Polymyxins

Russell E. Lewis, Pharm.D
Associate Professor of Infectious Diseases (MEDS-10/B)

russelledward.lewis@unipd.it
https://github.com/Russlewisbo
Slides and course materials: www.idpadova.com

Outline

• History — discovery, retirement, and the carbapenem-resistance renaissance
• Chemistry and mechanism — why charge matters
• Resistance — intrinsic, chromosomal, and the mcr revolution
• The CMS prodrug problem and modern PK/PD
• Dosing — loading doses, weight-based maintenance, the “polymyxin B over CMS” pivot
• Nephrotoxicity, neurotoxicity, and the pigmentation we don’t talk about
• Clinical use — CRE, CRAB, CRPA — and when not to use polymyxins
• Inhaled and intrathecal routes
• The AIDA trial, OVERCOME, and what monotherapy means in 2026
• Where polymyxins sit now that ceftazidime-avibactam, meropenem-vaborbactam, imipenem-relebactam, and cefiderocol are on the shelf

History

A case….

64-year-old liver transplant recipient, POD 14, develops shock and bacteremia. Cultures grow K. pneumoniae — KPC-3, ceftazidime-avibactam MIC 16 (resistant after 10 days of therapy), colistin MIC 1, meropenem MIC > 16. Creatinine has doubled from baseline

• What do you recommend?
• What dose?
• Mono or combination?
• Inhaled component?

This case has all the features that make polymyxins clinically interesting: salvage indication after newer agent failure, renal dysfunction limiting tolerance, the dosing-unit question, the monotherapy-vs-combination debate, and the inhaled-route question if it later progresses to pneumonia.

Discovery and first retirement

• Isolated in 1947 from Paenibacillus polymyxa (originally Bacillus polymyxa) — concurrent reports by Stansly, Benedict, Brownlee, and Bushby
• Five polymyxin variants (A–E) initially described; only polymyxin B and polymyxin E (colistin) advanced to clinical use
• Widely used through the 1960s for Gram-negative infection
• Largely abandoned in the 1970s as aminoglycosides and later β-lactams proved more effective and less nephro/neurotoxic
• Survived in topical preparations (Polysporin, ear/eye drops) and SDD regimens

Two of the five polymyxins survived; three never made it. The retirement was driven by the toxicity profile that became unacceptable once aminoglycosides and broad-spectrum β-lactams arrived. It took the resurgence of MDR Gram-negatives in the 2000s to drag them back.

The carbapenem-resistance renaissance

• Emergence of carbapenem-resistant Pseudomonas, Acinetobacter, and Enterobacterales (KPC, NDM, OXA-48) starting in the 2000s reopened the door
• Polymyxins became drugs of last resort — often the only remaining IV option for KPC-producing K. pneumoniae and CR-A. baumannii
• Modern era: PK/PD reinvestigation revealed the field had been dosing colistin incorrectly for decades
• Now partially displaced by ceftazidime-avibactam, meropenem-vaborbactam, imipenem-relebactam, and cefiderocol — but still in active use where these are unavailable or not indicated

“Agents of last resort” — The renaissance was forced; nobody wanted polymyxins back. What’s now clear: we were using subtherapeutic doses for years before the modern PK/PD work clarified targets (Kaye et al., 2016).

Chemistry and mechanism

Chemistry — cyclic cationic cipopeptides

• Cyclic decapeptide with a fatty acid tail (6-methyloctanoic or 6-methylheptanoic acid)
• Five positively charged α,γ-diaminobutyric acid (Dab) residues at physiologic pH
• Polymyxin B vs. colistin (polymyxin E) differ by a single amino acid (D-phenylalanine vs. D-leucine at position 6) — minor structural difference, major clinical difference
• Amphipathic structure → interacts with negatively charged Gram-negative outer membrane
• Colistimethate sodium (CMS) is the inactive methanesulfonate prodrug of colistin — formed by reacting colistin with formaldehyde and sodium bisulfite to mask the cationic charge and reduce toxicity

Upper panel- polymyxin B, lower panel- colistin (polymyxin E). The cyclic peptide with multiple cationic residues is the structural fingerprint. The single–amino-acid difference between polymyxin B and colistin is biochemically trivial but pharmaceutically large — colistin is given as the prodrug CMS, polymyxin B is given as the active drug. That’s the entire story of why we have two drugs that look the same on the spectrum chart but behave differently in patients.

Mechanism of action — Electrostatic disruption

The “self-promoted uptake” model: the cationic drug breaks down the outer membrane it’s trying to get through. The endotoxin neutralization effect is why polymyxin B hemoperfusion columns exist — though the EUPHRATES trial showed no mortality benefit in septic shock with endotoxin activity (From et al., 1979).

Spectrum at a glance

• Active: most Gram-negative bacilli — E. coli, Klebsiella, Enterobacter, Salmonella, Shigella, Pseudomonas aeruginosa, Acinetobacter baumannii, Stenotrophomonas (limited)
• Intrinsically resistant: Proteus, Providencia, Morganella, Serratia (modified lipid A), Burkholderia cepacia, all Gram-positive bacteria, anaerobes, Neisseria, Brucella, Helicobacter
• Not used against: any organism with a β-lactam alternative — toxicity is rarely worth it

The intrinsic resistance list is short but unforgiving. Proteus, Providencia, Morganella, and Serratia modify their lipid A so polymyxins don’t bind — this is intrinsic, not acquired. Burkholderia is also intrinsically resistant. Worth a mental rule: don’t use polymyxin for any organism on that list.

Resistance

Resistance overview

• Three mechanisms of acquired polymyxin resistance:
  1. Modification of lipid A to reduce its net negative charge — the dominant mechanism
  2. Capsule and outer membrane protein changes
  3. Efflux
• Plasmid-mediated mcr genes added a fourth mechanism in 2015 — and made colistin resistance horizontally transferable

(Cai et al., 2012)

Lipid A modification — The charge story

• Polymyxins bind lipid A through electrostatic attraction to negatively charged phosphate groups
• Bacteria modify lipid A by adding:
  • 4-amino-4-deoxy-L-arabinose (L-Ara4N) — neutralizes one phosphate
  • Phosphoethanolamine (pEtN) — adds a positive charge
• Net effect: outer membrane becomes less negatively charged → drug binds less avidly
• Regulated chromosomally via PmrAB and PhoPQ two-component systems
• Constitutive activation of these regulons → high-level polymyxin resistance

The PmrAB/PhoPQ regulons are the hub of polymyxin resistance. When stressed (low Mg²⁺, low pH, sub-MIC polymyxin exposure), bacteria turn on these regulons and start dressing their lipid A in chemical armor.

