Version 9
August 6, 2013
WP 1
D1.1
Multicenter open-label RCT to compare colistin alone vs.
colistin plus meropenem
BACKGROUND
Colistin has resurged in the last decade for the treatment of multidrug-resistant Gram-negative bacteria due to lack of other antibiotics. This antibiotic, firstly discovered in 1947, belongs to the polymyxin family and is a mixture of polymyxin E1 and polymyxin E2. The polymyxins act by disrupting the cell membrane. They have a strong positive charge and a hydrophobic acyl chain that confer them a high binding affinity to lipopolysaccharide (LPS) molecules. They interact electrostatically with these molecules and competitively displace divalent cations from them, causing disruption of the membrane. [1, 2] Electron microscopy studies show protrusions and bleb formation of the cell membrane with leakage of cell contents.[3-5] Colisitn is bacteriocidal; whether interaction with membranes is the cause of bacterial cell death is unknown. [1] Polymyxins also bind to the lipid A portion of the LPS and, in animal studies, block many of the biologic effects of endotoxin. [6] Colistin has broad in-vitro activity against Gram-negative bacteria, with the exception of Proteus spp., Providencia spp., Serratia spp., and rarer bacteria (Brucella spp., Edwardsiella spp., Pseudomonas mallei and Burkholderia cepacia). Breakpoints for susceptibility are defined for enterobacteriaceae, Acinetobacter spp. and Pseudmonas spp. The CLSI define 2 mg/L for all and EUCAST defines 2 mg/L for Acinetobacter sp. and enterobacteriaceae and 4 mg/L for Pseudomonas sp.[7]
Colistimethate sodium (CMS) is the preparation currently used for systemic treatment. CMS is a prodrug that undergoes spontaneous hydrolysis in-vivo or in aqueous solutions to the active drug, colistin. The existence of these two forms has complicated PK/PD studies, since old bioessays did not differentiate between the two forms, which have different half lives and modes of excretion. [8] To further complicate matters, colistin is measured using different units. One mg of CMS is equivalent to 12,500 international units (IU, where 1 IU is defined as the minimal concentration which inhibits the growth of E. coli 95 I.S.M in 1 ml broth at pH 7.2 [9]). One mg of colistin base activity (CBA, the unit of measurement used in the US formulation) is equivalent to 33,250 IU. Considering a 70kg adult, the classically maximal recommended daily dose of colistin in the US, 5 mg/kg CBA would translate to 11.5 mill IU, while in Europe 9 mill IU per day would translate to 3.9 mg/kg CBA.
Since its resurgence, observational studies have tried to examine the effectiveness of colistin. Although individual studies reported favourable results regarding both effectiveness and safety, a compilation of these studies shows higher mortality among patients treated with colistin or polymyxin B compared to patients given other antibiotics, mostly beta-lactams (Figure 1). [2] In most studies colistin was used in combination with other antibiotics, mainly carbapenems, and colistin was probably underdosed. In the largest study, conducted in Israel, colistin was given almost always as monotherapy at a mean dose of 6.1 ±2.3 MU/day and mortality was significantly higher with colistin when compared to carbapenems or ampicillin-sulbactam. [10] Pooling of adjusted results from multivariable analyses or matched studies shows similar results (Figure 2).
In the same comparative studies rates of nephrotoxicity were higher with colistin compared to other antibiotics (Figure 3). Rates of nephrotoxicity in recent studies designed to assess this outcome have ranged from 6-14% in some [11-15] to 32-55% in others[16-20]. The wide range of nephrotoxicity rates is explained at least partially by different definitions of renal failure. Both the daily dose [17, 20] and the total cumulative dose [15, 16, 21] have been associated with increased risk of nephrotoxicity. Among patients with colistin-induced nephrotoxicity between 0-1.5% [16, 20] to ~20% [14, 18, 19] required short-term renal replacement therapy. Studies monitoring patients up to 1-3 months after colistin last dose demonstrated reversibility of renal failure in at least 88% of patients [12, 16, 18]. The other feared toxicity of colistin is neurological. Manifestation range from dizziness, muscle weakness, paresthesias, hearing loss, visual disturbances and vertigo to confusion, hallucinations, seizures, ataxia, and neuromuscular blockade with apnea[22]. The latter manifestations are rare in clinical practice.
