Antimicrobial Agents: November 2018

Substituting linezolid for ethambutol in the 4-drug phase of therapy for pulmonary tuberculosis does not improve time to culture conversion. Linezolid has become one of the go-to drugs for the treatment of MDR tuberculosis, so with the drug having recently gone generic, why not try adding it in as first-line therapy?  The authors conducted an open-label RTC at three South Korean hospitals, recruiting adults with pulmonary tuberculosis not resistant to rifampin in their first week of treatment.  Subjects were randomized 1:1:1 to have 8 weeks of standard 4-drug antitubercular therapy, linezolid subbed for ethambutol for the first 2/8 weeks of 4-drug therapy, or linezolid subbed for ethambutol for the first 4/8 weeks of 4-drug therapy.  The primary outcome was the proportion of patients with a conversion to a negative sputum culture by week 8.

A total 429 patients were enrolled over a three-year period, of whom 401 were included in the primary analysis.  At week 8, conversion to a negative sputum cultures was observed in 77% of patients in the control group, 82% of patients in the 2-week linezolid group, and 76% of patients in the 4-week linezolid group; none of these differences were statistically significant. Adverse event rates were similar across groups, and no cases of new linezolid resistance were identified. So much for “RIPL” - though I suppose it’s nice to know linezolid is an effective alternative if my patient can’t tolerate ethambutol but I still want to give them optic neuritis. 30477961

Both cefepime and piperacilling-tazobactam effectively treat bacteremia due to organisms with an inducible AmpC beta-lactamase. AmpC beta-lactamases confer broad resistance to penicillins and cephalosporins while being resistant to beta-lactamase inhibitors. They are present in a wide variety of gram-negative bacteria, though in most they’re constitutively expressed at either low or high levels; when we talk about “AmpC organisms,” what we’re really talking about are organisms with the ampR regulatory element that demonstrate inducible resistance, meaning they can convert from a beta-lactam susceptible phenotype to a resistant phenotype after exposure to beta-lactams that induce AmpC gene expression. So if you have an organism with an inducible AmpC (specifically Enterobacter spp and Citrobacter freundi – there is less data to suggest this phenomenon is clinically relevant for other genera), your patient’s initial cultures may look pan-susceptible except for resistance to strong AmpC inducers like cefoxitin or cefazolin, yet beta-lactam treatment may lead to clinical failure or relapse with the newly-resistant version of the same isolate. For those who are interested, Dr. Jacoby published a definitive review of AmpC in 2009, which is freely accessible here.

Inducible AmpC organisms have been traditionally treated with a carbapenem or a non-beta-lactam agent. However, cefepime is stable against the AmpC beta-lactamase and piperacillin is a weak inducer of AmpC, so these are potential carbapenem-sparing options. The authors reviewed cases of bacteremia due to an inducible AmpC organism (which they defined as a species of Enterobacter, Citrobacter, or Serratia demonstrating isolated cefoxitin resistance) treated with one of these drugs at a single academic medical center over a 5-year period (n=132). The primary outcome of interest was clinical cure at the end of therapy (EOT); secondary endpoints included microbiological cure, rates of change in isolate susceptibility, and 7 and 30-day mortalities.

Most patients had catheter-related infections, and nearly 80% of infections were due to Enterobacter cloacae. A quarter of the patients were immunocompromised. The aggregate rate of clinical cure at EOT was 87%, and microbiologic cure was achieved in 93%; mortality was 4% at 7 days and 11% at 30 days. Rates of microbiologic cure were similar between cefepime and piperacillin-tazobactam (93% and 96%). Resistance emerged during cefepime treatment in 4/108 cases and during pip-taz treatment in 1/24 cases; of note, these were all severe infections with inadequate source control.