PmrAB and PhoPQ — The regulatory nub

• Two-component systems that sense environmental stress (low Mg²⁺, low pH, sub-MIC polymyxin)
• Activated regulons upregulate the arn operon (L-Ara4N addition) and eptA (pEtN addition)
• Constitutive activation by mutation in pmrB or phoQ → high-level polymyxin resistance
• Same regulators control virulence genes in Salmonella and K. pneumoniae — there are fitness costs to resistance
• Single AA substitutions in PmrB or PhoQ are sufficient for clinical resistance

The regulatory architecture is conserved across many Gram-negatives. The fitness cost of constitutive activation is one of the reasons mcr-mediated resistance — which doesn’t require regulatory upheaval — was such a clinically dangerous development.

mcr — Plasmid-mediated polymyxin resistance

• 2015 (Liu, Lancet Infect Dis): mcr-1 discovered in E. coli from Chinese livestock and humans — first plasmid-mediated colistin resistance gene
• Encodes a phosphoethanolamine transferase (homologue of chromosomal eptA) that modifies lipid A
• Horizontally transferable on IncX4, IncI2, and IncHI2 plasmids
• Now identified globally; mcr-1 through mcr-10 variants described
• Frequently co-located with carbapenemase genes (NDM, OXA-48, KPC) — pan-resistant organisms now reported

Liu 2016 was the headline paper of the decade for Gram-negative resistance. The mcr family broke a long-standing assumption that polymyxin resistance was chromosomally encoded and slow to spread. The co-occurrence with carbapenemases is the nightmare scenario — pan-resistant Enterobacterales for which no oral or IV agent reliably works (Beyrouthy et al., 2017; Nang et al., 2019; Y-Y et al., 2016; Zhong et al., n.d.).

mcr — The timeline

mcr variant timeline

Year	Variant	First report	Notable feature
2015	mcr-1	China, E. coli (livestock + human)	First plasmid-mediated colistin resistance
2016	mcr-2	Belgium, E. coli	Variant on IncX4 plasmid
2017	mcr-3	China, E. coli	High-prevalence emergence
2017	mcr-4	Italy, Salmonella	First non-E. coli host
2017	mcr-5	Germany, Salmonella
2018–24	mcr-6 to mcr-10	Global	Multiple host species

The pace of mcr variant discovery — 10 distinct variants in roughly 7 years after the first description — reflects how quickly the field looked once they knew what to look for. The genes were almost certainly circulating well before 2015.

Heteroresistance

• Heteroresistance: an isolate appears susceptible by routine AST but contains a small subpopulation (often <1%) of resistant cells
• Under polymyxin selection pressure, the resistant subpopulation expands → rapid emergence of resistance during therapy
• Particularly well-documented in A. baumannii and K. pneumoniae
• Implication: polymyxin monotherapy carries a real risk of within-patient resistance emergence — one of the arguments for combination therapy
• Difficult to detect with routine methods — population analysis profile (PAP) testing is the gold standard

Heteroresistance is the part of the polymyxin resistance story you don’t see on a regular susceptibility report. Routine AST misses it; the patient’s culture comes back susceptible, the patient gets monotherapy, and ten days later they relapse with a fully resistant isolate. This argued for combination therapy clinically until the AIDA and Kaye trials suggested combination didn’t actually help outcomes (LiJ et al., 2006; Meletis et al., 2011).

Susceptibility testing — A mess

• Broth microdilution is the only validated method — disk diffusion and gradient strips perform poorly for polymyxins
• Polymyxins adsorb to plastic and polystyrene in standard MIC plates → falsely elevated MICs
• Solution: polysorbate-80 added to broth, or use of specific polymyxin reference panels
• Breakpoint debate: EUCAST and CLSI use susceptible ≤2 mg/L; recent USCAST recommendation abandons “susceptible” entirely — only “intermediate” (≤2) and “resistant” (≥4)
• Practical: an “S” colistin result is not a guarantee — combine with PK considerations

The susceptibility testing situation is a mess. The drug sticks to plastic, gradient strips don’t work, and the dose required to achieve the MIC overlaps with the dose that causes nephrotoxicity. USCAST took the step of removing “susceptible” as a possible result — meaning even at the lowest MICs we cannot reliably promise efficacy (Pogue et al., 2020).

Pharmacology —
The CMS Prodrug Problem

Two drugs, Two pharmacokinetic stories

PK comparison — polymyxin B vs. colistin

Property	Polymyxin B	Colistin (as CMS)
Administered as	Active drug	Inactive prodrug (CMS)
Active species formation	Immediate	Slow hydrolysis to colistin in blood
Renal elimination	Mostly non-renal	Active drug ~70% non-renal; CMS ~70% renal
Renal dose adjustment	No (or minimal)	Yes (CMS, not colistin)
Urinary concentration	Low (not good for UTI)	High via CMS hydrolysis in urinary tract
PK predictability	Better	Worse (variable CMS conversion)

The dosing logic for polymyxin B and colistin is not the same — the renal handling is opposite. Colistin needs renal adjustment because most of the CMS prodrug is cleared by the kidney before it converts. Polymyxin B doesn’t, because it’s given as the active drug and cleared non-renally. The UTI exception runs the other way — colistin is preferred for UTI because the CMS hydrolyzed in the urinary tract delivers active drug locally (Garonzik et al., 2011; Plachouras et al., 2009; Sandri et al., 2013).

CMS — The prodrug trap

• Colistimethate sodium (CMS) is colistin with each of its 5 primary amines methanesulfonylated → inactive
• In aqueous solution, CMS spontaneously hydrolyzes to a mixture of partially-methanesulfonylated species and free colistin
• Hydrolysis is slow and incomplete in vivo — only ~30% of an administered CMS dose appears as colistin in blood
• At steady state, plasma colistin Cmax ≈ 2–3 mg/L on standard doses — barely at the MIC of many MDR organisms
• A second dose unit problem: CMS is dosed as IU or “mg colistin base activity” (CBA) — frequent unit-conversion errors

The CMS prodrug story is the historical disaster of colistin dosing. The drug was approved in the 1950s based on assumptions about conversion that turned out to be wrong — actual conversion is much lower than assumed, which meant decades of patients were under-dosed. Modern population PK studies (Garonzik, Plachouras, Sandri) recalibrated everything (Bergen et al., n.d.; Garonzik et al., 2011).

CMS Dosing Units — The unit confusion

• CMS is sold in multiple unit conventions:
  • International Units (IU)
  • Milligrams of colistimethate sodium (mg CMS)
  • Milligrams of colistin base activity (mg CBA)
• 1 million IU CMS ≈ 80 mg CMS ≈ 30 mg CBA
• US labels typically use CBA; European labels use IU
• Medication errors — including fatal under- and over-dosing — have been attributed to unit conversion mistakes
• Always verify which unit your hospital’s label uses before prescribing or verifying a colistin order

Unit confusion has killed patients. A patient transferred between hospitals or countries can easily get a 5-fold dose error if the receiving team assumes the unit convention of the prior order. Build a habit of writing the unit explicitly on every colistin order (LiJ and Nation, 2006).