Studies currently focus on improving the efficacy and safety profile of colistin. A first step is the optimization of dosing and schedule of administration. Recent PK studies demonstrate that it takes about 36-48 hours for colistin (rather than CMS) to reach therapeutic concentrations in plasma (≥2 mg/L) using classical dosing in patients with normal renal function [23, 24]. Thus, a loading dose, equaling to about the total daily dose is currently recommended. Furthermore, these studies demonstrate that once or twice daily dosing is probably sufficient[8]. For example, targeting a colistin steady state level of 2.5 mg/liter for a patient with a creatinine clearance of 70 ml/min/1.73, requires 337.5 mg CBA per day (11.2 mill IU). [23] With higher creatinine clearance rates the dose increases further. A recent study reported on the clinical experience of treating critically ill patients with colistin using a 9 mill IU loading dose followed by 4.5 mill IU q12h for normal renal function. [25] A response rate of 82.1% (23/28) and nephrotoxicity of 17.8% (5/28) was reported.
The suboptimal efficacy of colistin and the nephrotoxicity associated with high dosing regimens has led to the search for combination therapies that might improve clinical success via better killing or inhibition of the pathogen, more rapid killing, killing or inhibition at lower drug concentrations, thus avoiding toxicity, and prevention of resistance selection or emergence. Combinations suggested with colistin include various beta-lactams, azithromycin, co-trimoxazole, rifampin, doxycycline, minocycline, tigecyclin, vancomycin, aminoglycosides, quinolones, fosfomycin and sulbactam[26]. Most of the clinical experience exists with carbapenems that are sometimes used alone or in addition to colistin for carbapenem-resistant infections when MICs are relatively above the susceptibility breakpoint, mainly for Acinetobacter baumannii, in the assumption that high dosing might overcome resistance, but few data support this practice. In a mouse model, intratracheal meropenem was significantly more effective than colistin for carbapenem-resistant Acinetobacter baumannii pneumoniawith an MIC for meropenem of 32 μg/ml[27]. The main rationale for combination therapy lies in the existence of in-vitro synergy. Synergistic interaction between antibiotics is usually defined as a 2-log10-lower number of CFU/ml for the combination than for its most active component in time-kill studies. Antagonism is defined as 2-log10 increase in CFU/ml betweenthe combination and the most active single agent and additivity is defined as a 1 to <2-log10-lower number of CFU per milliliter for the combination. Other interactions are considered indifferent. [28-30] In the checkerboard and Etest methods, synergy is defined using the fractionalinhibitory concentrations index (FICI), where FICI is the sum of the FICs of individual antibiotics in a combination and the FIC of an antibiotic is defined as the combination’s MIC divided by the MIC of the antibiotic alone. The common convention is that FICIs of <0.5, >0.5–4, and >4 represent synergy, no interaction and antagonism, respectively, [31-33] although variations exist and older studies considered FICIs>1 as antagonistic. [34]
For in-vitro data, a systematic review and meta-analysis of the literature was performed as part of the background for the clinical trial. The following search string was used to locate all studies published in PubMed:
(colistin OR colisti* OR colistimethate OR polymyxin) AND (imipenem OR meropenem OR doripenem OR ertapenem OR carbapenem) AND (pharmacokinetic OR pharmacodynamic OR synergy OR synerg* OR antagonis* OR additive) AND (in-vitro OR checkerboard OR time-kill OR Etest OR E-test OR microdilution OR agar dilution OR susceptibility). A search was run also in Google scholar and the ICAAC, IDSA and ECCMID conference proceedings for the years 2007-2012. References of all included studies were reviewed for more eligible studies.
(colistin OR polymyxin) AND (imipenem OR meropenem OR doripenem OR carbapenem) AND (combination[ti] OR synergy[ti] OR synerg*[ti] OR combin*[ti]). In addition, the references of all included studies were searched for additional studies.
For each study, we sought to extract the method of in-vitro synergy testing, bacterial species, the type of carbapenem and polymyxin used, and number of isolates tested. Reported MICs of study isolates for the carbapenem and polymyxin tested were also extracted and susceptibility was assessed according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) published breakpoints. [35]We calculated synergy rates, where synergy was counted as an event and the sample size was the number of isolates tested. We used mixed-effects analysis in order to provide a pooled rate. The I2 statistic was used to test heterogeneity. Comprehensive Meta-Analysis V2.2 (Biostat, Englewood NJ, 2005) was used for analysis.
