August 2019: Antimicrobial Agents

Are noncarbapenem antibiotics effective for bacteremia due to Enterobacteriaceae producing ESBL or AmpC beta-lactamases? This is a fairly common question from non-ID clinicians, and it’s not an entirely baseless worry.  First, antibiotic resistance genes travel in clusters; for example, exposure to a fluoroquinolone can predispose a patient to having an ESBL-producing isolate later.  Even in cases when the antibiotic susceptibilities are known, resistance genes that are present but deactivated by transposons or chance mutation could revert to an active state under selective pressure. For example, if the ID community on Twitter is any indication, even specialists think twice before prescribing cefazolin to a patient with a phenotypically methicillin-susceptible S.aureus that PCR-based testing shows is harboring the mecA gene.  Anyway, wouldn’t it be nice to have some clinical data to point to when promising your colleagues that that awful MDR ESBL E.coli that’s somehow still susceptible to quinolones can be treated with ciprofloxacin instead of meropenem?

The authors conducted a retrospective review of patients treated at their hospital over an 11-year period.  They enrolled patients with bacteremia due to either ESBL- or AmpC-producing Enterobacteriaceae, excluding those who died within 72hr or had less than one month of subsequent followup; they identified patients using the hospital’s electronic records.  Their primary outcome of interest was 30-day mortality, and they stratified patients by the type of antibiotic prescribed.

The study included a total 101 patients, of whom 59 received a carbapenem and 42 received an alternative therapy – TMP/SMX in 25 cases, quinolones in 9 cases, aminoglycosides in 5 cases, and other drugs in 3 cases.  The patients had mean ages 72-79, presented with septic shock in 17-25% of cases, had infections due to either E.coli (80%) or K.pneumo (20%), and had mostly urinary sources of infection (58-71%).  When comparing the patients treated with carbapenems or alternative agents, those given alternative agents were more likely to have had a delay in appropriate therapy (2 vs 0 days; p<0.0006); nonetheless, they had shorter hospitalizations (7 vs 12 days; p<0.0001), similar rates of clinical cure (95% vs 85%), and similar rates of mortality (5% vs 10% favoring alternative treatment; p=0.46).  In unadjusted analysis, receipt of a noncarbapenem antibiotic was not associated with mortality (OR 0.27 with 95% CI 0.05 – 1.61); and in fact, after propensity score adjustment, carbapenem prescription seemed a risk factor for mortality (OR 4.95 with 95% CI 0.94-26.01).  Might this reflect physicians assigning treatment based on several of illness not captured in Table 1 (i.e. sick patients stay on a carbapenem while the folks who got better switch to orals)?  Certainly.  But, at least for now we can say that the best available evidence indicates that oral non-carbapenem antibiotics (predominantly TMP/SMX and quinolones) work at least as well as carbapenems for ESBL- and AmpC producing Enterobacteriaceae. 31075401

Ceftriaxone (at standard doses) is inferior to the new anti-staphylococcal cephalosporins for MSSA pneumonia. Count me among those skeptical of ceftriaxone as a treatment for methicillin-susceptible S.aureus (MSSA).  Aside from personal anecdotes of clinical failure, I’ve found the pharmacokinetic data for ceftriaxone at doses less than 2g every 12 hours not encouraging.  Here’s a roundup of clinical data that proves equally uninspiring. The authors performed a systematic review and meta-analysis of RTCs that compared clinical outcomes in patients given ceftriaxone or either ceftaroline or ceftobiprole (another of the new anti-MRSA cephalosporins, which is not currently approved in the US) for pneumonia due to MSSA.  They excluded trial enrollees who received ceftriaxone plus vancomycin.  The authors took clinical cure as their primary outcome.  They wrote that they were not able to do subgroup analysis or identify publication bias by funnel plot due to small sample size.

