March 2019: General ID

Elevated amoxicillin MICs portend a poor outcome in streptococcal infective endocarditis.  The authors performed a retrospective cohort study of streptococcal endocarditis (defined by Duke criteria) occurring in adults at 12 French hospitals over a 5-year period.  Microbiologic diagnoses were made by blood culture with species identification by MALDI-TOF and susceptibility testing by either disc diffusion or E-test using the EUCAST cutoffs.  Their aim was to examine patients’ clinical outcomes stratified by the amoxicillin MICs of the patients’ pathogens.

A total 414 patients were included in the study; 73% were men, the median age was 73, and 71% had native valve (mostly aortic valve) disease.  Most streptococci were in the viridans group (53%), of which the largest subset were the S.mitis/S.oralis group; of the remainder, the S. bovis/S.equinus group was the next most common.  Sixty percent of isolates had an MIC <0.125 mg/L, 13% had no susceptibility data reported, and the remaining 28% had MICs of 0.25 mg/L or greater.

Patients received combination antibiotic therapy in 71% of patients, with all but 3% of those being the addition of gentamicin.  The most commonly prescribed beta-lactams were amoxicillin (84%), and third-generation cephalosporins (12%).  When the authors compared survivors to non-survivors, they found that the nonsurvivors were more likely to have infections due to organisms with elevated amoxicillin MICs (57% vs 35%; p<0.01 in univariate analysis and p=0.04 in multivariate analysis); the only other strong predictor of death was not receiving surgery (73% of nonsurvivors vs 60% of survivors; p=0.04 in univariate analysis and p=0.0001 in multivariate analysis).  These data suggest that penicillin susceptibility cannot be assumed in streptococcal endocarditis, and that patients with pathogens that have higher amoxicillin MICs may need PK/PD optimized drug dosing and close clinical monitoring. 30851401

Prosthetic joint infections (PJI) with resistant gram-negative bacteria have poor outcomes and are best manage with joint replacement.  On the one hand, it’s insane to try managing PJI with debridement and implant retention (DAIR) when you don’t have any effective antibiotics to treat your patient’s infection, because DAIR asks antibiotics to do more of the therapeutic heavy lifting than they ought to in the first place.  On the other hand, in some cases staged implant removal and replacement may not be possible on account of the patient’s frailty or surgical factors.  Here, the authors analyzed data prospectively collected data from adults treated for multidrug-resistant (MDR) or extensively-drug-resistant (XDR) gram-negative PJI across 18 international medical centers over a 5-year period.  The authors defined “MDR” as resistance spanning at least three different drug classes (e.g. to ciprofloxacin, cefepime, and meropenem) and “XDR” as resistance to all but one or two drug classes.  Each patient was followed for 24 months after their initial surgery, and the primary outcome of interest was clinical success stratified by the surgical approach used and other treatment factors.

The analysis included 131 patients with a mean age of 73, two thirds of whom were women and half of whom had diabetes.  Nearly a third of patients had PJI with a sinus tract.  The most common infecting organisms were E.coli, P.aeruginosa, and K.pnuemoniae, and the most common resistance phenotypes were ESBL (72%), quinolone resistance (48%), and carbapenem resistance (9%).  Overall treatment success rates were low (58% for hips and 61% for knees).  The patients received extended durations of antibiotic therapy (mean 74 days), though antibiotic duration was not associated with outcome.  MDR infections generally responded better than XDR infections (treatment success rates 67% vs 39%; p=0.018), though no specific antimicrobial regimen was associated with a higher rate of cure than others.  Half of patients received DAIR, and this surgical approach was associate with increased treatment failure (OR 3.57; p<0.0001).  In subgroup analysis, non-DAIR approaches were superior to DAIR regardless antibiotic regimen, MDR vs XDR infection, and early versus late infection. 

So, the take-home message is that key to cure in resistant GNR PJI is surgery, and specifically removal of the infected prosthetic material. 30395988

A month of rifapentine plus isoniazid prevents tuberculosis in patients with HIV.  Note that I didn’t write “treats latent TB (LTBI),” which is on account of the quirky design of this (nonetheless important) clinical trial.  The authors randomized patients with HIV who either lived in an area of high TB prevalence or who had evidence of LTBI to receive either 9 months of isoniazid or one month of daily weight-dosed rifapentine plus isoniazid.  Pregnant and breast-feeding women were excluded, and all patients received NNRTI-based ART for at least the first month of treatment.  The primary outcome of interest was time to diagnosis of tuberculosis or death from an unknown cause, and the trial used a noninferiority design with a difference threshold of 1.25 events per 100 person-years.

Three thousand patients were enrolled and followed for a median 3.3 years; just over half were women, the median CD4 was 470, and half of patients were on ART.  The primary outcome occurred in 32/1488 patients in the 1 month rifapentine-INH arm and 33/1488 patients in the 9 months INH arm (0.65 vs 0.67 events per 100 person-years; rate difference -0.02 with 95% CI reaching 0.3 events per 100 person-years, meeting the prespecified criteria for non-inferiority).  Serious adverse events were similar between the two groups (6% vs 7% with p=0.07) but the patients assigned to the 1-month treatment group were much more likely to actually complete therapy (97% vs 90%; p<0.001).

These data suggest that a month of rifapentine and INH is probably as good as 9 months of INH for the treatment of LTBI, but I read this study with slight caution because of the inclusion of patients who merely lived in TB-endemic areas.  This choice potentially diluted out a true difference in the efficacy of the two regimens by including a sample of people who were not actually at high risk of developing TB disease (e.g. those living in an endemic area who nonetheless did not have LTBI at the start of treatment).  A counterargument to this is that difference in the rates of the primary outcome was not only not statistically significant, but barely present at all.  This suggests that either there truly isn’t a major difference in efficacy between the two regimens, or any difference between the regimens is made up for by the improved adherence in the rifapentine-INH arm. 30865794

Speaking of TB, NEJM published another trial this month focused on shortened treatment regimens for patients with rifampin-resistant TB.  Remember that rifampin’s unique mechanism of action makes it effective against metabolically quiescent cells – this gives it both the ability to sterilize infections involving biofilms (e.g. PJI and other hardware infections) and a particular potency against TB, which has dormant “persister” cells that can survive short durations of treatment.  Hence, when rifampin isn’t an option for TB, the traditional duration of treatment goes from a few months to a few years.  In this study, the authors used high-dose moxifloxacin in place of rifampin and looked to see whether a “short” regimen (11 months) would be as effective as a standard regimen (20 months) for treating TB.  Here, treatment success meant persistently negative TB cultures at 132 weeks after treatment with no other indicators of a poor clinical outcome.

A total 383/424 patients were included in the mITT analysis.  Long-term treatment achieved a favorable outcome in 80% of recipients, versus 79% for those in the short group (95% CI -7.5% to 9.5%, falling within the pre-specified 10% noninferiority margin).  There was no benefit with regards to the incidence of adverse events (grade 3 or higher, 48% for short duration vs 45% for long duration therapy); rates of death and emergent resistance to quinolones and/or aminoglycosides were also similar.  So, as long as your patient can tolerate a long duration of high-dose quinolones, you can safely shave off 9 months of antitubercular treatment without compromising treatment efficacy.  Not bad.  30865791