Thirty-eightpublished studies and 15 conference proceeding were included, reporting on 244 different tests on 1050 bacterial isolates. A summary of selected studies are presented in Table 1. In time-kill studies, combination therapy showed synergy rates of 77% (95% CI 64-87) for A.baumannii, 44% (95% CI 23-51%) for Klebsiella pneumoniae and 50% (95% CI 30-69%) for Pseudomonas aeruginosa with low antagonism rates for all. For A. baumannii, meropenem was more synergistic than imipenem, whereas for P. aeruginosa the opposite was true. Checkerboard and Etest studies generally reported lower synergy rates than time-kill. Comparisons of resistance development between monotherapy and combination therapy were found in one study on 3 A. baumannii isolates and four studies on 14 P. aeruginosa isolates, all recent studies. Use of combination therapy led to less resistance development in-vitro.
Thus, in-vitro studies show variable results, but overall synergy is substantial. Carbapenem-polymyxin synergy is probably more likely when isolates are more susceptible to one or both of the drugs in the combination. It was observed more frequently with A. baumannii than with K. pneumonia or P. aeruginosa strains and this could be related to lower MICs for A. baumannii to carbapenems in general. Difference between carbapenems is less clear and depended on bacteria type, with doripenem having some advantage.
Learning from in-vitro studies on clinical effects is difficult because the bacterial inocula differ, drug levels may be affected by practical constraints of antibiotic administration and clinical effects are confounded by underlying conditions and adverse effects. Furthermore, poor correlation has been shown between differentin-vitro methods for synergy testing.[34] Indeed, despite strong in-vitro proof of synergy and prevention of resistance induction for beta-lactam-aminoglycoside combinations for various Gram-negative and Gram-positive bacteria, randomized controlled trials do not show a clinical benefit for the same combinations compared with beta-lactams alone in the treatment of sepsis by the same bacteria[36]. Detriments of combination therapy may comprise of further resistance induction, increased toxicity and antagonistic interactions between antibiotics. Thus, the effects of combination therapy must be tested in clinical studies
Data from in-vivo and human studies on combination therapy is weak. Three in-vivo studies examined the role of carbapenem-polymyxin combination (Table 2), all examining the effect of combining imipenem and colistin. While two studies P. aeruginosa studies found improved outcome with combination, the third tested on A. baumannii showed no benefit with this combination. Three studies were found reporting on the clinical effects of combination therapy (Table 3)[37-39].Two were retrospective comparative studies, comparing carbapenem-colistin combination therapy to colistin monotherapy. One showed worst survival with combination therapy [37], but there was an inherent difference between patient groups in that patients with P. aeruginosa were treated with colistin monotherapy while combination therapy was given mostly to patients infected by A. baumannii. The second very small study showed improved survival in five patients receiving combination therapy compared to seven patients treated with polymyxin monotherapy among patients with K. pneumoniae bacteremia[38]. The last study compared any combination therapy (the most common combination was tigecycline and colistin) to any monotherapy (the most common was tygecycline) and found an overall advantage to combination therapy[39]. Colistin monotherapy was given to 22 patients and colistin-meropenem combination therapy to 6 patients in this study.
The objective of the current trial is to examine the clinical effects of colistin-carbapenem combination therapy in the optimal trial design. Basing on the review of PK studies we will select the currently optimal dosing regimen for colistin, including a loading dose. Given no difference in the expected interactions, we will select meropenem as the carbapenem tested since high doses can be given to critically-ill patients and is the carbapenem of choice in the trial centers. To avoid bias we will conduct a randomized controlled trial, but given the expected difficulties in obtaining informed consent we will prospectively collect data from all eligible patients, documenting their treatment regimen if not recruited into the randomized controlled trial.
