 66 ± 53 54 12 428
with some patients presenting a positive culture on the first days
day, whereas in others, a positive culture was obtained Hospitalization 111 31.2 ± 25 26 0 139
time to positive
only after 140 days of hospitalization. Duration of treat- culture, days
Duration of 114 12.5 ± 6.3 11.5 3 43
therapy, days
Creatinine level,
mg/dL
baseline 114 1.04 ± 0.74 0.8 0.3 3.7
highest 114 1.59 ± 1.21 1.1 0.4 6.4
at discharge/ 67 1.35 ± 1.15 1 0.3 6.5
death
Dose, mga
daily 114 96.7 ± 20.3 100 50b 150
total 114 1228 ± 838 1025 1500 6450
a
One milligram = 10,000 units.
b
Dose used for patients with previous renal failure, as defined by the at-
tending physicians.
Figure 1. Included and excluded patients. SCr = serum creatininer.
theannals.com The Annals of Pharmacotherapy I 2009 December, Volume 43 I

4. CAC Mendes et al.
PREVALENCE OF AKI IN THE STUDY POPULATION LOGISTIC REGRESSION FOR THE IDENTIFICATION OF
AKI-RELATED RISK FACTORS
The prevalence of AKI was 22% (n = 25). Conversely,
78% of patients (n = 89) maintained stable renal function The variables significantly and independently associated
during antibiotic administration. In the AKI group, 32% of with the progression to AKI in patients using intravenous
the patients were oliguric. polymyxin were presence of previously abnormal renal
function (OR 3.51; 95% CI 1.11 to 11.2; p = 0.033); use of
COMPARISON BETWEEN THE AKI AND NON-AKI GROUPS
In the AKI group, the baseline SCr was 1.6 ± 1.1 mg/dL
and the highest SCr during treatment was 3.4 ± 1.2 mg/dL Table 2. Univariate Analysis of Parameters Comparing
(an increase of 103%). In the non-AKI group, the baseline Patients With and Without AKI
SCr was 0.9 ± 0.4 mg/dL and the highest SCr was 1.1 ±
AKI, % No AKI, % p
0.5 mg/dL (an increase of 22%). Parameter (n = 25) (n = 89) Value
The patients who progressed to AKI were older (62 ± Male 57 72 0.18
15 vs 54 ± 19 y in non-AKI; p = 0.02); had higher baseline Age, y (mean ± SD) 62 ± 15 54 ± 19 0.02
SCr (1.6 ± 1.1 vs 0.9 ± 0.4 mg/dL; p = 0.002); had more Abnormal baseline creatininea 40 12 0.001
hypotension (50% vs 27%; p = 0.016); used vasoactive Mechanical ventilation 92 58 0.002
drugs more often (47% vs 12%; p < 0.001); used more me- Furosemide use 80 45 0.001
chanical ventilation (92% vs 58%; p = 0.002); and used Hypotensionb 50 27 0.016
more nephrotoxic drugs (88% vs 57%; p = 0.002) and Oliguria 32 1 <0.001
furosemide (80% vs 45%; p = 0.001). DEPANA showed Ranitidine use 84 74 0.3
that progression to renal failure was significantly more ACE inhibitor use 20 35 0.16
probable when the bacteria were isolated in the abdomen, Hypertension 32 38 0.57
catheter, or blood (p = 0.004), respectively. Diabetes mellitus 32 23 0.4
Vancomycin use 68 43 0.012
Of the 93 patients with normal baseline renal function
Amphotericin B use 36 16 0.013
(SCr <1.5 mg/dL), 15 (16%) progressed to AKI, whereas
Amikacin use 36 21 0.06
10 of the 21 (47.6%; p = 0.001) patients with previous re-
Dobutamine use 16 10 0.2
nal dysfunction developed AKI.