Loading dose — The Garonzik/Nation insight

• Population PK work by Garonzik 2011 and Nation 2017 established that without a loading dose, colistin steady-state plasma concentrations are not reached for 2–3 days — too slow for serious infection
• Loading dose recommendation:
  • 300 mg CBA (9 million IU CMS) IV as a single dose
  • or weighted by total body weight: 5 mg CBA/kg up to maximum 300 mg
• Maintenance: 300 mg CBA/day in divided doses, adjusted for renal function
• Critically ill patients require the loading dose — failing to give it virtually guarantees subtherapeutic concentrations during the first 48 hours

The loading-dose recommendation is the single biggest change in colistin practice from the resurgence era. If you take only one prescribing pearl from this lecture: in a critically ill patient with a serious CR-Gram-negative infection, the loading dose is mandatory. Otherwise you’re effectively treating without antibiotics for the first 48 hours (Garonzik et al., 2011; Nation et al., 2017, 2016; Tsuji et al., 2019).

Colistin dosing — A practical walkthrough

For a critically ill 70-kg patient with normal renal function:

• Loading dose: 300 mg CBA IV over 1 hour (or 9 MIU CMS)
• First maintenance dose: start 12 hours after loading dose
• Maintenance: 150 mg CBA q12h (= 300 mg CBA/day, divided)
• Adjust for renal function based on CrCl — Nation 2017 table
• Re-dose for CRRT: 440 mg CBA/day in divided doses (high clearance)
• Target plasma steady-state concentration: ≥ 2 mg/L for MIC 2, ≥ 4 mg/L for MIC 4 — often unattainable safely

The Nation 2017 dosing table is the most widely cited modern reference. CRRT actually increases colistin clearance via convective transport of the formed colistin, so the dose for a patient on CRRT is higher than for a stable dialysis patient. Easy to under-dose if you don’t think about this (Nation et al., 2017; Tsuji et al., 2019).

Polymyxin B dosing

• Polymyxin B is given as the active drug — no prodrug delay
• Loading dose: 2.0–2.5 mg/kg IV over 1 hour
• Maintenance: 1.25–1.5 mg/kg q12h
• Generally no renal dose adjustment — non-renal clearance dominates
• Preferred over CMS for most systemic indications in the 2019 International Consensus Guidelines — more predictable PK, possibly less nephrotoxicity
• Exception: UTI — polymyxin B does not achieve adequate urinary concentrations; CMS is preferred for upper UTI

This is the modern pivot — for bacteremia, pneumonia, or other non-urinary infection, polymyxin B is now favored. For UTI, colistin (CMS) wins because the urinary route hydrolyzes CMS to colistin locally, delivering active drug to the site of infection (Sandri et al., 2013; Tsuji et al., 2019).

Renal dose adjustment — Colistin only

CMS renal adjustment (after loading dose) — adapted from Nation 2017

CrCl (mL/min)	CMS daily dose (mg CBA)	Frequency
> 90	300	divided q12h
50–89	270	divided q12h
30–49	240	divided q12h
10–29	180	divided q12h
< 10 (no HD)	100	once daily
HD	60–80 + supplement post-HD	once daily + HD-day boost
CRRT	440	divided q8–12h

The dialysis-day supplemental dose is critical — CMS is dialyzable, so without a post-HD top-up, exposure crashes after each session. CRRT, paradoxically, requires more drug than HD because of continuous clearance (Nation et al., 2017).

PK/PD — AUC/MIC is the target

• Concentration-dependent killing with significant post-antibiotic effect
• Free drug AUC/MIC is the primary PK/PD index
• Target: fAUC/MIC ≈ 12–20 for stasis or 1-log kill against most isolates
• At standard doses against an organism with MIC 2 mg/L, target may not be achievable
• Pulmonary epithelial lining fluid concentrations of colistin are poor with IV CMS — argues for inhaled colistin in pneumonia
• Bone, CSF, prostate, eye: poor penetration

The PK/PD target is achievable for the lowest MICs but not the highest within the susceptible range. This is the core reason monotherapy fails so often — we’re at the edge of efficacy at the doses we can tolerate (Bergen et al., 2011; Cheah et al., 2015; Landersdorfer et al., 2018).

Adverse Effects

Nephrotoxicity — The big one

• Incidence: 30–60% across modern cohort studies
• Mechanism: polymyxin uptake by megalin-mediated proximal tubular reabsorption → tubular cell damage
• Risk factors:
  • Higher cumulative dose / higher plasma levels (trough > 2.4 mg/L predicts AKI)
  • Baseline renal dysfunction
  • Concomitant nephrotoxins (vancomycin, aminoglycosides, IV contrast, NSAIDs)
  • Higher daily dose × longer duration
• Usually reversible with discontinuation
• Frequently dose-limiting — the reason “use just a little longer” rarely helps in non-responders

30–60% is not a typo. Polymyxin nephrotoxicity is the rule, not the exception, in modern cohorts. The Sorli/Forrest work established trough-level thresholds — and the dose required to achieve the PK/PD target overlaps with the dose that causes AKI. This is the central pharmacologic problem with polymyxins (Forrest et al., 2017; Hartzell et al., 2009; Kubin F Elimen et al., 2012; Pogue et al., 2011; Sorli et al., 2013).

Trough-level monitoring — Why it’s hard

• Therapeutic drug monitoring is not routinely available in most clinical labs
• LC-MS/MS required; turnaround days not hours
• Target trough ≤ 2.4 mg/L to minimize AKI risk (Sorli 2013)
• Target plasma steady-state concentration ≥ 2 mg/L for efficacy against MIC 2 mg/L (Garonzik)
• These targets overlap — the therapeutic window is essentially zero
• A few academic centers in Europe and US offer TDM; most clinicians dose by weight and renal function alone

The therapeutic window is the central frustration. TDM would help, but it isn’t available. We dose by formula and hope for the best.

Neurotoxicity

• Paresthesias (perioral, peripheral) — most common, dose-related, usually mild
• Ataxia, dizziness, vertigo
• Neuromuscular blockade — potentiates non-depolarizing blockers; rare respiratory paralysis reported with high-dose IV
• Generally reversible
• Inhaled polymyxins: bronchospasm rather than systemic neurotoxicity

Neurotoxicity is the textbook adverse effect, but in modern dosing it’s rarely dose-limiting. Paresthesias are common but tolerable. The neuromuscular blockade concern is real for patients receiving anesthesia or already on non-depolarizing agents.

