July 2019: General ID

Patients given more than 5-7 days of antibiotics for their pneumonia (be it CAP or HAP) suffer a higher rate of adverse events.  The most recent IDSA guidelines suggest 7 days of antibiotic therapy for HAP/VAP, and several studies show CAP can be treated effectively with 5 days of antibiotics, so outside of a few unique use cases (e.g. non-lactose fermenting GNR pneumonia in an immunocompromised host) that ought to be managed by ID anyway, there’s little reason to give more than 5-7 days of antibiotics for pneumonia.  But what are the harms?  This group out of Michigan sought to quantify just that.

The researchers used a retrospective cohort design, including adult patients at 43 Michigan medical centers over a one-year period discharged with a diagnosis of either community-onset pneumonia (CAP) or the now-defunct healthcare-associated pneumonia (HCAP).  Patients had to have had documented clinical and radiographic evidence of pneumonia, and to have received at least 4 days of antibiotics; the researchers excluded patients who had been admitted ot the ICU, were pregnant, were immunocompromised, had other concurrent infections, had unusual pathogens isolated (e.g. Legionella, fungi), or had conditions requiring prolonged treatment (e.g. empyema, bacteremia).  The primary outcome of interest was rate of excess antibiotic treatment; secondary outcomes included death, ED visit, and antibiotic-associated adverse events, all measured at 30 days after discharge.

A total 6481 patients were included in the analysis (median age 70; 51% women; 57% with a severe / PSI class 4 or 5 pneumonia; 34% with either a concurrent COPD or CHF exacerbation).  A bacterial etiology was established in just 8% of cases (mostly pneumococcus).  Two thirds of patients received excess antibiotics (median 2 days), for a total 2526 excess antibiotic days per 1000 patients with pneumonia; the per-hospital rates of antibiotic overuse ranged from 38% to 95%.  In multivariable analysis, factors associated with excess antibiotic prescription included having a positive respiratory culture, longer length of stay, no documentation of total antibiotic duration in the discharge summary, receipt of antibiotics in the preceding 90 days, and diagnosis of HCAP vs CAP.  With regards to patient outcomes, excess antibiotic use was not associated with 30-day mortality, readmission, or ED visits (which suggests that extending the course of treatment “just to be on the safe side” doesn’t impact meaningful outcomes), but excess prescription did increase patient-reported adverse events at a rate of 5% per excess antibiotic day.

So, if you’re looking to give your patient diarrhea or vulvovaginal candidiasis, by all means continue to prescribe extended courses of antibiotics for pneumonia; otherwise, stick to a total 5-7 days. 31284301

Unnecessary anaerobic coverage promotes colonization with ceftriaxone-resistant organisms.  The authors retrospectively examined all patients with a >72hr stay in their French surgical ICU, comparing those who received noncarbapenem antibiotics with anaerobic coverage (i.e. pip-taz, amox-clav, and metronidazole) to those who did not.  The patients received rectal screening for ceftriaxone-resistant GNRs (Enterobacteriaceae with AmpC or ESBL, Psuedomonas, Stenotrophomonas, and Acinetobacter) on admission and then weekly; the primary outcome of interest was acquisition of such organisms during the admission.

A total 352 patients were included in the study, of whom 10% had one or more type of ceftriaxone-resistant GNR on admission; the patients’ median ICU stay was 16 days, and 12% died in the ICU.  A third of the patients acquired a new ceftriaxone-resistant GNR during their stay, mainly AmpC-producing Enterobacteriaceae, and patients who had intestinal colonizations with these pathogens were 10-25% more likely to develop infections with them.  To adjust for clinically important differences between patients given vs not given anti-anaerobic antibiotics, the impact of anti-anaerobic antibiotics was assessed in a propensity score matched cohort (n=77 for each arm).  The authors found that receipt of antibiotics with anaerobic activity was associated with ceftriaxone-resistant GNR acquisition (aHR 3.9; 95% CI 1.1-13.7). In comparison, neither receipt of ceftriaxone nor receipt of carbapenems were associated with carriage acquisition.  The authors found similar results when assessing the entire cohort using propensity score weighted multivariate regression analysis.