Dopamine use 52 16 <0.0005
The AKI and non-AKI groups were similar as to the Noradrenaline use 16 10 0.2
presence of diabetes mellitus, respectively (32% vs 23%; p Other nephrotoxic drugs use 88 57 0.005
= 0.4), hypertension (32% vs 38%; p = 0.57), use of raniti- Vasoactive drugs use 47 12 0.0005
dine (84% vs 74%; p = 0.3), dobutamine (16% vs 10%; p (Nora or Dopa)
= 0.2), and ACE inhibitors (20% vs 35%; p = 0.16). There Death 92 53 <0.0005
was also no difference as to the sex (p = 0.18). ACE = angiotensin-converting enzyme; AKI = acute kidney injury.
The mortality of the patients who progressed to AKI a
Abnormal serum creatinine values defined as greater than or equal
was significantly higher than that of the non-AKI group to 1.5 mg/dL and less than or equal to 4.0 mg/dL.
b
Systolic blood pressure less than 90 mm Hg.
(92% vs 53%, respectively; p < 0.001). These data are
summarized in Table 2.
Twelve of the 25 patients with AKI died dur-
ing the treatment and 13 patients completed the
therapy with polymyxin B (between 6 and 33
days). Of these 13, 8 had a baseline SCr less
than 1.5 mg/dL, and 7 of the 8 had a creatinine
measurement after the discontinuation of the an-
tibiotic. Of these 7, only 2 had normalized crea-
tinine levels (SCr <1.5 mg/dL) and they were
the only survivors in this group. Nine patients re-
ceived dialysis. Of these, 8 had abnormal base-
line creatinine values and 1 had normal baseline
creatinine.
All of the patients with previous renal dys-
function who progressed to AKI and all of the pa-
tients undergoing dialysis died. The progression Figure 2. Evolution of patients with acute kidney injury. AKI = acute kidney injury;
of the patients with AKI is shown in Figure 2. BSCr = baseline serum creatinine; SCr = serum creatinine.
I The Annals of Pharmacotherapy I 2009 December, Volume 43 theannals.com

5. Acute Kidney Injury Associated with Parenteral Polymyxin B
vasoactive drugs (OR 3.03; 95% CI 1.02 to 9.04; p = 0.047); A limitation of our study, as well as most of the previous
and infection site of abdomen, catheter, or blood (OR 3.82; ones, is lack of a control group of patients not exposed to
95% CI 1.21 to 12.13; p = 0.023). There was also a trend to polymyxin B. Considering the use of polymyxin B as sal-
progression into AKI when polymyxin B was given concur- vage therapy for multiresistant organisms and the retro-
rently with other nephrotoxic drugs (OR 3.88; 95% CI 0.96 spective nature of the study, the selection of an adequate
to 15.61; p = 0.057). Table 3 summarizes these data. control group was not feasible. It should be acknowledged
This is one of the few studies assessing the safety of in- that the majority of patients used other potentially nephro-
travenous polymyxin B sulfate, and to the best of our toxic drugs concurrently with polymyxin B and that some
knowledge, it had the largest number of patients so far. In of the risk factors, such as need for vasoactive drugs, are
the past 10 years, almost all publications have referred to plausible for any population of critically ill patients. So it is
colistimethate sodium, the sulfated form of polymyxin E possible that the observed prevalence of AKI overestimat-
(colistin).15 Colistimethate sodium has supposedly been ed the true rate of polymyxin-related AKI.
