April 2019: Antimicrobial stewardship and infection control

Clindamycin-resistant Staphylococcus aureus is increasing in incidence among patients with surgical site infections.  At least, that’s the case at Johns Hopkins, where this retrospective study took place.  The authors examined 109 cultures of S. aureus taken from patients with surgical site infections after a number of various procedures between 2012 and 2017.  The authors used the Cochran-Armitage test to calculate the statistical significance of time trends in the susceptibility of isolates to each antibiotic.  They found that, from 2012 to 2017, clindamycin resistance increased (20% to 38%; p < 0.01), tetracycline resistance decreased (20% to 0%; p < 0.01), and trimethoprim-sulfa resistance remained stable (10% to 13%; p=0.63).  Thirteen of these patients had documented penicillin allergies; five of these were given clindamycin, even though only one had a documented severe reaction to penicillin, and two went on to develop clindamycin-resistant infections.

The authors write that “unnecessary deviations from recommended B-lactam containing regimens should be reduced as much as possible,” and they’ll get using no objection from me.  On the other hand, they state that vancomycin prophylaxis is not recommended on account of, among other things, the risk of producing vancomycin-nonsusceptible S. aureus and evidence that vancomycin is inferior to beta-lactams for surgical prophylaxis.  To the former point – one of the most (only?) remarkable attributes of vancomycin has been its resistance to resistance among the staphylococci, so if anything I see this as a strong argument in favoring of using vancomycin preferentially, protecting other antibiotics with lower barriers to resistance.  To the latter point – is vancomycin’s inferiority to beta-lactams for surgical prophylaxis a marker of vancomycin’s inferiority, or of beta-lactams’ superiority (that is – would TMP/SMX or doxycycline be any better than vancomycin)?  I’m not aware of data convincingly arguing one way or another.  If we’re certain that a patient cannot receive a beta-lactam for surgical prophylaxis – is there an antibiotic with proven superiority, in terms of either safety or efficacy, to vancomycin? 30782221

Prospective audit and feedback can improve ID consultation rates and treatment in S. aureus bacteremia.  The authors implemented a prospective audit and feedback system aimed at improving adherence to a S. aureus bacteremia (SAB) treatment bundle at their single inpatient center.  The bundle included ID consultation, following blood cultures until cleared, echocardiography, source control procedures when feasible, and appropriate selection and duration of antibiotic therapy.  The study had a pre-post intervention design; the primary outcome was adherence to the bundle, and secondary outcomes included LOS, 30-day readmission rates, and 30-day mortality.

A total 199 patients were included (preinterventipm, n=62; postintervention; n=137).   Antimicrobial stewardship pharmacists made 148 recommendations in the post-intervention period, of which 93% were accepted by the treating team.  Overall adherence to bundle recommendations increased by 44%, driven by increases in ID consultation (57% to 83%), appropriateness of definitive antimicrobial therapy (84% to 99%), rates of echocardiography (73% to 96%) and adequate durations of therapy (87% to 100%) (p<0.001 for all comparisons).  In multivariate regression analysis, the prospective audit and feedback intervention was independently associated with adherence to bundle recommendations (aOR 5.6; 95% CI 2.8-11.3; p<0.001).  Fewer patients were readmitted (9% vs 18%; p=0.07) or died (22% vs 29%; p=0.28) within 30 days in the post-intervention period.  Given the general agreement of the previous literature on expert consultation in SAB and improvement in clinical outcomes, it seems reasonable to expect that a larger study would have found statistical significance in those differences in outcomes.  On the other hand, this approach may be more cumbersome and costly than simply having an automatic and mandatory ID consultation triggered any time a blood cultures turns positive with S. aureus.  30949538

Is it time to curb our enthusiasm about fecal microbiota transplantation (FMT) for recurrent C. difficile?  This is a meta-analysis of thirteen trials of FMT for C. difficile infection (total n=610), which the authors performed after noting that cure rates in observational studies, often >90%, seemed higher than those achieved in clinical trials.  Sample sizes of the individual trials were small (n=20 to 127), diagnostic criteria were variable (PCR-only in two studies, EIA-only in three studies, either/or in three studies, and not stated in the rest), and the durations of follow-up were less than 90 days in all but three studies.

The authors found that 439/610 (76%) of patients achieved a clinical cure overall, and that these rates were lower in randomized vs open-label trials (139/216 or 68% vs 300/394 or 83%), in patients who received FMT via enema vs colonoscopy or oral therapy (66% vs 87% and 81%, respectively), and in studies that included refractory CDI vs only recurrent CDI (64% vs 79%).  There were no differences in outcome with FMT using fresh vs frozen stool (80% vs 77%), nor any differences based on whether EIA alone or other strategies were used for diagnosis (79% vs 78%).

Any time we see better efficacy with a treatment in observational studies versus clinical trials, or in open-label versus randomized clinical trials, we should be suspicious of researcher bias producing overly exuberant conclusions (and here I don’t mean any sort of malicious intent or research malfeasance, but the sort of bias that creeps in when you have a researcher who genuinely believes in the efficacy of their new procedure categorizing data points by subject clinical outcomes).  On the other hand, an efficacy rate of two thirds for recurrent and refractory CDI is still pretty good, and would mean FMT is probably still the best option available for these recalcitrant infections. 30957161