The authors identified 5 RTCs (four in adults, one in children; four with ceftaroline, one with ceftobiprole) with a total 77 patients (39 in the ceftriaxone arms vs 38 in the new agent arms).  All but one were double blind; the last was blinded only to the observers.  All were drug company funded trials, all used similar durations of therapy (5-7 days in three trials, 3-7 days in one trial, and 7.5 vs 9 days in one trial) and all defined clinical cure as total resolution of signs & symptoms of infection with no further need for antibiotic therapy.  The ceftriaxone doses used were 35mg/kg in the pedi study and either 1g or 2g daily in the adult studies (n=2 each).  Patients given ceftriaxone had a lower rate of clinical cure vs those given the newer drugs (total weighted risk difference 28.5% with 95% CI 3.4% to 53.5%).

These are not high quality data – the authors have aggregated several small trials with single-digit numbers of patients meeting the inclusion criteria per arm.  However, previous studies of serious MSSA infections have found similar results, and there are also no high-quality data demonstrating the non-inferiority of ceftriaxone to traditional anti-MSSA agents like nafcillin/oxacillin or cefazolin, so I remain hesitant to use ceftriaxone for serious MSSA infections. 31173864

in vitro data suggests that ceftaz-avibactam synergizes well with multiple agents for carbapenem-resistant K.pnuemoniae and P.aeruginosa.  We’ve talked about the ceftaz-avi + aztreonam for metalo-betalactamases story a few times, but what about combining ceftaz-avi (CZA) with other drugs?  This paper out of the Rybak lab at Wayne State looked at combining CZA with a number of other drugs, including aztreonam (AZT), meropenem (MEM), amikacin, colistin, and fosfomycin for 21 strains each of carbapenem-resistant K.pneumoniae and carpanemen-resistant P.aeruginosa using MIC reduction testing and time-kill assays.  For those of you who can parse beta-lactamase acronym salad, the K.pneumo isolates were mostly carrying the resistance gene KPC alongside a smattering of CXT-M, TEM, SHV, and in a few cases OXA; most of the P.aeruginosa isolates had no additionally acquired beta-lactamases, though there were a few carrying OXAs, IMP, or KPC  (note that this doesn’t mean these P.aeruginosa strains were walking around unarmed – a bunch had carbapenem resistance mediated by OprD porin mutations or overexpression of chromosomal beta-lactamases like AmpC).

The MIC reduction assays are summed up in a really nice figure 1, which is a bar graph showing the mean and standard deviation of the fold-reduction in CZA MIC with each tested combination, as well as the MIC50 of each combination, for K.pnuemo and P.aeruginosa.  Long story short, the most effective combinations for K.pneumo were CZA-AZT, CZA-amikacin, and CZA-MEM, whereas the only meaningfully synergistic combination for P.aeruginosa was CZA-MEM; fosfomycin and colistin were not particularly synergistic for either organism.  In the time-kill analysis (done with 2 strains each of K.pneumo and P.aeruginosa), CZA was synergistic with amikacin, AZT, and fosfomycin for both strains of K.pneumo, with meropenem and colistin for only one strain of K.pneumo, with amikacin, AZT, and meropenem for both P.aeruginosa strains, and for colistin with one P.aeruginosa strain.

So what does this mean? Well, for carbapenem-resistant K.pnuemonia, CZA-AZT rseems like the way to go: not only does it overcome resistance in metallo-betalactamases, it offers synergistic activity for KPC based on data from both assays.  For P.aeruginosa, CZA-MEM seems to be the most consistently synergistic combination.  And for both organisms, it looks like the most reliable alternative combination is CZA-amikacin.  Here’s hoping someday we get clinical validation of these in vitro findings. 31182535

High dose trimethoprim-sulfa plus clindamycin yielded clinical outcomes equivalent to standard IV therapy in this before-after study of patients with S.aureus endocarditis.  So at the beginning of the year we had the POET trial, which showed us that in patients with infective endocarditis, switch to PO antibiotics after at least 10 days of IV therapy was non-inferior to continued IV antibiotic treatment with regard to a composite outcome including death and various measures of treatment failure.  One of the criticisms of this study is that it had few injection drug users and no cases of MRSA.  Well, here we have a study to fill in those gaps!