considered to be less nephrotoxic than polymyxin B. How- The association between previous renal failure and
ever, a recent review emphasized that there are more simi- polymyxin B21 or colistimethate sodium22-24,26,28,34,39 nephro-
larities (mechanism of action, antimicrobial spectrum, clin- toxicity was previously observed. In contrast, other studies
ical use, toxicity) than differences (chemical structure, an- did not confirm this result, some using polymyxin B18,19
timicrobial power, dosage, pharmacokinetic properties) and others using colistimethate sodium.30,31 One of the pos-
between the 2 drugs.16 sible explanations for the association of previous renal dys-
function with AKI in our study might be related to the ad-
Discussion ministration of an excessive polymyxin B dose in patients
with previous renal injury. This excessive dose might be a
Similar to populations in previous studies, this popula- result of the lack of experience with the use of this antibi-
tion predominantly included ICU patients.17-20 Of the 114 otic and of imprecise information for the correction of the
patients studied, 22% developed AKI, whereas in the other dose in the manufacturer’s package insert.40 Another factor
studies, the frequency of polymyxin B–associated AKI potentially associated with a high incidence of AKI in our
ranged from 6% to 54% (the number of patients analyzed patients is an inadequate assessment of the patient’s body
in these studies ranged from 11 to 74).17-21 In studies as- weight, since it was often estimated by the physician pre-
sessing patients receiving colistimethate sodium, the scribing the antibiotics. Many of the patients were in the
prevalence of AKI also ranged broadly, from 0% to ICU and could not be weighed. An additional variable that
37%.22-36 This large disparity in data is probably due to the might have contributed to the development of nephrotoxic-
fact that the studied populations were not homogeneous ity is the 24-hour continuous infusion of the drug in pa-
and that definitions for the diagnosis of AKI varied. In our tients with decreased baseline renal function. This regimen
study, we used an AKI definition previously published by was adopted because it was recommended by the Hospital
our group.13 We did not use the more recent definitions for Infection Control Committee at the time the patients were
AKI, such as RIFLE or AKIN, because they had not yet treated. An in vitro experiment suggested that colistin-in-
been published when we started data collection.37,38 duced cell toxicity is time dependent.41 On the other hand,
Few studies have assessed the risk factors for AKI in a recent study comparing polymyxin B continuous infu-
patients receiving polymyxin B or colistimethate sodium. sion and twice-daily administration showed the same fre-
In our study, with logistic regression analysis, we found quency of nephrotoxicity in both regimens.42
previous renal dysfunction, use of vasoactive drugs, and The use of vasoactive drugs (noradrenaline, dopamine)
site of infection (abdomen, catheter, blood) to be indepen- indicates the presence of shock and/or hemodynamic insta-
dent variables for risk of AKI. bility. This situation is classically associated with drug
nephrotoxicity, since it causes renal hypoperfusion and is-
chemia. In fact, many nephrotoxic drugs interfere with the
synthesis of energy molecules, increasing the risk of kid-
Table 3. Logistic Regression Analysis
ney injury during an ischemic episode.
Odds Something that was not observed in prior studies was
Parameter Ratio 95% CI p Value
the location of infection sites as a risk factor for AKI. The
Baseline serum creatinine 3.51 1.11 to 11.2 0.033
≥1.5 mg/dL and ≤4 mg/dL
presence of abdominal infection, usually related to major
Vasoactive drug use 3.03 1.02 to 9.04 0.047 pathologies, might lead to a more severe presentation of
(noradrenaline and/or dopamine) sepsis, which is also a well-known factor in progression to
Infection site 3.82 1.21 to 12.13 0.023 AKI.43
(abdomen, catheter, blood)
Only 2 of the 25 patients who developed AKI survived,
Other nephrotoxic drug use 3.88 0.96 to 15.61 0.057
and they were the only ones who recovered renal function.
theannals.com The Annals of Pharmacotherapy I 2009 December, Volume 43 I

6. CAC Mendes et al.
Polymyxin therapy was not discontinued in these 2 pa- Financial disclosure: Dr. Burdmann is partially supported by grants
from the National Council for Scientific and Technological Develop-
tients upon onset of renal dysfunction. They received ment (CNPq, Brazil).
polymyxin B (1,000,000 units/day) for 25 and 35 days, re- Part of this work was presented at the American Society of Nephrol-
spectively. Other studies showed the recovery of renal ogy meeting in 2006 and published as an abstract in the Journal of
function in most of the patients who survived after drug the American Society of Nephrology, 2006;17:164A.