CMS storage and stability

• Powder vials: stable at room temperature
• Reconstituted CMS in saline: stable refrigerated for limited periods (manufacturer typically states 24 hours)
• Spontaneous hydrolysis to colistin continues in solution — bag age = more active drug = more local toxicity on infusion
• Premix policy: mix immediately before each dose where feasible
• IV bags hung > 24 hours should be discarded — both efficacy (oxidative degradation) and toxicity (formed colistin) concerns

The stability quirks are pharmacy operational details that bedside clinicians often don’t see. A bag of CMS that sat in the satellite pharmacy fridge for 36 hours is not the same as a freshly mixed bag — and the formed colistin in it can hurt the patient.

Drug interactions

• Other nephrotoxins — vancomycin, aminoglycosides, IV contrast, NSAIDs, calcineurin inhibitors — additive AKI risk
• Non-depolarizing neuromuscular blockers — polymyxins potentiate; prolonged paralysis reported
• No CYP-mediated interactions of clinical significance
• No QT prolongation — a rare oasis among antibiotics
• Premix instability — do not premix CMS in IV bags more than 24 hours before use; hydrolysis to active colistin in the bag increases toxicity

The lack of CYP and QT effects makes polymyxins relatively easy to combine with most other drugs except nephrotoxins. The premix instability is a pharmacy-level concern worth flagging — formed colistin in a sitting IV bag is more toxic than freshly mixed CMS.

Skin hyperpigmentation — The polymyxin B quirk

• Diffuse skin and head/neck darkening reported specifically with polymyxin B
• Onset typically within first weeks of therapy- higher incidence in transplant patients
• Histology: dermal melanophages with melanin deposition
• Reversible over months after discontinuation
• Mechanism poorly understood; may involve histamine release and inflammatory cytokine-driven melanocyte activation
• Not seen with colistin to the same degree
• Important to counsel patients before initiating polymyxin B

The hyperpigmentation is the under-counseled adverse effect. Patients on prolonged polymyxin B can develop striking darkening of the head, neck, and upper torso — distressing if unexpected. Reassure that it reverses over months (Mattos et al., 2016; Zavascki et al., 2015; Zhou et al., 2022).

Clinical Use

Time-to-effective therapy

• For MDR Gram-negative bacteremia, each hour of delay in effective antibiotic therapy is associated with measurable mortality increase
• Empirical polymyxin coverage may be warranted when:
  • Prior MDR colonization or recent CR-Gram-negative infection
  • Local CR rates > 20% among likely pathogens
  • Sepsis or septic shock with high pretest probability of MDR organism
• De-escalate within 48–72 h once susceptibilities allow — empirical use must be time-limited
• The Tsuji 2019 guidelines emphasize early adequate dosing (loading dose) over delayed perfect dosing

(Tsuji et al., 2019)

When polymyxins are considered

• Carbapenem-resistant Enterobacterales (CRE) — KPC, NDM, OXA-48 producers without availability of newer β-lactam/inhibitors
• Carbapenem-resistant A. baumannii (CRAB) — limited alternatives; cefiderocol, sulbactam-durlobactam now competing
• Carbapenem-resistant P. aeruginosa (CRPA) — competing with ceftolozane-tazobactam, ceftazidime-avibactam, imipenem-relebactam, cefiderocol
• MDR Stenotrophomonas — limited supportive data; TMP-SMX remains first-line
• Empiric coverage for septic patient with prior MDR Gram-negative colonization

The list is shrinking as newer agents arrive. In 2026, polymyxin use is increasingly concentrated in settings where ceftazidime-avibactam, meropenem-vaborbactam, imipenem-relebactam, and cefiderocol are unavailable, contraindicated, or inactive against the specific isolate(Kaye et al., 2016).

Pediatric considerations

• Pediatric data are sparse — most dosing extrapolated from adult studies
• Loading dose: 4 mg CBA/kg colistin (max 300 mg) suggested for children with serious infection
• Maintenance: 5–7.5 mg CBA/kg/day in divided doses, renal-adjusted
• Polymyxin B: 1.5–2.5 mg/kg/day divided q12h
• Inhaled colistin widely used in pediatric CF (Ratjen et al.)
• Nephrotoxicity rates may be lower in children — but data sparse

(Ratjen et al., 2006)

Pregnancy and breastfeeding

• Pregnancy category C for both polymyxin B and colistin
• Animal data inconclusive; human data limited to case reports
• Use only when no safer alternative exists for life-threatening MDR Gram-negative infection
• Cross placenta; potential for fetal renal exposure
• Breast milk transfer minimal but oral bioavailability of polymyxins is negligible — relevant exposure unlikely
• Document risk-benefit discussion explicitly

Polymyxins in pregnancy are an “absolute last resort” decision. The toxicity profile is concerning enough in adults that fetal exposure considerations are real, but life-threatening MDR infection in pregnancy can force the issue.

Carbapenem-resistant enterobacterales

• First-line in 2026 (where available): β-lactam/inhibitor combinations active against the carbapenemase
  • KPC: ceftazidime-avibactam, meropenem-vaborbactam, imipenem-relebactam
  • OXA-48: ceftazidime-avibactam
  • NDM: ceftazidime-avibactam + aztreonam, or cefiderocol
• Polymyxin role: salvage when newer agents unavailable, contraindicated, or when isolate also resistant to them (e.g., emergent ceftazidime-avibactam resistance in KPC-3 variants)
• Increasingly the second drug of last resort, not the first

For KPC, the modern β-lactam/inhibitors have largely replaced polymyxins as first-line. The polymyxin is now a salvage choice or for resistance to newer agents — emergent ceftazidime-avibactam-R KPC variants are a clinical reality.

Carbapenem-Resistant Acinetobacter baumannii (CRAB)

• Historically: colistin ± meropenem ± tigecycline ± ampicillin-sulbactam
• Sulbactam-durlobactam (2023, FDA-approved): first agent designed specifically for CRAB; superior to colistin in the ATTACK trial
• Cefiderocol active in vitro against most CRAB but inconsistent clinical performance
• High-dose ampicillin-sulbactam still has a role
• Combination therapy historically standard; recent trials (AIDA, OVERCOME, Kaye 2023) challenge the necessity

CRAB is the hardest pathogen on the list. Sulbactam-durlobactam is the new agent of choice where available — superior to colistin in the ATTACK trial, with much better tolerability. In settings where it’s not available (most of the world), polymyxin-based combinations remain a default (Betrosian A. P Frantzeskaki et al., 2008; Hartzell et al., 2007; Maragakis and Perl, 2008; Oliveira et al., 2008).

Synergy testing — Mostly unhelpful clinically

• In vitro synergy between colistin and many partners (carbapenems, rifampin, fosfomycin, tigecycline) demonstrated repeatedly
• Methods: checkerboard, time-kill, E-test crossing
• Problem: in vitro synergy does not reliably predict in vivo benefit (AIDA, Kaye 2023 are the clearest disproof)
• Synergy testing is not standardized, not endorsed by CLSI/EUCAST/USCAST for clinical decision-making
• Should not delay therapy or anchor treatment selection
• Reserve for truly pan-resistant isolates where every option is empirical

Synergy testing was popular in the early 2010s. The negative outcome trials have largely closed that book — in vitro synergy is interesting biochemistry but unreliable as a clinical guide (Bergen et al., 2011; Owen et al., 2007; Poudyal et al., 2008).