1
Version 9
August 6, 2013
Table 1: Studies examining colistin-carbapenem combination therapy
First author / Year published / Polymyxin tested / Carbapenem tested / Bacteria type / no. of isolates / Carbapenem resistance / Polymyxin Resistance / Synergy methods / Outcome reportedChan[34] / 1987 / colistin / imipenem / P. aeruginosa, S. maltophilia / 33 / R,S / R / checkerboard, time-kill / FICI
Rynn[40] / 1999 / colistin / meropenem / P. aeruginosa / 2 / S / S / time-kill / AUKBC
Yoon[41] / 2003 / polymyxin B / imipenem / A.baumannii / 8 / R / R / checkerboard, time-kill / FICI, time-kill synergy, bactericidality
Landman[42] / 2005 / polymyxin B / imipenem / P. aeruginosa / 10 / R / S / time-kill / bactericidality
Bratu[43] / 2005 / polymyxin B / imipenem / K. pneumoniae / 16 / R / R,S / time-kill / bactericidality, time-kill synergy
Timurkaynak[44] / 2006 / colistin / meropenem / A. baumannii, P. aeruginosa / 10 / R,S / S / checkerboard / FICI
Wareham[45] / 2006 / polymyxin B / imipenem / A. baumannii / 5 / R / S / Etest / FICI
Tateda[46] / 2006 / polymyxin B / imipenem / P. aeruginosa / 12 / R / R / checkerboard breakpoint / FICI
Biancofiore[47] / 2007 / colistin / meropenem / A. baumannii / 1 / R / S / checkerboard / FICI
Cirioni[48] / 2007 / colistin / imipenem / P. aeruginosa / 2 / R,S / R / checkerboard, time-kill / FICI, time-kill synergy
Tripodi[49] / 2007 / colistin / imipenem / A. baumannii / 9 / R / S / time-kill / bactericidality, time-kill synergy
Pankuch[50] / 2008 / colistin / meropenem / P. aeruginosa, A. baumannii / 102 / R,S / R,S / time-kill / time-kill synergy
Tascini[51] / 2008 / colistin / imipenem / E. cloaca / 1 / S / S / checkerboard / FICI
Guzel[52] / 2008 / colistin / meropenem / P. aeruginosa / 50 / S / S / checkerboard / FICI
Guelfi[53] / 2008 / polymyxin B / meropenem / P. aeruginosa, A. baumannii / 20 / R,S / S / checkerboard / FICI
Burgess[54] / 2008 ICAAC / colistin / meropenem / A. baumannii / 5 / R / S / time-kill / bactericidality, time-kill synergy
Ullman[55] / 2008 ICAAC / colistin / meropenem / A. baumannii / 3 / R,S / S / PK/PD time-kill / bactericidality
Pankey[56] / 2009 / polymyxin B / meropenem / A. baumannii / 8 / R / S / Etest, time-kill / FICI, bactericidality, time-kill synergy
Souli[57] / 2009 / colistin / imipenem / K. pneumonia / 42 / R,S / R,S / time-kill / time-kill synergy
Burgess[58] / 2009 ICAAC / colistin / imipenem / A. baumannii / 5 / R / S / time-kill / bactericidality, time-kill synergy
Hilliard[59] / 2009 ICAAC / colistin / doripenem / P. aeruginosa / 2 / S / S / checkerboard / FICI
Milne[60] / 2010 / colistin / meropenem, imipenem / P. aeruginosa / 144 / R,S / R,S / Etest, SBPI / FICI, SBPI
Pongpech[61] / 2010 / colistin / meropenem, imipenem / A. baumannii / 30 / R / S / checkerboard, time-kill / FICI
Rodriguez[62] / 2010 / colistin / imipenem / A. baumannii / 14 / R,S / R,S / time-kill / bactericidality, synergy
Elemam[63] / 2010 / polymyxin B / imipenem / K. pneumoniae / 12 / R / R / checkerboard / FICI
Lin [64] / 2010 / colistin / imipenem / E. cloaca / 1 / S / S / time-kill / bactericidality, synergy
Shields[65] / 2010 / colistin / imipenem, doripenem / A. baumannii / 17 / R / S / Etest, time-kill / FICI, bactericidality, synergy
Sopirala[66] / 2010 / colistin / imipenem / A. baumannii / 8 / R / S / checkerboard, Etest, time-kill / FICI, time-kill synergy
Urban[67] / 2010 / polymyxin B / doripenem / K. pneumoniae, A. baumannii, P. aeruginosa, E. coli / 20 / R,S / R,S / time-kill / bactericidality
Pankuch[68] / 2010 / colistin / doripenem / A. baumannii, P. aeruginosa / 50 / R,S / R,S / time-kill / time-kill synergy
Steed[69] / 2010 ECCMID / colistin / imipenem / A. baumannii / 8 / R / S / time-kill / bactericidality, time-kill synergy
Souli[70] / 2010 ECCMID / colistin / meropenem, ertapenem / K. pneumoniae / 55 / R,S / R,S / time-kill / time-kill synergy
Khuntayaporn[71] / 2010 ICAAC / colistin / imipenem, meropenem, doripenem / P. aeruginosa / 57 / R / - / checkerboard / FICI
