Antimicrobial agents: December 2018

Rapid IV push of cefepime is a risk factor for neutropenia in patients receiving OPAT.  I suppose that means that high peak concentrations of cefepime induce neutropenia?  The authors identified a cluster of patients receiving cefepime via OPAT who developed neutropenia.  To follow up on this, they reviewed the records of all non-neutropenic adults receiving OPAT from a single home care service between 2016 and May 2018.  Patients with missing lab data or who had recently received chemotherapy likely to induce neutropenia were excluded.  They defined cefepime-induced neutropenia as a drop in the ANC from a baseline >1700 cells/ul to <1000 during receipt of cefepime, followed by recovery of the ANC after withdrawal of cefepime.

A total 326 patients received 402 courses of cefepime during the study period.  The incidence of cefepime-induce neutropenia increased during the study period, from 0.9% in the first 18 months of the study to 5.4% in the last 12 months.  Among the patients who did versus did not develop cefepime-induced neutropenia in the latter period, the cumulative doses were similar (103g vs 136g) but the durations were more often >2wk in the neutropenia group (p=0.014) and these patients were more than three times as likely to have received their cefepime via IV push (78% vs 25%; p=0.002).  In multivariate regression analysis, receipt of cefepime by IV push emerged as the predominant risk factor (RR 9.3; 95% CI 2 to 43).  No one who developed cefepime-induced neutropenia died.

In conclusion: administering cefepime via IV push is associated with neutropenia.  It’s also, pharmacokinetically speaking, the absolute worst way to dose that drug, so maybe don’t do it. 30590400


Cefiderocol is as or more effective than imipenem-cilastin for complicated gram-negative urinary tract infections.  Cefiderocol is the first siderophore-antibiotic conjugate drug to get anywhere near FDA approval. If “siderophore-antibiotic conjugate” is a new term for you, the concept is pretty cool.  Basically, gram-negative bacteria need iron to grow and actively scavenge it from the environment using iron-binding small molecules called siderophores. These drugs have built-in siderophores that they use to “Trojan horse” their way through bacterial active transport channels and into the bacterial periplasmic space.  This has the effect of dramatically enhancing concentrations of antibiotic at the site of action (the bacterial cell wall).  If that wasn’t enough, cefiderocol is also stable to all known carbapenemases and active against carbapenem-resistant P. aeruginosa, S. maltophilia, and A. baumannii.  Be still my beating heart!

Anyway, for some reason every new gram-negative antibiotic has be shown to be non-inferior to another antibiotic for the treatment of complicated UTI before the FDA will let us turn around and use it for HAP/VAP and CLABSI. In this case, Shionogi (who is developing cefiderocol and funded the study) decided to put their new drug up against imipenem-cilastin.  The authors recruited adults who had either complicated UTI with or without pyelonephritis or acute uncomplicated pyelonephritis for a double-blind randomized-controlled trial.  They allowed the enrollment of immunosuppressed patients, including renal transplant patients, to capture more patients with MDR infections.  The two main exclusion criteria were polymicrobial, carbapenem-resistant, or fungal organisms on urine culture and end-stage renal disease.  Patients were randomized 2:1 to receive the renally dose-adjusted equivalent of either cefiderocol 2g IV every 8 hours or imipenem-cilastin 1g IV every 8 hours for a clinician-directed duration of 7-14 days. No sequential or oral step-down therapy was allowed. The primary endpoint was a composite of clinical and microbiologic response at a test of cure visit 5-9 days after finishing antibiotic therapy.

A total 452 patients were enrolled in the study from 67 hospitals in 15 countries; 371 patients (cefiderocol, n=252; imi-cilastin, n=119) had a qualifying uropathogen and were included in the modified intention-to-treat analysis.   The baseline characteristics of the two groups were similar; about 20% in each group had clinically severe disease, about 20% in each had moderate to severe renal disease, and about 33% in each had obstructive uropathy.  Uropathogens isolated were also similar, including ~60% E.coli,  ~20% K.pnuemoniae, ~5% P.aeruginosa, and a smattering of other organisms.  Median duration of treatment was 9 days for both groups. 

At the test of cure visit, more patients in the cefiderocol group than the imi-cilastin group had reached the primary outcome (73% vs 55%; adjusted treatment difference 18.6% with p<0.001).  This difference was driven by higher rates of microbiologic response with cefiderocol (74% vs 56%) rather than a  difference in clinical response (90% vs 87%).  Sustained microbiologic eradication was also more frequent in the cefiderocol group (57% versus 44%; 95% CI of treatment difference 3.% to 24.6%).   Rates of adverse events and severe adverse events were similar between groups, though those percentages were numerically lower for cefiderocol in every case; the most common side effect of cefiderocol was diarrhea (4%).