Sulbactam-durlobactam — The new standard for CRAB

• β-lactam + diazabicyclooctane β-lactamase inhibitor combination, FDA-approved 2023
• Sulbactam has intrinsic activity against PBP3 of A. baumannii; durlobactam protects sulbactam from class A, C, and D β-lactamases
• ATTACK trial (Kaye 2023, Lancet ID): sulbactam-durlobactam non-inferior to colistin for CRAB pneumonia/bacteremia, with substantially less nephrotoxicity (12% vs. 38%)
• 28-day mortality similar; secondary outcomes favored sulbactam-durlobactam
• Where available, preferred over polymyxin for CRAB
• Cost and availability limit global access — polymyxin remains the default in much of the world

Sulbactam-durlobactam is the most important recent addition to the CRAB armamentarium. The ATTACK trial design used colistin as comparator and showed clear tolerability advantage — exactly the kind of comparison polymyxins were waiting to lose.

Carbapenem-resistant Pseudomonas aeruginosa

• First-line in 2026: newer agents
  • Ceftolozane-tazobactam (most CRPA isolates not carbapenemase-mediated)
  • Ceftazidime-avibactam
  • Imipenem-relebactam
  • Cefiderocol
• Polymyxin role: XDR strains resistant to newer agents; consider in combination
• Inhaled colistin: adjunct for ventilator-associated pneumonia
• The “use the new β-lactams first” pivot is even stronger for CRPA than CRE

(Linden et al., 2009)

Combination Therapy

The old Argument for combinations

• Polymyxin monotherapy:
  • Sub-MIC at the achievable AUC for many “susceptible” isolates
  • High risk of within-patient emergence of resistance (heteroresistance → constitutive resistance)
  • Therapy failure rates concerning in observational series
• Combinations proposed:
  • Colistin + meropenem (synergy in vitro despite carbapenem resistance — Bergen)
  • Colistin + rifampin
  • Colistin + tigecycline
  • Colistin + fosfomycin
• In vitro synergy and clinical benefit appeared to track in observational data

(Bergen et al., 2011; Owen et al., 2007; Poudyal et al., 2008)

AIDA — Paul 2018

• Open-label RCT, 406 patients, severe CR-Gram-negative infections (mostly Acinetobacter, some Enterobacterales and Pseudomonas)

• Colistin alone vs. colistin + meropenem

• Primary endpoint: clinical failure at 14 days

• Result: no difference in clinical failure (79% colistin alone vs. 73% combination, p = 0.17)

• No difference in 28-day mortality

• Combination arm: more nephrotoxicity

• Conclusion: routine colistin + meropenem combination does not improve outcomes and may increase harm

AIDA was the trial that started the unwinding of the combination-therapy paradigm. Open-label, but the result was clean: no benefit, more nephrotoxicity. The Acinetobacter dominance limits generalizability to Enterobacterales, but it’s still the best RCT we have (Paul et al., 2018).

Kaye 2023 — OVERCOME-like NEJM Evidence Trial

• Multicenter RCT, 464 patients with CR-Gram-negative infections (mostly CRAB)

• Colistin alone vs. colistin + meropenem

• Mortality at 28 days: no significant difference

• Clinical cure: no significant difference

• Combination arm: more nephrotoxicity and more antibiotic exposure

• Confirmed AIDA’s findings in a separate cohort

Kaye 2023 was the second nail in the routine combination coffin. Same conclusion as AIDA — no benefit, more harm. The modern teaching is: don’t routinely combine colistin + carbapenem for CRAB unless you have a specific reason (e.g., uncertain susceptibility, very high inoculum) (Kaye et al., 2023).

So — mono or combination?

• For CRAB and Enterobacterales: RCT evidence (AIDA, Kaye 2023) does not support routine combination with carbapenem
• Combinations may still have a role: when polymyxin MIC is at or above the breakpoint, when source control is delayed, in pneumonia with poor lung penetration
• Combination with newer agents (ceftazidime-avibactam + colistin in salvage settings) is observational only
• Default in 2026: if a newer β-lactam works, use it as monotherapy; if you must use a polymyxin, use it alone unless a specific reason exists to combine

The clinical posture has shifted. Combination is no longer the default — it’s the exception. The exception cases (heteroresistance, very high inoculum, no source control) still exist but are narrower than we used to think.

Cefiderocol — The siderophore cephalosporin

• Catechol siderophore moiety hijacks bacterial iron transport (TonB-dependent receptors) → bypasses outer membrane permeability barriers
• Active against most CRE (including NDM), CRPA, CRAB, Stenotrophomonas
• CREDIBLE-CR (Bassetti 2021): mortality numerically higher in cefiderocol arm vs best available therapy in nosocomial pneumonia — unexplained signal
• APEKS-NP (Wunderink 2021): non-inferior to high-dose meropenem in nosocomial pneumonia
• Use as monotherapy for severe MDR Gram-negative infection where active; preferred over polymyxin where availability and cost permit
• Mixed adoption — the CREDIBLE-CR mortality signal made some centers cautious

Cefiderocol is one of the strongest competitors to polymyxin for MDR Gram-negative infection. The unexplained mortality signal in CREDIBLE-CR is real and has slowed broader adoption — different trials, different patient populations, but the signal can’t be entirely dismissed.

Inhaled and intrathecal routes

Inhaled polymyxins

• CMS aerosolized via vibrating-mesh or jet nebulizer
• Long-standing use in cystic fibrosis for chronic P. aeruginosa colonization (Ratjen)
• Adjunctive role in ventilator-associated pneumonia caused by MDR Gram-negatives — Lu 2012 demonstrated efficacy of high-dose nebulized colistin
• Dosing (CF, EMA insert): 1–2 MIU (33–66 mg CBA) three times daily for ≥2 years
• VAP adjunct dosing: highly variable; 75–150 mg CBA q12h commonly used
• Premix only immediately before nebulization (bronchoconstriction risk; CMS hydrolysis in vitro)

(Lu et al., 2092; Ratjen et al., 2006; Smet et al., 2009)

Inhaled colistin — practical notes

• Use vibrating-mesh nebulizer (more efficient, less drug loss) where available
• Premix only immediately before nebulization — formed colistin in standing solution causes bronchospasm and increases toxicity
• Pretreat with short-acting β2-agonist to prevent bronchoconstriction
• Monitor for bronchospasm, cough, dyspnea
• Systemic absorption from inhaled CMS is real but minor — does not substitute for IV therapy in pneumonia
• Role in VAP: adjunct only; meta-analyses suggest modest microbiologic benefit, uncertain clinical impact

The nebulizer choice matters more than people realize. Vibrating mesh delivers about 60% of the dose to the lungs; jet nebulizer maybe 10%. If your respiratory therapy team is using jet, the patient is getting a fraction of what you ordered.

