September 2019: General ID

Clofazamine facilitates shorter durations of treatment for MDR tuberculosis.  So says this RTC out of China that randomized 135 patients to receive either standard of care (6 months of 6-drug therapy followed by 5-drug consolidation therapy for a year or more) or a shortened clofazamine-based regimen (6 months of 6-drug therapy with clofazamine replacing ethambutol in the SoC regimen followed by 5-drug therapy with clofazamine for another 6 months).  The primary outcome was treatment outcome based on standard WHO definitions for MDR-TB (basically, >3 consecutive negative sputum cultures for “cure”, and bacteriologic conversion with <3 consecutive negative sputum cultures by the end of therapy for “treatment completion”), and the study used a superiority design. 

The median age was 39 years, 79% had previously been treated for a median 15 months, and none of the patients had HIV.  In total, 69% of patients in the clofazamine arm had a favorable outcome (cure or treatment completion as defined above) versus 65% in the standard of care arm, with p=0.7.  Rates of treatment failure and mortality were also similar between groups.  Of note, patients in the clofazamine arm of the study more often had sputum conversion by three months (69% vs 56%; p=0.04) and were more likely to have resolved their cavitary disease by the end of treatment (38% vs 24%; p=0.06).  Adverse event rates and severities were similar.  So, in conclusion, 12mo of a clofazamine-based regimen appears to be as good or better than 18mo of standard of care therapy for MDR TB, with the caveat that this study was powered for superiority and not non-inferiority or drug safety outcomes. 31549147

Speaking of tuberculosis, this RTC out of Korea randomized adults with pulmonary TB to either standard 4-drug antitubercular therapy with discontinuation of ethambutol after the results of culture-based phenotypic susceptibility testing, or early discontinuation of ethambutol as soon as a genotypic assay proved the susceptibility of the patient’s isolate to both isoniazid and rifampin.  The primary outcome was treatment success at the end of therapy, and the study was powered to show non-inferiority.  The mean duration of ethambutol was 2 weeks in the genotype-based de-escalation arm versus 11 weeks in the culture-based de-escalation arm.  Of the 600 patients randomized, treatment outcome could be evaluated in 493; of these, the treatment success rates were 94% in both study arms (RR 1.0 with 95% CI 0.95-1.06), meeting the prespecified threshold for noninferiority.  One year recurrence rates and adverse event profiles were also similar.  So, TB genotyping for rifampin and INH susceptibility can drastically shorten the necessary duration of ethambutol therapy in pulmonary TB.  Good news for people who want to be able to distinguish Christmas decorations everywhere! 31527020

This meta-analysis of 9 retrospective studies and a single RTC suggests that after PJI managed with debridement and implant retention (DAIR), shorter durations of antibiotic therapy (8 weeks for hips, 12 weeks for knees) may be as effective as longer courses. The data cited in the meta-analysis comprised 856 patients (just 63 of whom were in the single RTC) and the durations compared were quite variable, particularly because several studies focused on the duration of IV antibiotic therapy versus duration of total antibiotic therapy.  That said, when the authors applied a common definition of failure (death, re-infection, or persistent infection), they observed no difference between the short and long-course antibiotic recipients (RR 0.9 with 95% CI 0.6-1.2).  I think I’m going to wait until I’ve gotten all of my orthopedic surgery colleagues onboard with early switch to PO as per OVIVA before springing this on them, but I’ll be interested to see what new evidence comes out and if the recommended durations for PJI with DAIR change with the next update to the guidelines.  31549147

Speaking of PJI, this retrospective cohort study examined the outcomes of patient with PJI managed with DAIR who had their infections within 90 days of the initial arthroplasty surgery.  The question the authors sought to answer was whether DAIR is any less successful when performed more than 30 days after surgery (the traditional cutoff for expecting a good outcome – the idea here is that if you can catch an infection introduced at surgery before it can set up a mature biofilm, you have a decent chance of being able to eradicate it).  The authors excluded patients with bacteremia, arthroscopic rather than open debridements, and those who had less than a year of follow-up; they defined failure as the patient needing another surgery for PJI, dying, or being placed on long-term suppressive antibiotics.

A total 769 patients were included in the cohort, of whom 294 (38%) had a treatment failure.  The rates of treatment failure by week from initial surgery were 39%, 42%, 41%, 32%, 30%, 28% for the first six weeks and then 42% for weeks 7-12.  Put another way, time between initial arthroplasty and development of PJI did NOT predict success with managing PJI using DAIR, either when time to infection was considered as a continuous or stratified variable.  What did predict treatment failure with DAIR?  CRP as a continuous variable (OR 1.01 per single unit elevation; 95% CI 1-1.01), the arthroplasty having been a revision (OR 4.3 with 95% CI 1.6-11.6) or done for a fracture (OR 2.4 with 95% CI 1.3-4.5), and female gender (OR 1.9 with 95% CI 1.1-3.1).  This makes sense to me- if the concern is about mature biofilm making an infection untreatable with antibiotics alone, the real question isn’t how long it’s been from arthroplasty placement to infection, but how long it’s been from infection of the arthroplasty to the infection being discovered. 31504331