In summary, cefiderocol is at least as efficacious as imipenem-cilastin for the treatment of complicated UTI and pyelonephritis, and is significantly better at achieving microbiologic cure.  I suppose it makes sense to trial new antibiotics in a “lower stakes” setting like UTI, but it’s frustrating to not get the version of this study that’s cefiderocol versus best available therapy for MDR gram-negative HAP/VAP and bacteremia – because that’s how we’re going to use this drug, right?  That’s neither the authors’ nor Shionogi’s fault, though. 30509675


Both ceftazidime-avibactam and ceftolozane-tazobactam are fairly reliable agents for MDR P. aeruginosa in the United States.  Usually I pass on these “we tested a gajillion clinical isolates for susceptibility to our drug!” studies, because I think local antibiograms more meaningfully inform our empiric antibiotic selections.  However, I gave this study a look because I’d been taught “ceftaz-avi is the CRE drug and ceftolo-tazo is the MDR Pseudomonas drug” and I was curious to see whether the data bore that out.

The authors collected 1,909 P. aeruginosa isolates from 70 US medical centers in 2017, restricting themselves to isolates that the centers’ micro labs deemed likely to be causes of infection by local criteria.  Species-level identifications were confirmed by MALDI-TOF, and then the authors determined the antimicrobial susceptibility of each agent via microbroth dilution.

The less interesting results of this study were the overall susceptibility rates, which were 97% for ceftaz-avi and 98% for ceftolo-tazo.  Other than both agents being less reliable on the west coast, these numbers didn’t vary much geographically (see the pretty map in Fig 1 for details).  The more interesting results are in Table 1, where the authors stratify the isolates by resistance to other beta-lactams: specifically, they give us data for the subset of MDR P. aeruginosa resistant to ceftazidime, cefepime, meropenem, and pip-tazo.  Among these isolates, both drugs have fair activity, but susceptibility rates are clearly higher for ceftolo-tazo (79%) than for ceftaz-avi (70%).  The only more reliably active drug was colistin (100%); amikacin was about as reliable as ceftolo-tazo (82%) and tobramycin was about as reliable as ceftaz-avi (71%).  Gentamicin was a lousy choice (48%), as were ciprofloxacin (31%) and levofloxacin (23%) – which is just more evidence that empiric double coverage adding a quinolone to a carbapenem is nonsense because it doesn’t meaningfully improve probability of appropriate empiric therapy.

So yes, both ceftaz-avi and ceftolo-tazo are highly active against P. aeruginosa as a whole, but between the two ceftolo-tazo is clearly the more reliable agent for MDR P. aeruginosa. 30224535


Rifaximin prophylaxis promotes colonization with rifampin-resistant staphylococci.  This makes sense, because both are rifamycins, and I think it’s important because we use a lot of rifaximin as prophylaxis for hepatic encephalopathy in patients with decompensated cirrhosis.  The authors examined a cohort of patients undergoing colorectal surgery in Spain (n=74), each of whom received preoperative prophylaxis with three doses of 200mg rifaximin every two hours.  They collected and cultured perianal and nasal swabs from each patient in the perioperative period as well as two and four weeks after surgery, and those patients who acquired a rifampin-resistant staph were followed with monthly swabs until their cultures turned negative.

No rifampin-resistant staphylococci were isolated from any of the preoperative cultures, but 43% of patients had become colonized 2 weeks later, 20% were still colonized at 4 weeks, and two patients remained colonized at 8 weeks; by 12 weeks, everyone had become decolonized. Interestingly, all of the rifampin-resistant isolates were coagulase-negative staphylococci, mostly S. epidermidis.  This could simply reflect CONS having been present in greater quantities versus S. aureus on patients’ skin to begin with, or perhaps CONS have some genetic predisposition to be able to develop rifamycin resistance more quickly.

This study suggests that rifaximin ought to be avoided where possible in patients who are at risk for indwelling device infections that might need rifampin in the future (e.g. prosthetic joint infection, prosthetic valve endocarditis).  If >40% of patients became colonized with rifampin-resistant CONS after three doses of rifaximin, probably 100% of patients on prolonged rifaximin become colonized, and it seems reasonable that eventually patients’ S. aureus would become resistant to rifampin as well.  The good news is that the resistant organisms were all found on perianal swabs, so the authors’ findings may not necessarily be generalizable to the rest of the skin microbiome (i.e. the poop, and hence the rifaximin, has more contact with the bacteria on the perianal skin than the rest of the body unless you have very poor hand hygiene indeed). And in fact, the authors point out that a previous study suggests the emergence of rifampin resistance might be restricted to the perianal skin flora. 30249689