administration discontinuation, but causal links among sur-
vival, withdrawal of a drug, and renal function improve- References
ment are impossible to assess at this time.7,19 1. Falagas EM, Kasiakou SK. Colistin: the revival of polymyxins for the
In the present study, we found a mortality rate of 32.4% management of multidrug-resistant Gram-negative bacterial infections.
at the end of treatment and a hospital mortality rate of Clin Infect Dis 2005;40:1333- 41. DOI 10.1086/429323
2. Koike M, Iida K, Matsuo T. Electron microscopy studies on the mode of
61.4%. Increased mortality rates in patients treated with
action of polymyxin. J Bacteriol 1969;97:448-52.
polymyxin B were also observed by other authors.17,20 3. Hancock RE, Chapple DS. Peptide antibiotics. Antimicrob Agents
Studies in patients receiving colistimethate sodium also re- Chemother 1999;43:1317-23.
ported a high mortality rate, ranging from 40% to 61%.30,33 4. Storm DR, Rosenthal KS, Swanson PE. Polymyxin and related peptide
antibiotics. Annu Rev Biochem 1977;46:723-63.
The high rate of mortality found in our study might be ex-
DOI 10.1146/annurev.bi.46.070177.003451
plained by the severity of illness in these patients (the ma- 5. Gales AC, Jones RN, Sader HS. Global assessment of the antimicrobial
jority hospitalized in the ICU, under mechanical ventila- activity of polymyxin B against 54,731 clinical isolates of Gram-nega-
tion) and by the large number of patients infected by P. tive bacilli: report from the SENTRY antimicrobial surveillance pro-
gramme (2001–2004). Clin Microbiol Infect 2006;12:315-21.
aeruginosa. Because the antibiotic was usually only re- DOI 10.1111/j.1469-0691.2005.01351.x
leased by the Hospital Infection Control Committee after 6. Bratu S, Tolaney P, Karumudi U, et al. Carbapenemase-producing Kleb-
the confirmation of bacteria that were susceptible only to siella pneumoniae in Brooklyn, NY: molecular epidemiology and in vit-
ro activity of polymyxin B and other agents. J Antimicrob Chemother
polymyxin B, treatment might have been delayed 3 to 4
2005;56:128-32. DOI 10.1093/jac/dki175
days, potentiating the occurrence of negative results in pa- 7. Falagas ME, Bliziotis IA, Kasiakou SK, Samonis G, Athanassopoulou P,
tients with severe infection. Patients who progressed to Michaloupoulos A. Outcome of infections due to pandrug-resistant
AKI had a 2-fold increase in mortality when compared (PDR) Gram-negative bacteria. BMC Infect Dis 2005;5:24.
DOI 10.1186/1471-2334-5-24
with those who had preserved renal function. This finding 8. Horton J, Pankey GA. Polymyxin B, colistin, and sodium colistimethate.
is consistent with previous studies, which also found a Med Clin North Am 1982;66:135- 42.
higher mortality rate in patients with polymyxin B–related 9. Archibald L, Phillips L, Monnet D, McGowan JE, Tenove F, Gaynes R.
AKI.18,24 Antimicrobial resistance in isolates from inpatients and outpatients in the
United States: increasing importance of the Intensive Care Unit. Clin In-
In conclusion, polymyxin B had a prevalence of nephro- fect Dis 1997;24:211-5.
toxicity similar to that for colistimethate sodium described 10. Levin AS, Mendes CM, Sinto SI, et al. An outbreak of multiresistant
in several published studies. Patients with previous renal Acinetobacter baumanii in a university hospital in São Paulo, Brazil. In-
fect Control Hosp Epidemiol 1996;17:366-8.
dysfunction seem to be more vulnerable to nephrotoxicity.