Intrathecal / Intraventricular Polymyxin

• For CSF shunt and external ventricular drain infections with multidrug-resistant Gram-negative pathogens
• Polymyxin penetrates CSF poorly with IV administration → topical route
• 2019 Consensus Guidelines: intraventricular CMS 125,000 IU (4.1 mg CBA) daily
• Polymyxin B 5–10 mg intrathecally also used
• Aseptic chemical ventriculitis can complicate (fever, CSF pleocytosis without infection)
• Specialist neurosurgical / neuro-ID territory

(Tsuji et al., 2019)

Other Considerations

Selective decontamination of the digestive tract (SDD)

• de Smet 2009 (NEJM): SDD with topical polymyxin + tobramycin + amphotericin B in oropharynx and gut reduced ICU mortality
• Subsequent trials (e.g., SuDDICU 2022) showed more modest effects
• Adoption uneven; widespread in Dutch and German ICUs, controversial in US
• Polymyxin’s role here exploits gut decontamination without systemic absorption (oral polymyxin is essentially non-absorbed)
• Resistance concerns continue to limit uptake

(Smet et al., 2009)

Source control aatters

• Polymyxin efficacy depends heavily on adequate source control:
  • Drained abscess
  • Removed infected catheter
  • Debrided wound
  • Resolved obstruction
• Without source control, polymyxin alone rarely succeeds — the bug regenerates faster than the drug can suppress it
• Mortality data: source-controlled CR-Gram-negative bacteremia has dramatically better outcomes than uncontrolled
• ID consult role: insist on source control discussion at every polymyxin start

This is generic ID wisdom but especially true for polymyxins given the marginal PK/PD index. A polymyxin in an uncontrolled source is rarely a winning combination.

Polymyxin B Hemoperfusion

• Toraymyxin / PMX-DHP — extracorporeal columns containing immobilized polymyxin B fibers
• Binds circulating endotoxin via lipid A
• Used in Japan and parts of Europe for endotoxic septic shock
• EUPHRATES (Dellinger 2018, JAMA): primary endpoint negative; post-hoc subgroup (endotoxin activity 0.6–0.9) suggested possible benefit
• Not routinely used in US ICUs; remains controversial
• Not the same as IV polymyxin — completely different mechanism (endotoxin binding, no antimicrobial effect on the patient’s flora)

Worth a brief mention for completeness — patients who arrive on Toraymyxin from another center are not receiving polymyxin antimicrobial therapy. The mechanism is endotoxin neutralization, not infection treatment.

Place in 2026

Polymyxin B preferred — Practical Examples

Polymyxin B vs CMS practical decisions

Clinical scenario	Preferred agent	Why
CRE bacteremia	Polymyxin B	Faster, more predictable plasma concentrations
CRAB pneumonia (when SUL-DUR not available)	Polymyxin B ± inhaled colistin	Better PK predictability
CR-Pseudomonas catheter-related BSI	Polymyxin B	Non-renal clearance; line management more important than antibiotic choice
Lower UTI / cystitis	Colistin (CMS)	Urinary CMS hydrolysis delivers active drug locally
CSF shunt infection	Intraventricular CMS	Direct administration; polymyxin B not standardized intrathecally
CRRT patient	Either; polymyxin B simpler dosing	Avoids the CMS conversion variability

The CRRT case is interesting — both work but polymyxin B is logistically simpler because you don’t have the CMS-conversion uncertainty in addition to the dialytic clearance question (Tsuji et al., 2019)

The Modern hierarchy for CR gram-negatives

Polymyxin role in 2026

Pathogen	First-line (where available)	Polymyxin role
KPC	Meropenem-vaborbactam, ceftazidime-avibactam, imipenem-relebactam	Salvage
NDM	Ceftazidime-avibactam + aztreonam, cefiderocol	Salvage
OXA-48	Ceftazidime-avibactam	Salvage
CRAB	Sulbactam-durlobactam, cefiderocol	Adjunct / where newer unavailable
CRPA	Ceftolozane-tazobactam, ceftazidime-avibactam, imipenem-relebactam, cefiderocol	Salvage / inhaled adjunct
CR UTI	Colistin (CMS)	First-line by route

Italian epidemiology — CRAB and CRE

• Italy and Greece historically have highest CRAB rate in Europe (>80% of Acinetobacter isolates in some surveys)
• KPC-producing K. pneumoniae dispersed across Italian ICUs since mid-2010s
• Colistin remained the default agent through the late 2010s
• Recent Italian uptake of ceftazidime-avibactam, meropenem-vaborbactam, and cefiderocol has reduced polymyxin use — but not eliminated it
• mcr genes documented in Italian livestock and clinical isolates — surveillance ongoing

The Italian CRAB and CRE burden is one of the reasons polymyxin pharmacology is still front-of-mind in Italian ICUs. The shift to newer agents has been incomplete

Polymyxin allergy

• Polymyxin hypersensitivity is rare but does occur — type I (urticaria, angioedema, anaphylaxis) and type IV reactions documented
• Topical polymyxin (e.g., Neosporin) is a common cause of contact dermatitis
• Cross-reactivity between polymyxin B and colistin is essentially complete
• No reliable skin testing protocol
• True systemic allergy → desensitization can be considered but expertise is limited
• Suspected allergy in a clinical record should not be dismissed but should be evaluated carefully — alternatives may carry greater risk

When to refuse a polymyxin recommendation

• β-lactam alternative active and available → use the β-lactam
• Isolate intrinsically resistant — Proteus, Providencia, Morganella, Serratia, Burkholderia, any Gram-positive, anaerobes
• Baseline AKI with no source control — toxicity expected, benefit unlikely
• Source amenable to control — drainage of a CRE-positive abscess often resolves the issue without escalation
• Patient already on combination of nephrotoxins — vancomycin + colistin + IV contrast is a recipe for dialysis
• Polymyxin MIC at or above the breakpoint — efficacy is unreliable; consider a different agent or combination

The consult that gets pushback most often: “ICU wants colistin started for the CR Klebsiella in the sputum, but the patient has trach colonization, not pneumonia, and there’s a working antibiotic alternative.” Declining a polymyxin in such cases is good stewardship.