Dorobisz[72] / 2010 ICAAC / colistin / doripenem / Acinetobacter / 6 / R / R / checkerboard, time-kill / FICI, bactericidality
Srispha-Olarn[73] / 2010 ICAAC / colistin / meropenem / A. baumannii / 3 / R / S / PK/PD time-kill / bactericidality, time-kill synergy
Ly[74] / 2011 ICAAC / colistin / doripenem / P. aeruginosa / 3 / S / R,S / PK/PD time-kill / bactericidality
Liang[75] / 2011 / colistin / meropenem / A. baumannii / 4 / R / S / time-kill / bactericidality, synergy
Pankey[76] / 2011 / polymyxin B / meropenem / K. pneumoniae / 14 / R,S / R,S / Etest, time-kill / FICI, bactericidality, time-kill synergy
Sheng[77] / 2011 / colistin / imipenem / A. baumannii / 18 / R / S / checkerboard, time-kill / FICI, bactericidality, time-kill synergy
Bergen[78] / 2011 / colistin / imipenem / P. aeruginosa / 6 / R,S / R,S / time-kill / bactericidality, time-kill synergy
Bergen[79] / 2011 / colistin / doripenem / P. aeruginosa / 2 / R,S / R,S / PK/PD time-kill / bactericidality, time-kill synergy
Santimaleeworagun[80] / 2011 / colistin / imipenem / A. baumannii / 8 / R / S / checkerboard / FICI
Lim[81] / 2011 / polymyxin B / meropenem / P. aeruginosa / 22 / R / S,R / time-kill / bactericidality
Morosini[82] / 2011 ECCMID / colistin / meropenem / K. pneumoniae / 1 / S / S / time-kill / bactericidality, FICI
Poudyal[83] / 2011 ECCMID / colistin / doripenem / A. baumannii / 3 / R,S / S / PK/PD time-kill / bactericidality, time-kill synergy
Teo[84] / 2011 ICAAC / polymyxin B / doripenem / P. aeruginosa / 16 / R / - / time-kill / bactericidality, time-kill synergy
Principe[85] / 2011 ICAAC / colistin / doripenem / A. baumannii / 24 / R,S / - / checkerboard / synergy
Mohamed[86] / 2011 ICAAC / colistin / meropenem / P. aeruginosa / 2 / R,S / S / PK/PD time-kill / bactericidality, time-kill synergy
Peck[87] / 2012 / colistin / imipenem / A. baumannii / 6 / R / R,S / time-kill / bactericidality, synergy
Jernigan[88] / 2012 / colistin / doripenem / K. pneumoniae / 12 / R / S,R / time-kill / bactericidality, time-kill synergy, AUBKC
Deris[89] / 2012 / colistin / doripenem / K. pneumoniae / 4 / R,S / R,S / PK/PD time-kill / bactericidality, time-kill synergy
Ozseven[90] / 2012 / polymyxin B / imipenem, meropenem / A. baumannii / 34 / R / S / checkerboard / FICI
He[91] / 2012 / colistin / doripenem / P. aeruginosa / 100 / R / S / Etest, time-kill / FICI
R - resistant, S - sensitive, MDR – multidrug resistant, XDR – extremely drug resistant, AUBKC – area under the bacterial killing curve
Table 2 – in-vivo studies
Study / Methods / Type of carbapenem1 / Bacteria (MIC in mg/L) / Outcome / Effect on defined outcome / Effect summaryCirioni 2007 [31] / In-vivo randomized (BALB/c male mice with bacteremia following IV injection of P. aeruginosa) / Imipenem / P. aeruginosa, 1 quality control strain: imipenem MIC 0.5, colistin MIC 4)
One CR MDR clinical isolate: imipenem MIC 32 colistin MIC 8 / Deaths (colistin vs. combi)
Positive blood culture at 24h / Control strain: 8/20 vs. 2/20
Clinical strain: 6/20 vs. 3/20
Control strain: 8/20 vs. 2/20
Clinical strain: 13/20 vs. 3/20 (p<0.05 for all) / significant effects in-vivo on survival and bacteremia clearance
Aoki 2008[92] / In-vivo - BALB/c female mice pneumonia model (intranasal and subcutaneous) / imipenem / P. aeruginosa, 1 PAO1 strain and 6 clinical strains / Survival, lung bacterial burden / 90% survival in combination vs 10% in monotherapy, reduced bacterial burden / Significant effects on survival and bacterial burden with colistin
Song 2009[93] / In-vivo randomized (neutropenic mice with pneumonia following tracheal A. baumannii inoculation) / Imipenem / A. baumannii, 1 clinical CR isolate OXA-51 positive. Imipenem MIC 64, colistin ≤0.5 / Lung bacterial loads at 48h
Bacteremia eradication at 48h
Mortality at 48h / Combi 7.15 ± 3.56 vs.
Colistin alone 6.35 ± 0.98
Combi 0/3 vs.
Colistin alone 0/3
Combi 1/3 vs.
Colistin alone 1/3 / In-vivo bacterial load, bacteremia and mortality
Table 3- observational studies