11. Evans ME, Feola DJ, Rapp RP. Polymyxin B sulfate and colistin: old an-
The use of vasoactive drugs, as well as abdominal, blood, tibiotics for emerging multiresistant gram-negative bacteria. Ann Phar-
and catheter infection sites, were identified as independent macother 1999;33:960-7. DOI 10.1345/aph.18426
variables for the development of AKI. Patients who have 12. Pedersen MF, Pedersen JF, Adsen PO. A clinical and experimental com-
parative study of sodium colistimethate and polymyxin B sulfate. Invest
these characteristics and/or previous renal failure and re- Urol 1971;9:234-7.
ceive polymyxin B must have renal function monitored 13. Santos WJ, Zanetta DM, Pires AC, Lobo SM, Lima EQ, Burdmann EA.
carefully and receive intensive measures to prevent the de- Patients with ischaemic, mixed and nephrotoxic acute tubular necrosis in
the intensive care unit—a homogeneous population (abstract)? Crit Care
velopment of renal failure. The development of nephrotox-
2006;10:R68. DOI 10.1186/cc4904
icity was associated with a poor prognosis, and a signifi- 14. Greenacre MJ, ed. Theory and applications of correspondence analysis.
cant number of patients who survived did not totally re- London: AC Press, 1984.
cover from renal failure. 15. Falagas ME, Kasiakou SK. Toxicity of polymyxins: a systematic review
of the evidence from old and recent studies (abstract). Crit Care 2006;10:
R27. DOI 10.1186/cc3995
Carlos AC Mendes MD, Assistant Professor, Internal Medicine
16. Kwa A, Kasiakou SK, Tam VH, Falagas ME. Polymyxin B: similarities
Department, Division of Internal Medicine, Hospital de Base, São
José do Rio Preto Medical School, São José do Rio Preto, São to and differences from colistin (polymyxin E). Expert Rev Anti Infect
Paulo, Brazil Ther 2007;5:811-21. DOI 10.1586/14787210.5.5.811
José A Cordeiro PhD, Professor of Statistics, Epidemiology and 17. Sobieszczyk ME, Furuya EY, Hay CM, et al. Combination therapy with
Public Health Department, São José do Rio Preto Medical School polymyxin B for the treatment of multidrug-resistant Gram-negative res-
Emmanuel A Burdmann MD PhD, Professor of Medicine, Divi- piratory tract infections. J Antimicrob Chemother 2004;54:566-9.
sion of Nephrology, Hospital de Base, São José do Rio Preto Med- DOI 10.1093/jac/dkh369
ical School 18. Ouderkirk JP, Nord JA, Turett GS, Kislak JW. Polymyxin B nephrotoxi-
Reprints: Dr. Burdmann, São José do Rio Preto Medical School, city and efficacy against nosocomial infections caused by multiresistant
Av. Brigadeiro Faria Lima, 5416, São José do Rio Preto, São Paulo, Gram-negative bacteria. Antimicrob Agents Chemother 2003;47:2659-62.
Brazil 15090-000, fax 55-17-32162227, burdmann@famerp.br DOI 10.1128/AAC.47.8.2659-2662.2003
I The Annals of Pharmacotherapy I 2009 December, Volume 43 theannals.com

7. Acute Kidney Injury Associated with Parenteral Polymyxin B
19. Holloway KP, Rouphael NG, Wells JB, King MD, Blumberg HM. 36. Kallel H, Hergafi L, Bahloul M, et al. Safety and efficacy of colistin
Polymyxin B and doxycycline use in patients with multidrug-resistant compared with imipenem in the treatment of ventilator-associated pneumo-
Acinetobacter baumannii infections in the intensive care unit. Ann Phar- nia: a matched case–control study. Intensive Care Med 2007;33:1162-7.
