Antimicrobial Stewardship and Infection Control: September 2018

When it comes to encouraging colonization with MDR pathogens, not all antibiotics are created equal. I like this study because it clearly demonstrates the harm (in terms of antibiotic resistance acquisition) of choosing antibiotics unwisely. Stewardson et al performed a multinational prospective cohort study involving primary care clinics around Europe between 2011 and 2013. They recruited outpatients with suspected UTI (i.e. those exposed to antibiotics), outpatients seen for other problems who were not exposed to antibiotics, and 1-3 household contacts for each outpatient. Fecal samples were obtained from the outpatients and their household contacts at the baseline visit, at the end of the antibiotic prescriptions, and 4 weeks later. Samples were incubated on selective media supplemented with ciprofloxacin or nitrofurantoin to identify carriage of resistant bacteria, and E.coli isolates were tested for the ESBL phenotype.

The study recruited 716 participants from 300 households, including 205 outpatients given antibiotics (42% received nitrofurantoin and 37% received quinolones). Increased household prevalence of ciprofloxacin-resistant bacteria in stool at 4 weeks was associated with use of quinolones (aPr 2.0; 95% CI 1.18-3.36; Number Needed to Harm = 12) but not nitrofurantoin (aPr 0.98; 95% CI 0.53-1.81), and shorter courses of quinolones were no less “resistance-ogenic” than longer durations. In contrast, use of nitrofurantoin was not associated with an increased prevalence of nitrofurantoin-resistant bacteria in stool, and in fact 11/12 of the nitrofurantoin resistance cases came from antibiotic-unexposed households. Neither nitrofurantoin nor ciprofloxacin use was associated with ESBL colonization, though 12% of ciprofloxacin resistant E. coli isolates in the study had the ESBL phenotype, versus 0% of the nitrofurantoin-resistant E. coli isolates. 29331548

It’s unfortunate, then, that Nostrum Laboratories recently chose to increase the cost of liquid nitrofurantoin, a 60-year-old product, to more than $2300 a bottle. Nostrum’s CEO has responded to criticism of the move by saying that the company has a “moral requirement to sell the product at the highest price.” We should have an international standard unit for moral vacuity in pharmaceutical executives; I suggest the Shkreli.

Continuous low-dose antibiotic prophylaxis reduces the incidence of recurrent UTIs in patients who undergo self-catheterization at the cost of a significant increase in carriage of antimicrobial-resistant bacteria. This trial out of the UK randomized patients who use clean intermittent urinary self-catheterization 1:1 to receive either daily antibiotic prophylaxis or placebo for one year. The primary outcome of interest was the incidence of symptomatic UTI requiring antibiotic treatment over the study period, and the antimicrobial susceptibility profile of urinary and fecal isolates was monitored as a secondary outcome.

Of the 318 patients included in the primary analysis (n=159 for each group), the incidence of UTI was 1.3 cases per person-year in the prophylaxis group and 2.6 cases per person-year in the placebo group. One way of describing this data is to say that antibiotic prophylaxis reduced the incidence of symptomatic UTI by 48%. Another way is to say that it took an average of 280 days of antibiotic prophylaxis to prevent one symptomatic UTI. Use of prophylaxis was well-tolerated; however, antimicrobial resistance to the agent used for prophylaxis increased in the prophylaxis recipients by months 9-12 (absolute increases in resistance prevalence in the prophylaxis versus placebo groups: nitrofurantoin 15%, trimethoprim 34%, and co-trimoxazole 29%; p < 0.05 for all differences).

I don’t know about you, but to me giving nine months of antibiotic prophylaxis to prevent one 1-7 day course of antibiotics, with the added cost of increasing the patient’s carriage of antimicrobial-resistant pathogens, seems like a bad deal.  30037647

Clinical assessment reliably identifies patients with a positive stool C. difficile PCR and a negative toxin A/B assay for whom CDI treatment can be safely forgone. So says this small Canadian retrospective cohort study. The authors reviewed the cases of all patients with diarrhea and indeterminate results of a C. difficile stool testing algorithm (i.e. a positive PCR and GDH assay but a negative toxin immunoassay) seen at two Vancouver hospitals over a one-year period. These patients were evaluated by an antimicrobial steward (MD/DO or PharmD) and classified as colonized or infected with C. difficile based on the presence of >3 loose stools per 24hr without an alternate explanation (e.g. laxative use); this assessment was communicated to the primary physician, and the patient was reclassified as infected if the primary physician decided to treat anyway. The outcomes of interest were need for subsequent CDI treatment, follow-up CDI testing, all-cause mortality, and development of CDI complications in the 8 weeks after the initial indeterminate test result.

A total 110 patients with indeterminate test results met the study’s inclusion criteria (patients classified as colonized, n = 51; classified as infected, n=59). Patients with an immunocompromising condition were more likely to be classified as infected rather than colonized; otherwise the two groups were similar. Comparing the patients classified as colonized - and hence untreated at the time - versus infected, there were no statistically significant differences in the rates of subsequent treatment for CDI (2% vs 11%), followup CDI testing (33% vs 39%), all-cause mortality (11.8 vs 8.5), or CDI complications including ICU care, megacolon, and colectomy (p > 0.05 for all). 30238342

An EMR nudge to obtain an ID consult prior to PICC insertion for IV antibiotics reduces the incidence of inappropriate PICC placement. In a letter to the editor published in ICHE, Liu et at from Mount Sinai report their experience with adding a prompt to their hospital’s electronic medical record to improve antimicrobial stewardship. Specifically, they added a prompt to the PICC insertion order set with the question “Will the PICC be used for antibiotics?”, and if the answer was yes, “Is ID consulted?”, with the clinician being required to select “ID will be consulted immediately” to continue in the order set if the answer was “no.” The group had two ID physicians independently review the cases in which a PICC was ordered without ID consultation to determine whether the PICC placement for IV antibiotics in those cases was appropriate.

In a six month period prior to the intervention, nine out of 181 PICCs were placed without an ID consultation, and these were inappropriate in 8/9 cases. In the six months after the intervention, the prompt provoked three additional ID consultations, and in total only four of 234 PICC orders were placed without ID consultation, with only one resulting in an inappropriate PICC placement. I must confess, I despise pop-up messages in an EMR, but these at least seem to be doing some good. 30231950