In subgroup analysis, the only predictor of clinical failure was use of pip-tazo for isolates exhibiting baseline resistance to third-generation cephalosporins rather than to cefoxitin alone (0% vs 91% clinical success, p<0.001). This makes sense to me: piperacillin isn’t stable to the AmpC beta-lactamase, which also isn’t inhibited by tazobactam – so if the AmpC gene is already a little bit turned on at baseline, pip-tazo probably isn’t going to be effective. Your mileage may vary, but if I’m going to eschew a carbapenem for a serious infection with an inducible AmpC organism, I’ll pick cefepime over pip-taz any day. 30125680

How do patients with MRSA bacteremia fare when their antibiotic regimens are switched from IV to oral agents?  The IDSA’s 2011 guidelines on the treatment of MRSA infection only recommend intravenous antibiotics for bacteremia. At my institution, I’ve found that actual practice varies considerably between ID specialists; some are dogmatically adherent to IV-only treatment, while others are happy to switch to a highly bioavailable oral agent after the patient has clinically stabilized. Little published evidence supports the latter approach; this study aimed to change that.

The authors reviewed the outcomes of patients with MRSA bacteremia from a single center over a 10-year period who received either oral or IV outpatient antibiotic treatment. The primary outcome of interest was 90-day clinical failure, which was defined as recurrent MRSA bacteremia, development of a new deep-seated MRSA infection, or death. A total of 492 patients were included in the study (oral antibiotic treatment, n=70; IV antibiotic treatment, n=422), and inverse probability of treatment weighting was used to balance key prognostic factors for infection outcome (this is a form of propensity score weighting, at which some statisticians cry foul). In the weighting-adjusted analysis, 90-day clinical failure rates were non-significantly lower in the patients who received oral antibiotics (aHR 0.38; 95% CI 0.13-1.1); in addition, the patients who received oral antibiotics had lower rates of 90-day readmission (aHR 0.6; 95% CI 0.39-0.94). So, these data suggest that in at least some patients MRSA bacteremia can be safely treated with oral antibiotics. 30418557

Speaking of S. aureus, a retrospective study in CID this month reports the import of S. aureus bacteriuria (SABU). I was always taught that SABU indicated S. aureus bacteremia (SAB). In this study of 2540 urine cultures positive for S. aureus from 2054 patients, the total incidence of SAB was 7%. The patients who had SAB were more likely to be male, hospitalized, to have diabetes, cirrhosis, and/or cancer, to have recently had a urologic procedure, to have lab findings suggesting a systemic infection, and to have pure S. aureus in the urine culture. On the other hand, patients with SABU alone were more likely to be old, demented, and to have abnormal urinalyses. In-hospital mortality was substantial for both groups, occurring in 9% of the patients with SABU and 18% of the patients with SABU+SAB; mortality was highest among patients with SAB that went undetected until >48hr after detection of SABU. Between that and the not-so-small 7% incidence of bacteremia among patients with SABU, I’m gonna go ahead and keep ordering blood cultures whenever I see S. aureus in the urine. 30476003

High-dose sertraline has antifungal activity in animal models of Aspergillus fumigatus infection. Several studies suggest that the SSRI antidepressant sertraline possesses antifungal activity, the proposed mechanism for which is inhibition of eukaryotic translation initiation factor Tif3.  The strongest supporting clinical data is a small open-label Ugandan trial of adjunctive sertraline for cryptococcal meningitis, in which recipients had faster CSF clearance and lower incidences of IRIS and relapsed infection than a historical reference cohort (reference: 26971081). However, while the data isn’t yet published, I’ve been told that a subsequent head-to-head RTC of adjunctive sertraline versus placebo showed no benefit.  Anyway, if sertraline does have clinically meaningful antifungal activity, that’d be a big deal for the developing world and it’s often quite limited antifungal armamentarium.

In this paper, the authors used escalating doses of sertraline to treat Aspergillus fumigatus infections in two animal models of invasive aspergillosis: the waxy moth and the mouse. In the moth larvae, sertraline improved both survival and a standardized index of larval health, particular at the higher doses (10mg/kg and 15mg/kg). In the mice, 10mg/kg of sertraline was as effective as 3mg/kg amphotericin or 10mg/kg voriconazole at reducing pulmonary fungal burdens. Note that this would be the equivalent of 700-1400mg of sertraline a day for a human – well in excess of the 400mg given in the reportedly negative trial above. Who knows whether those dosages would even be tolerable in people. 30403787

This literature review concludes that a single dose of aminoglycoside is effective for urinary tract infections. Fewer and fewer antibiotics are reliable for uropathogens in the community. Nitrofurantoin and fosfomycin remain excellent choices for cystitis, but beyond these the outlook is grim; where I practice, outpatient E.coli resistances to trimethoprim-sulfa, amox-clav, first-generation cephalosporins, and fluoroquinolones are all in the 20-40% range. What about aminoglycosides?