macother 2006;40:1939- 45. Epub 3 Oct 2006. DOI 10.1345/aph.1H353 DOI 10.1007/s00134-007-0675-2
20. Furtado GH, d’Azevedo PA, Santos AF, Gales AC, Pignatari AC, Medeiros 37. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P; Acute Dialysis
EA. Intravenous polymyxin B for the treatment of nosocomial pneumonia Quality Initiative workgroup. Acute renal failure—definition, outcome
caused by multidrug-resistant Pseudomonas aeruginosa. Int J Antimicrob measures, animal models, fluid therapy and information technology
Agents 2007;30:315-9. DOI 10.1016/j.ijantimicag.2007.05.017 needs: the Second International Consensus Conference of the Acute
21. Pastewski AA, Caruso P, Parris AR, et al. Parenteral polymyxin B use in Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204-12.
patients with multidrug-resistant gram-negative bacteremia and urinary DOI 10.1186/cc2872
tract infections: a retrospective case series. Ann Pharmacother 2008;42: 38. Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: re-
1177-87. Epub 29 Jul 2008. DOI 10.1345/aph.1K346 port of an initiative to improve outcomes in acute kidney injury (ab-
22. Levin AS, Barone AA, Penço J, et al. Intravenous colistin as therapy for stract). Crit Care 2007;11:R31. DOI 10.1186/cc5713
nosocomial infections caused by multidrug-resistant Pseudomonas 39. Wolinsky E, Hines JD. Neurotoxic and nephrotoxic effects of colistin in
aeruginosa and Acinetobacter baumannii. Clin Infect Dis 1999;28:1008- patients with renal disease. N Engl J Med 1962;266:759-62.
11. DOI 10.1086/514732 40. Product information. Polymyxin B sulfate sterile, 500,000 units. Bed-
23. Koch-Weser J, Sidel VW, Federman EB, Kanarek P, Finer DC, Eaton AE. ford, OH: Bedford Laboratories, July 2009.
Adverse effects of sodium colistimethate. Manifestations and specific reac- 41. Lewis JR, Lewis SA. Colistin interactions with the mammalian urotheli-
tion rates during 317 courses of therapy. Ann Intern Med 1970;72:857-68. um. Am J Physiol Cell Physiol 2004;286:C913-22.
24. Michalopoulos AS, Tsiodras S, Rellos K, Mentzelopoulos S, Falagas DOI 10.1152/ajpcell.00437.2003
ME. Colistin treatment in patients with ICU-acquired infections caused 42. Teng CB, Koh PT, Lye DC, Ang BS. Continuous versus intermittent in-
by multiresistant Gram-negative bacteria: the renaissance of an old an- fusion of polymyxin B in the treatment of infections caused by mul-
tibiotic. Clin Microbiol Infect 2005;11:115-21. tidrug-resistant Gram-negative bacteria. Int J Antimicrob Agents 2008;
DOI 10.1111/j.1469-0691.2004.01043.x 31:80-2. DOI 10.1016/j.ijantimicag.2007.08.004
25. Ledson MJ, Gallagher MJ, Cowperthwaite C, Convery RP, Walshaw MJ. 43. Leblanc M, Kellum JA, Gibney RT, Lieberthal W, Tumlin J, Mehta R.
Four years’ experience of intravenous colomycin in an adult cystic fibro- Risk factors for acute renal failure: inherent and modifiable risks. Curr
sis unit. Eur Respir J 1998;12:592-4. Opin Crit Care 2005;11:533-6.
DOI 10.1183/09031936.98.12030592
26. Garnacho-Montero J, Ortiz-Leyba C, Jimenez-Jimenez FJ, et al. Treat-
ment of multidrug-resistant Acinetobacter baumannii ventilator-associat-
ed pneumonia (VAP) with intravenous colistin: a comparison with Frecuencia e Factores de Riesgo Para el Desarrollo de Lesión Renal
imipenem-susceptible VAP. Clin Infect Dis 2003;36:1111-8.
Aguda Asociada a uso Intravenoso de Polimixina B
DOI 10.1086/374337
27. Reina R, Estenssoro E, Saenz G, et al. Safety and efficacy of colistin in CAC Mendes, JA Cordeiro, y EA Burdmann
Acinetobacter and Pseudomonas infections: a prospective cohort study. Ann Pharmacother 2009;43:xxxx.