Key Pearls

What I want you to remember — Pharmacology

• Colistin is given as CMS, an inactive prodrug that hydrolyzes slowly to colistin
• Polymyxin B is the active drug — predictable PK, non-renal clearance, no renal adjustment
• Loading dose is mandatory in critically ill patients — without it, the first 48 hours are subtherapeutic
• Unit confusion (IU vs. CBA vs. mg CMS) has killed patients — always specify the unit on the order
• Polymyxin B preferred over CMS for systemic infection; CMS preferred for UTI

What I want you to remember — Resistance

• Intrinsic resistance list: Proteus, Providencia, Morganella, Serratia, Burkholderia, all Gram-positives, anaerobes
• Lipid A modification (L-Ara4N, pEtN) is the dominant acquired mechanism, chromosomally regulated via PmrAB/PhoPQ
• mcr genes (2015–) brought plasmid-mediated, horizontally transferable colistin resistance — globally distributed, sometimes co-located with carbapenemases
• Heteroresistance is common in CRAB and CRKP — explains within-patient resistance emergence on monotherapy
• Susceptibility testing is unreliable — drug adsorbs to plastic; broth microdilution with polysorbate-80 is the reference method

What I want you to remember — Clinical

• 30–60% nephrotoxicity — therapeutic and toxic concentrations overlap
• Polymyxin B causes diffuse hyperpigmentation — counsel patients
• AIDA (Paul 2018) and Kaye 2023: colistin + meropenem does not improve outcomes vs. colistin alone for CR-Gram-negatives, and adds nephrotoxicity
• Inhaled colistin has a role in CF and VAP adjuncts; intrathecal CMS for shunt infections
• In 2026, polymyxins are reserve agents — use a newer β-lactam/inhibitor when active and available

Topical and ophthalmologic uses

• Topical (skin) polymyxin B + bacitracin ± neomycin — minor wounds, post-procedural prophylaxis
• Ophthalmic — bacterial conjunctivitis, blepharitis (polymyxin B + trimethoprim drops)
• Otic — external otitis (polymyxin B + neomycin + hydrocortisone), provided tympanic membrane intact
• Topical contact dermatitis is the dominant adverse effect
• These uses predate the IV resurgence and remain widespread — most pharmacists encounter polymyxin chiefly in topical form

The topical applications are a useful reminder that polymyxin B has been quietly in use the whole time, even when the IV form was off the formulary. The “Bacitracin allergy” patients are often actually reacting to neomycin or polymyxin in the same ointment.

You cannot step down to oral

• Polymyxins are essentially non-absorbed when given orally
• No oral step-down option exists — once you commit to IV polymyxin therapy, the patient is on IV until completion
• Implications:
  • Long courses require sustained IV access
  • Outpatient parenteral antibiotic therapy (OPAT) possible but requires daily-multiple infusions and TDM/safety monitoring
  • Duration choices should weigh the IV-access burden
• Distinct from many other Gram-negative–active agents (fluoroquinolones, TMP-SMX, doxycycline, oral fosfomycin) where step-down is feasible

The “you cannot go oral” point is easy to forget when reasoning about duration. Patients with KPC bacteremia who clinically improve on day 5 still need 7–8 more days of IV; there’s no oral exit.

Stewardship posture

• Always ask: is there a safer or more effective alternative?
• De-escalate aggressively once susceptibilities and source control allow
• Document the indication explicitly in the chart
• Short courses where evidence supports (5–7 days for uncomplicated bacteremia with source control may be feasible)
• Avoid prolonged empirical use while awaiting cultures — high toxicity for unproven benefit
• Local cumulative antibiograms for CR organisms should inform empirical regimens — when newer agents are reliably active locally, polymyxin should not be empirical

Polymyxin use is the kind of decision that should be reviewed prospectively by the ID team. In our institution, an explicit stewardship audit of every new polymyxin start would catch a non-trivial proportion of unnecessary or over-long courses.

Future agents

• Polymyxin analogues with reduced nephrotoxicity in development (e.g., SPR741, CB-182,804) — early-phase only
• Combination repurposing — polymyxin + rifampin, fosfomycin, ceftazidime-avibactam under investigation; not RCT-validated
• Phage therapy for pan-resistant Gram-negatives — case reports, no RCT data yet
• Antimicrobial peptides (LL-37 derivatives) and PMB-derivatized nanoparticles in preclinical testing
• The polymyxin pipeline is thin — improvements likely incremental, not paradigm-shifting

The polymyxin pipeline is small. Phage therapy and antimicrobial peptides are the more exciting frontiers; improved polymyxin analogues may emerge but the basic toxicity is mechanistic, not easily engineered away.

References

Bergen PJ, Forrest A, Bulitta JB, et al. Clinically relevant plasma concentrations of colistin in combination with imipenem enhance pharmacodynamics activity against multidrug-resistant pseudomonas aeruginosa at multiple inocula. Antimicrob Agents Chemother 2011;55:5134–42.

Bergen PJ, Li J, Rayner CR, et al. Colistin methanesulfonate is an inactive prodrug of colistin against pseudomonas aeruginosa. Antimicrob Agents Chemother n.d.

Betrosian A. P Frantzeskaki F, Xathaki A, et al. Efficacy and safety of high-dose ampicillin/sulbactam vs. Colistin as monotherapy for the treatment of multidrug resistant acinetobacter baumannii ventilator-associated pneumonia. J Infect 2008;56:432–6.

Beyrouthy R, Robin F, Lessene A, et al. MCR-1 and OXA-48 in vivo acquisition in KPC-producing escherichia coli after colistin treatment. MCR-1 and OXA-48 in Vivo Acquisition of KPC-Produeing Escherichia Coli After Colistin Treatment 2017;61:e02540–16.

Cai Y, Chai D, Wang R, et al. Colistin resistance of acinetobacter baumannii: Clinical reports, mechanisms and antimicrobial strategies. J Antimicrob Chemother 2012;7:1607–15.

Cheah SE, Wang J, Nguyen VT, et al. New pharmacokinetic/pharmacodynamic studies of systemically administered colistin against pseudomonasy@eruginosa and acinetobacter baumannii in mouse thigh and lung infection models: Smaller response in lung infection. J Antimicrob Chemother 2015;70:3291–7.

Forrest A, Garonzik SM, Thamlikitkul V, et al. Pharmacokinetic/toxicodynamic analysis of colistin-associated acute kidney injury in critically ill patients. Antimicrob Agents Chemother 2017;61:e01367–17.

From AH, Fong JS, Good RA. Polymyxin b sulfate modification of bacterial endotoxin: Effects on the development of endotoxin shock in dogs. Infect Immun 1979;23:660–4.

Garonzik SM, Li J, Thamlikitkul V, et al. Population pharmacokinetics of colistin methanesulfonate and formed colistin in critically ill patients from a multicenter study provide dosing strategies for variogytategories of patients. Antimicrob Agents Chemother 2011;55:3284–94.

Hartzell JD, Kim AS, Kortepeter MG, et al. Acinetobacter pneumonia: A review. Medgenmed 2007;9:4.