When I read a paper like this I think back to the undergraduate course I took in microbial ecology.  What is the field of infectious diseases, after all, except applied ecology and evolution?  Here, our ecosystem is the gastrointestinal tract, inhabited by normal flora - including anaerobes - that prevent immigration and colonization by nosocomial pathogens by taking up physical space on the intestinal surface, competing for nutrients, stimulating the immune response (aka, shifting the environment toward a state more hospitable to normal flora and less hospitable to invaders), and probably serving other important biochemical/metabolic functions we haven’t characterized that support biodiversity.  Give broad-spectrum antibiotics and you’re essentially carpet-bombing the Amazon rainforest – destroying a complex and resilient ecosystem, and in doing so vacating ample ecologic niches into which invasive (and pathogenic) species can take up residence.  This is the same reason that it’s no surprise FMT seems able to clear intestinal carriage of MDR pathogens – in a sufficiently developed and robust ecosystem, every ecologic niche is occupied by species highly adapted to thrive in that specific habitat and role, and “ecologic opportunist” pathogens find themselves crowded out.  Anyway, that’s a longwinded way of saying don’t prescribe antibiotics that kill microbes you don’t specifically want dead – you’ll just make life harder for you and your patient later on.  31289826 

More frequent preventative treatment of malaria in Ugandan children yielded unexpected post-treatment protection in a recent RTC.  Intermittent preventative treatment (IPT) with dihydroartemisinin-piperaquine (DHT) consists of one 20mg/160mg tablet given daily for three days on a recurring basis and is a novel approach preventing malaria in African children. This double blind RTC recruited pregnant women and followed their children from birth through year 3, with children given the drugs from 8 weeks to 24 months of age.  The women and their children were randomized into four groups, all of whom received DHT on various schedules: pregnant women every 8 weeks and then kids every 4 weeks, every 4 weeks for both, pregnant women every 8 weeks and kids every 12 weeks, and pregnant women every 4 weeks and kids every 12 weeks.

A total 191 children were born to the enrolled mothers, of whom 183 reached 8 weeks of age and were included in the primary analysis.  The authors observed 96% fewer episodes of symptomatic malaria among the children given IPT every 4 versus every 12 weeks (aIRR 0.041; 95% CI 0.012-0.15), as well as an 89% reduction in the prevalence of parasitemia.  Moreover, after IPT was discontinued at 24 months, the children who had received monthly treatment continued to have a lower incidence of malaria than those who had been treated every 12 weeks (aIRR 0.62; 95% CI 0.4-0.95).  Adverse event rates were similar between groups, most often consisting of cough.

The findings of this study are interesting because one of the original objections to IPT was that it might hamper development of natural immunity by decreasing exposure to malaria.  In contrast, it seems that IPT – or at least, monthly IPT with DHT – promotes the development of a sustained, partially protective malaria immunity. 31307883

A small cohort study suggests that treating streptococcal PJI for six months or more reduces the rate of treatment failure.  This prospective study was performed at a large tertiary care center in Germany that has a multidisciplinary orthopedic infection unit caring for approximately 300 patients a year.  In 2016, the center began using >6 month durations of antibiotic suppression for PJI.  This study compared the outcomes of patients with streptococcal PJI treated before and after the change (i.e. with a prospective cohort enrolled 2016-2018 and compared to a retrospective cohort of patients seen from 2009-2015). Patients with polymicrobial infections were excluded.  The center’s surgical approach is I&D with mobile component exchange and implant retention for infections <4 weeks after arthroplasty duration and <4wks in duration (“acute infection”), and explantation with a 1 or 2-stage reimplantation for all others.  Antimicrobial suppression involved amoxicillin (preferred) or doxycycline (alternate). The primary outcome, treatment success, was defined wound healing with no clinical evidence of PJI, no further need for surgery, and no PJI-related death; followup to assess this outcome occurred at months 3, 6, and 12.