Int Care Med 2005;31:1058-65. DOI 10.1007/s00134-005-2691-4
28. Bassetti M, Repetto E, Righi E, et al. Colistin and rifampicin in the treat-
EXTRACTO
ment of multidrug-resistant Acinetobacter baumannii infections. J An-
timicrob Chemother 2008;61:417-20. DOI 10.1093/jac/dkm509 TRASFONDO: El principal efecto adverso de la polimixina B es nefro-
29. Falagas ME, Rizos M, Bliziotis IA, Rellos K, Kasiakou SK, Michalo- toxicidad, pero hay pocos datos sobre lesión renal asociada a polimixinas.
poulos A. Toxicity after prolonged (more than four weeks) administra- OBJETIVO: El objetivo principal de ese estudio fue evaluar la frecuencia y
tion of intravenous colistin (abstract). BMC Infect Dis 2005;5:1. los factores de riesgo para insuficiencia renal aguda (IRA) en pacientes
DOI 10.1186/1471-2334-5-1 tratados con polimixina B.
30. Falagas ME, Fragoulis KN, Kasiakou SK, Sermaidis GJ, Michalopoulos A. MÉTODOS: La población de estudio incluyó 114 pacientes que han recibido
Nephrotoxicity of intravenous colistin: a prospective evaluation. Int J An- por lo menos 3 días consecutivos de polimixina B intravenoso y tenían
timicrob Agents 2005;26:504-7. DOI 10.1016/j.ijantimicag.2005.09.004 creatinina sérica basal (SCr) y al mínimo otra creatinina sérica durante el
31. Kasiakou SK, Michalopoulos A, Soteriades ES, Samonis G, Sermaides
tratamiento. IRA fue definida cuando la SCr presentase aumento de más
que 1.8 mg/dL en pacientes con SCr basal < 1.5 mg/dL o si hubiese au-
GJ, Falagas ME. Combination therapy with intravenous colistin for man-
mento ≥50% en la SCr si la SCr basal ya presentase valor > 1.8 mg/dL o
agement of infections due to multidrug-resistant Gram-negative bacteria también si hubiese necesidad de diálisis.
in patients without cystic fibrosis. Antimicrob Agents Chemother 2005;
RESULTADOS: IRA se desarrollo en 22% de los pacientes. Esos pacientes
49:3136-46. DOI 10.1128/AAC.49.8.3136-3146.2005
eran más viejos, presentaban una creatinina sérica basal más elevada,
32. Kallel H, Bahloul M, Hergafi L, et al. Colistin as a salvage therapy for noso- una mayor frecuencia de SCr basal ≥1.5, usaron otras drogas nefro-
comial infections caused by multidrug-resistant bacteria in the ICU. Int J tóxicas y furosemida más frecuentemente, utilizaron más drogas vaso-
Antimicrob Agents 2006;28:366-9. DOI 10.1016/j.ijantimicag.2006.07.008 activas y ventilación mecánica, y los sitios de infección más frecuentes
33. Hachem RY, Chemaly RF, Ahmar CA, et al. Colistin is effective in treat- fueron el abdomen, sangre y catéteres intravenosos. Los pacientes que
ment of infections caused by multidrug-resistant Pseudomonas aerugi- evolucionaron para IRA han tenido mayor mortalidad (92 vs 53%; p <
nosa in cancer patients. Antimicrob Agents Chemother 2007;51:905-11. 0.0005). Regresión logística identificó SCr basal alterada (OR 3.71),
DOI 10.1128/AAC.01015-06 necesidad de drogas vasoactivas (OR 4.38) y sitios de infección en el
34. Pintado V, San Miguel LG, Grill F, et al. Intravenous colistin sulphome-
abdomen, sangre o catéter (OR 3.06) como factores de riesgo independi-
entes para evolución para IRA.