This study out of France used a before-after design.  In 2012, the authors’ hospital instituted a new treatment strategy for patients with S.aureus endocarditis at their hospital, consisting of one week of IV therapy with high dose TMP/SMX (equivalent to 2 DS tabs TID) plus clindamycin 900mg q8h, followed by an additional 5 weeks of monotherapy with oral TMP/SMX (6 tabs a day; I presume given 2 DS tabs TID).  Patients with persistent bacteremia or cardiac abscess also received a week of gentamicin plus rifampin on top of the IV TMP/SMX and clinda.  Prior to this they’d been giving 6 weeks oxacillin (for MSSA) or vancomycin (for MRSA) as per the standard approach, with adjunctive gentamicin for 1 week in both cases.  For this paper, the authors compared patients with definitive S.aureus endocarditis by modified Duke’s criteria who were treated before (2001-2011) and after (2012-2016) institution of the new protocol; their primary endpoints were 30-day, 90-day, and global mortality.

A total 341 patients were included in the study (170 before, 171 after); 48 of these patients used injection drugs, and 40 had infections due to MRSA.  Everyone given TMP/SMX+clinda had isolates susceptible to both drugs. Potentially important differences in the group who received TMP/SMX+clinda included older age (64 vs 59 years) a higher incidence of septic shock at presentation (14% vs 7%) more spondylodiscitis (11% vs 5%), fewer mycotic aneurysms (2% vs 7%), fewer vegetations seen on echocardiogram (64% vs 82%), and fever surgeries (52% vs 67%).  In other regards the groups were similar. 

After a median 166 days of followup, global mortality was substantially lower in the TMP/SMX-clinda group (19% vs 30%, OR 0.56 with 95% CI 0.32-0.92), as was in-hospital mortality (10% vs 18%, OR 0.49 with 95% CI 0.26-0.93) and 30-day mortality (7% vs 14%; OR 0.46 with 95% 0.22-0.96); only 90-day mortality did not show a substantial difference between groups (16% vs 21%; p=0.32).  Length of hospital stay for folks who didn’t die in the hospital was much shorter in the TMP/SMX+clinda group (5 vs 30 days, p=0.005).  Infection relapse occurred in 13 of the TMP/SMX+clinda vs 22 of the standard therapy patients (13% vs 8%; p=0.11). Nine patients (5%) developed acute renal failure in the TMP/SMX+clinda arm, which honestly is less than I expected with 6 DS tabs a day; importantly, this was no worse a rate than observed in the patients given oxacillin or vancomycin.

This wasn’t an RTC, and table 1 suggests that the patients treated in the TMP/SMX plus clinda era may have had less severe infections overall.  That said, it’s a large study that aligns with the findings of a well-designed RTC, showing significant reductions in LOS with improved or at least similar clinical outcomes using early stepdown to orals versus standard 6wk IV therapy, and it did so in a cohort including both injection users and patients with MRSA (though, admittedly, subgroup analyses of these were not provided).  So, here’s several years’ worth of real-world data showing us oral antibiotics can adequately treat endocarditis. 31181351

We’ll end with some quick hits:

Vancomycin and Ceftaroline work equally well for acute bacterial skin and soft tissue infections This conclusion is based on a retrospective cohort study of 724 patients; the authors found that patients who received ceftaroline were more likely to meet their criteria for discharge by day three, but actual differences in infection-related length in stay were similar. Anyway, I sure would love it if we could all stop pretending that ceftaroline is a drug used primarily for skin and soft tissue infections, and not failure or intolerance of vancomycin (or dual therapy with dapto) in patients with bacteremia and other serious infections due to MRSA. 31405859

Omadacycline induces tachycardia but doesn’t prolong the QTc. The authors found a 17 bpm increase in heart rate occurred half an hour after administration of an 100mg IV dose of omadacycline.  I have never heard of antibiotic-induced tachycardia - I wonder if this is true of other tetracyclines? 31405867

Giving Australian children with cellulitis ceftriaxone via OPAT was more effective and 1/3 as costly as treating them with inpatient IV flucloxacillin. Someone should tell the Aussies that cellulitis can be managed with oral antibiotics. 31420292

Bedaquiline can be dosed at 100mg daily after a 200mg daily 8-week wash-in with drug exposures similar to the on-label 3x weekly regimen, can select for resistance to clofazimine, and seems to be active against a wide variety of NTMs including M.abscessus.  Interesting facts about a drug I’ve never used. 31451504, 31262765, and 31182533