The authors searched the English-language literature for studies involving the use of single-dose aminoglycosides for either upper or lower urinary tract infection. They identified 13 articles (total N = 13,804 patients, median N = 37), of which seven had a comparator arm and seven included children only. Most of the studies focused on cystitis, though two included patients with symptoms of pyelonephritis; none of the studies included cases with sepsis or bacteremia, and five excluded patients with fever. Netilmicin, amikacin, and gentamicin were the most commonly used drugs.

Microbiologic cure was reported in 11/13 studies and was >85% in every case; the overall microbiologic cure rate was 95%, with no difference in outcomes between children and adults. Among the studies with an adequate duration of followup, relapse or reinfection within 30 days occurred in 19% (84/433 patients) – however, this was driven primarily by patients with urinary tract abnormalities (43% vs 12% in those with normal GU anatomy). In the studies with a comparator arm, outcomes between the aminoglycoside and alternate agent were similar (95% vs 96% initial microbiologic cure; 71% vs 73% sustained microbiologic cure). Adverse events attributed to the aminoglycoside occurred in 0.5% of cases and were mostly vestibular (e.g. tinnitus or disequilibrium); in comparison, the adverse event rate in patients receiving comparator antibiotics was 3.5%.

So, while the evidence is drawn largely from reports of children with cystitis, a single dose of aminoglycoside appears as effective and at least as well tolerated as other agents for the treatment of urinary tract infection. I’ll probably never use this in the clinic setting, and will continue to reach preferentially for nitrofurantoin or fosfomycin for that odd patient with true cystitis due to a MDR/ESBL-producing organism. But maybe aminoglycosides have a limited role as a carbapenem-sparing option for nonsevere ESBL UTI with hints of upper tract disease? 30397061

Phage therapy efficiently eliminates epidemic Vibrio cholera from the GI tract and prevents clinical disease… in rabbits.  It was nice to see this article after reading about the Phagoburn trial being marred by technical problems last month. The authors used Phi_1, a phage sourced from a UK researcher, to treat V. choleae infection in infant rabbits. After demonstrating the phage’s breadth of activity against several strains of V. cholera, they infected neonatal rabbits with pathogenic V. cholerae O1 (classical biotype) in the presence of an antacid, with or without the addition of Phi_1 either prophylactically or therapeutically (i.e. phage given either 6 hours before or after bacterial inoculation). The primary outcome of interest was diarrhea in the 24 hours following inoculation, and the secondary outcome was the intestinal V. cholerae bacterial load. 

Both prophylactic and therapeutic administration of Phi_1 conferred complete protection from V. cholera disease in the rabbit model (0/19 and 0/22 animals developed disease); untreated rabbits developed symptomatic infection in 11/17 cases. Receipt of phage was associated with 2-4 log reductions in V. cholera throughout small intestine. Interestingly, when the V. cholera recovered from phage-treated rabbits was recultured it remained susceptible to Phi_1, and the authors’ subsequent in vitro attempts to generate phage-resistant V. cholerae were unsuccessful.

The two human clinical trials of phage therapy for cholera I could find were done in the 1970s and were disappointing (4946956 and 4988693 – historical sidebar: it seems that this Pakistani cholera research laboratory was getting technical support from the USSR Ministry of Health at the same time it was operating under grant funding from the NIH!). These papers reference three reportedly successful trials of phage performed by Soviet researchers in the 1960s, but said papers aren’t in the English language literature. Anyway, these two studies both used phage cocktails that hadn’t been well-characterized and may not have had adequate lytic activity or host range. In addition, neither study included an antacid prior to administering the phage, so the actual infectious dose of phage getting to the V. cholerae in the small intestine was probably pretty low. This is all to say that performing a new human trial with a preparation of Phi_1 seems like an eminently reasonable idea. 30395214