A total 69 patients were included in the study, of whom 24 received the long-term antimicrobials.  The infections largely involve the knee and hip and did so with similar frequencies; in both groups, more than half of patients had had prior revision surgeries before their present infections. Patients who received long-term antibiotics were more likely to have been treated with implant retention (63% vs 27%; p=0.005), received an additional week of IV antibiotics on average (mean 21 vs 14 days; p=0.04), and were half again as likely to be treated with monotherapy (75% vs 49%; p=0.06).  Thirty eight (58% of) patients remained infection-free over a median 13 months. In multivariate analysis adjusting for surgical strategy, duration of IV antibiotics, addition of rifampin, previous revision surgery, and type of prosthesis, use of >6 months of antibiotic treatment was associated with a lower rate of treatment failure (OR 0.07; 95% CI 0.01-0.72). In a Kaplan-Meier survival analysis of extended followup, by month 36 the rate of treatment failure-free survival was 95% in the suppression arm versus 54% in the comparison group (p=0.02 despite this 40% difference, which belies the small sample size).  Four of the 24 patients given amoxicillin suppression (e.g. 1 in 6 patients) developed a rash; two of these were switched to clindamycin, and they both subsequently developed C.difficile infection.

So, antimicrobial suppression (or rather, extended oral treatment) for 6 months or more may improve the clinical outcome of streptococcal prosthetic joint infection.  I have two concerns about interpreting data from this study.  First, the sample size is small – less than 50 patients in each arm.  Second, the median duration of followup was not that long – meaning in many cases the patients receiving antibiotic suppression would have been on or just recently stopped antibiotic therapy at their last captured followup visit.  And maybe that’s fine if your takeaway from this paper is “OK, I’ll just give my patients with streptococcal PJI a 50-gallon bucket of amoxicillin and have them stay on it for life” – a reasonable approach for those patients with recalcitrant and relapsing GBS bone and joint infections.  But if, on the other hand, if your takeaway is “I should give all of my streptotoccal PJI patients six months of amoxicillin because it will improve their likelihood of durable clinical cure once they come off antibiotics,” I’m unsure whether data adequately support that conclusion. 31310778

What is the epidemiology of CSF leakage in patients with bacterial meningitis?  The authors examined cases of community-acquired bacterial meningitis in a prospective national cohort study in order to define the epidemiology of CSF leak-associated meningitis. Sixty-five of 2022 meningitis episodes (3%) were associated with a CSF leak, and 59% of these represented recurrent episodes. The cause of CSF leak was identified in 75% of cases, and was most often due to ENT surgery (29%) or a remote history of head trauma (23%); in 61%, the CSF leak was evident on admission, usually on account of liquorrhea.  Only a third of these patients had the triad of fever, neck stiffness, and altered mentation on admission; this may be accounted for by their shorter time to presentation, which was <24 hours in two thirds of cases. Among the recurrent episodes, the CSF leak had already been identified in most cases (71%), and half of the patients had already had a surgery to try to address the leak. CT and/or MRI was able to identify the anatomic source of the CSF leak in 92% of cases.

The most common pathogen was pneumococcus, isolated in 51% of cases, followed by unencapsulated H.influenzae (17%) and other streptococci (16%).  An LP was performed in each case; the median opening pressure was 42 cm H20 and the median WBC count was 4870 (>1000 in 89% of cases).  Patients were given beta-lactam therapy (e.g. ceftriaxone and amoxicillin) in most cases; 88% had a favorable outcome (vs 63% in patients with meningitis in the absence of a CSF leak), and no patients died.  So – keep in mind that meningitis in the presence of a CSF leak has a relatively high rate of recurrence, even in cases that have been previously referred for surgical repair, and that they more often present without the classic triad of clinical symptoms, perhaps because they seek care more quickly after the onset of symptoms.  31300817