Hartzell JD, Neff R, Ake J, et al. Nephrotoxicity associated with intravenous colistin (colistimethate sodium) treatment at a tertiary care medical center. Clin Infect Dis 2009;48:1724–8.

Kaye KS, Marchaim D, Thamlikitkul V, et al. Colistin monotherapy versus combination therapy for carbapenem-resistant organisms 2023.

Kaye KS, Pogue JM, Tran TB, et al. Agents of last resort: Polymyxin yoane. Infect Dis Clin North Am 2016;30:391–414.

Kubin F Elimen TM, Phadke V, et.al. Incidence and predictors of acute kidney injury associated with intravenous»polymyxin b therapy. J Infect 2012;65:80–7.

Landersdorfer CB, Wang J, Wirth V, et al. Pharmacokinetics/pharmacodynamics of systemically administered polymyxin b against klebsiella pneumoniae in mouth thigh and lung infection models. J Antimicrob Chemother 2018;73:462–8.

LiJ, Nation RL. Defining the dosage units for colistin methanesulfonate: Urgent need for international harmonization. Antimicrob Agents Chemother 2006;50:4231–2.

LiJ, Rayner CR, Nation RL, et al. Heteroresistance to colistin in multidrug-resistant acinetobacter baumannii. Antimicrob Agents Chemother 2006;50:2946–50.

Linden PK, Parenteral PDL, Clin Infect Dis. 2006;43(suppl 2):S89- S94. doi: 10.1086/504485. 43. de Smet AMGA inhaled colistin for treatment of ventilator- associated pneumonia., Kluytmans JAJW, Cooper BS, et al. Decontamination of the digestive tract and oropharynx in ICU patients. N Engl J Med 2009;360:20–31.

Lu Q, Luo R, Bodin L, et al. Efficacy of high-dose nebulized colistin in ventilator+ associated-pneumonia caused by multidrug-resistant pseudomonas aeruginosa and acinetobacter baumannii. Anesthesiology 2092;117:1335–47.

Maragakis LL, Perl TM. Acinetobacter baumannii: Epidemiology, antimicrobial resistance, and treatment options. Clin Infect Dis 2008;46:1254–63.

Mattos KP, R LloretG, Cintra ML, et al. Acquired skin hyperpigmentation following intravenous polymyxin b treatment: A cohort study. Pigment Cell Melanoma Res 2016;29:888–390.

Meletis G, Tzampaz E, Sianou E, et al. Colistin heteroresistance in carbapenemase-producing klebsiella pneumoniae. J Antimicrob Chemother 2011;66:946–7.

Nang SC, Li J, Velkov T. The rise and spread of mcr plasmid-mediated polymyxin resistance. Crit Rev Microbiol 2019;45:131–61.

Nation RL, Garonzik SM, Li J, et al. Updated US and european dose recommendations for intravenous colistin: How do they perform? Updated US and European Dose Recommendations for Intravenous Colistin: How Do They Perform? Clin Infect Dis 2016;62:552–8.

Nation RL, Garonzik SM, Thamlikitkul V, et al. Dosing guidance for intravenous colistin in critically-ill patients. Clin Infect Dis 2017;64:565–71.

Oliveira MS, Prado GV, Costa SF, et al. Ampicillin/sulbactam compared with polymyxins for the treatment of infections caused by carbapenem-resistant acinetobacter spp. 2008.

Owen RJ, Li J, Nation RL, et al. In vitro pharmacodynamics of colistin against acinetobacter baumannii clinical isolates. J] Antimicrob Chemother 2007;59:473–3.

Paul M, Daikos GL, Durante-Mangoni E, et al. Colistin alone versus colistin plus meropenem for treatment of severe infections caused by carbapenem-resistant gram-negative bacteria: An open-label, randomised controlled trial. Lancet Infect Dis 2018;18:391–400.

Plachouras D, Karvanen M, Friberg LE, et al. Population pharmacokinetic analysis of colistin methanesulfonate and colistin after intravenous administration in critically ill patients with infections*caused by gram-negative bacteria. Antimicrob Agents Chemother 2009;53:3430–6.

Pogue JM, Jones RN, Bradley JS, et al. Polymyxin susceptibility testing and interpretive breakpoints: Recommendations from the united states committee on antimicrobial susceptibility testing (USCAST). Polymyxin Susceptibility Testing and Interpretive Breakpoints: Recommendations from the United States Committee on Antimicrobial Susceptibility Testing (USCAST) 2020;64:e01495–19.

Pogue JM, Lee J, Marchaim D, et al. Incidence of and,risk factors for colistin- associated nephrotoxicity in a large academic health system. Clin Infect Dis 2011;53:879–84.

Poudyal A, Howden B. P Bell JM, et al. In vitro pharmacodynamics of colistin against multidrug-resistant klebsiella pneumoniae. J Antimicrob Chemother 2008;62:1311–8.

Ratjen F, Rietschel E, Kasel D, et al. Pharmacokinetics of‘inhaled colistin in patients with cystic fibrosis. J Antimicrob Chemother 2006;57:306–11.

Sandri AM, Landersdorfer CB, Jacob J, et al. Population pharmacokinetics of intravenous polymyxin b in critically ill patients: Implication for selection of dosage regimens. Clin Infect Dis 2013;57:524–31.

Smet AMGA de, Kluytmans JAJW, Cooper BS, et al. Decontamination of the digestive tract and oropharynx in ICU patients. N Engl J Med 2009;360:20–31.

Sorli L, Luque S, §, et al. Trough colistin plasma level is an independent risk factor for nephrotoxicity: A prospective observational cohort study. BMC Infect Dis 2013;13:380.

Tsuji BT, Pogue JM, Zavascki AP, et al. International consensus guidelines for the optimal use of the polymyxins: Endorsed by the american college of clinical pharmacy (accp), european society of clinical microbiology and infectious diseases (escmid), infectious disease society of america (idsa), international society for anti- infective pharmacology (isap), society of,critical care medicine (sccm), and society of infectious diseases pharmacists (SIDP). Pharmacotherapy 2019;39:10–39.

Y-Y L, Wang Y, Walsh TR, et al. Emergence of plasmid-mediated colistin resistance mechanism MCR-1 in animals and human beings in china: A microbiological and molecular biological study. Lancet Infect Dis 2016;16:161–8.

Zavascki AP, Manfro RC, Maciel RA, et al. Head and neck hyperpigmentation probably associated with polymyxin b therapy. Ann Pharmacother 2015;49:1171–2.

Zhong LL, Zhang YF, Doi Y, et al. Coproduction of MCR-1 and NDM-1 by colistin-resistant escherichia coli isolated from a healthy individual n.d.

Zhou Y, Li Y, Xie X, Song L, Lan G, Sun B, et al. Higher incidence of neurotoxicity and skin hyperpigmentation in renal transplant patients treated with polymyxin b. Br J Clin Pharmacol 2022;88:4742–50.