thate sodium for therapy of infections due to multidrug-resistant Gram-
negative bacteria. J Infect 2008;56:185-90. CONCLUSIONES: Pacientes que desarrollaron IRA han tenido una elevada
DOI 10.1016/j.jinf.2008.01.003 mortalidad. Polimixina B debe ser utilizada con mucha precaución en
pacientes con SCr alterada, en los que presentan necesidad de uso de
35. Oliveira MS, Prado GV, Costa SF, Grinbaum RS, Levin AS. Ampi-
drogas vasoactivas y cuando el foco de infección está localizado en el
cillin/sulbactam compared with polymyxins for the treatment of infec- abdomen, sangre, y catéter.
tions caused by carbapenem-resistant Acinetobacter spp. J Antimicrob
Chemother 2008;61:1369-75. DOI 10.1093/jac/dkn128 Traducido por Emmanuel A Burdmann
theannals.com The Annals of Pharmacotherapy I 2009 December, Volume 43 I

8. CAC Mendes et al.
La Prévalence et les Facteurs de Risque Associés à une Atteinte RÉSULTATS: Une toxicité rénale aigue a été notée chez 22% des patients
Rénale Induite par la Polymyxine B de l’étude. L’observation du profil de ces patients a permis l’identifica-
tion de certaines caractéristiques communes: un âge élevé, une valeur
CAC Mendes, JA Cordeiro, et EA Burdmann prétraitement élevée de la créatinine sérique, la présence plus fréquente
Ann Pharmacother 2009;43:xxxx. d’une ventilation artificielle, l’identification d’un site d’infection [qu’il
soit au niveau abdominal, sanguin ou mécanique (cathéter)], et l’utilisation
accrue de furosémide, de médicaments ayant des effets néphrotoxiques
RÉSUMÉ ou de médicaments vasoactifs. Les patients chez qui une néphrotoxicité
INTRODUCTION: La néphrotoxicité constitue le principal effet indésirable était documentée avaient une plus grande mortalité (92% vs 53%; p <
relié à l’administration parentérale de la polyxymine B. Peu d’informations 0.0005). Une analyse de régression logistique a permis d’identifier les
sont toutefois disponibles quant à la prévalence et les facteurs de risque facteurs de risque indépendants suivants: une valeur élevée prétraitement de
associés à une telle toxicité. la créatinine sérique (rapport des cotes de 3.71), la présence de médi-
caments vasoactifs (rapport des cotes de 4.38) et l’identification d’un
OBJECTIF: L’objectif de cette étude était d’évaluer la prévalence et les site d’infection [abdominal, sanguin ou mécanique (cathéter)- rapport
facteurs de risque associés à la néphrotoxicité induite par l’utilisation de des cotes de 3.06].
la polymyxine B.
CONCLUSIONS: Les patients qui développent une néphrotoxicité suite à
MÉTHODOLOGIE: Une population de 114 patients ayant reçu la polymyxine l’administration parentérale de la polymyxine B ont un plus haut taux de
B sur 3 jours consécutifs et plus ont été inclus dans cette étude rétro- mortalité. La polymyxine B devrait être utilisée avec prudence chez les
spective. Tous les patients devaient avoir une valeur documentée pré- patients dont les valeurs prétraitement de la créatinine sérique sont élevées,
traitement de leur créatinine sérique ainsi qu’au moins une autre valeur chez les patients requérant l’utilisation concomitante de médicaments
durant le traitement. Une toxicité rénale aigue était définie par une aug- vasoactifs ou chez qui une infection d’origine abdominale, sanguine ou
mentation de la créatinine sérique de 1.8 mg/dL chez les patients dont la mécanique (cathéter) a été diagnostiquée.
créatinine prétraitement était inférieure à 1.5 mg/dL ou une augmentation
de 50% de la valeur prétraitement de la créatinine sérique pour tous les Traduit par Sylvie Robert
autres patients.
I The Annals of Pharmacotherapy I 2009 December, Volume 43 theannals.com