Multiple doses of ivermectin are no better than a single dose for Strongyloides stercoralis infection in a multicenter phase 3 RTC.  This study recruited 309 patients with a positive fecal test or serology for strongyloidiasis and randomized them to receive either one or four doses of 200ug/kg ivermectin.  The authors excluded folks with evidence of hyperinfection or CNS disease and the immunocompromised. Their primary outcome was clearance of infection at 12 months, defined as a negative stool agar plate or PCR test and either a negative serology or at least a 4-fold drop in strongyloides antibody titer.  Everyone in the trial had eosinophilia (Table 1 shows the lowest AEC enrolled was 450 cells/ul and the average was 770-800), which I mention because it suggests they were treating currently and not just previously infected patients.  After a year of enrollment, an interim analysis showed the primary outcome occurred in 86% of the single-dose recipients versus 85% of the 4-dose recipients, and the study was terminated early for futility. 31558376

The outcomes of suppressive antibiotics for PJI are pretty dismal.  So says a multicenter retrospective cohort study (n=302) out of Spain, in which the authors defined treatment failure as appearance or persistence of a fistula tract, need for debridement, removal of a prosthesis, or uncontrolled symptoms of infection.  Two thirds of these patients received antibiotic monotherapy, mostly a tetracycline (40%) or TMP/SMX (35%).  Over a median 37 months, suppressive antibiotics were successful in only 59% of cases, with only 50% of patients still having control of their infection at five years.  Of the failures, resistance to the suppressive antibiotic emerged in a quarter of cases.  Now, I might argue that the definition of “treatment failure” was pretty permissive here – given that suppression is an approach of last resort, perhaps we ought not look beyond the patient surviving, not needing further surgery, and still being able to use their prosthesis.  But still, these data suggest what I’ve been beginning to expect more and more on the orthopedic ID service – long-term antibiotic suppression is a lousy endgame strategy.  31539638

Five days of antibiotics for cystitis in men, anyone?  I suspect male UTI is mostly overtreated, though it’s hard to say so with confidence given how little attention/data the subject has received.  Here’s a nice review of the efficacy data for pivmecillinam, a 1st line treatment option for UTI that’s not available in the US, that sheds some light on the issue.  The authors report a retrospective study examining the outcome of E.coli UTIs empirically treated with pivmecillinam in Denmark over a 6-year period.  They defined treatment failure as receipt of a second prescription or hospitalization due to UTI, and measured this outcome at 14 and 30 days after the initial prescription.  A total 21,864 cases of E.coli UTI treated with pivmecillinam were found, of which 2524 (12%) occurred in men.  The authors found that the most common durations prescribed were 3, 5, and 7 days, and that 3-day treatment versus 5-day treatment was associated with a higher rate of 14-day failure, with HRs ranging from 1.4 to 1.9 in women and 1.5 to 2.0 in men when stratified for age and pregnancy status (p<0.05 for all).  However, outcomes were no different with 5 versus 7 days treatment for either men or women (HRs 0.6 to 1.1 in women and 0.9 to 1.2 in men; p > 0.05 for all).  So, at least when the organism is E. coli and the drug is pivmecillinam, it seems extending treatment of lower tract UTI in men beyond 5 days offers no benefit. 31098630

And now a few short notes:

A third of patients admitted from the ED with a diagnosis of pneumonia are re-diagnosed with something else later.  This study examined the 875 adults hospitalized in an Israeli internal medicine ward, of whom 195 had been admitted from the ED with a diagnosis of pneumonia.  Of these, 56 (29%) received a different diagnosis at discharge, including upper respiratory tract infection (54%), urinary sepsis (14%), and noninfectious problems (30%).  ED diagnoses of pneumonia were more likely to be accurate when the patient had an CRP >10mg/dl (56% vs 30%; p<0.0001) and when the patient actually had an infiltrate on chest X-ray (86% vs 4%; p<0.0001).  I don’t mean to throw shade by highlighting this article – the first and second most important jobs of the ED doc are to stabilize the critically ill and triage who needs to come into the hospital, respectively.  Rather, I point it out to remind that other clinicians’ diagnoses shouldn’t be accepted uncritically (particularly for ID docs, as we’re often consulted because the patient is poorly with the current treatment plan).  31479761   

Two weeks of antibiotics may be adequate for bone and joint infections in children.  This retrospective study out of France examined 56 cases of osteomyelitis, 95 cases of septic arthritis, and 25 case of concurrent osteo and septic arthritis; the most frequently identified pathogens were S.aureus (39%) and Kingella kingae (27%).  The mean duration of IV therapy was just four days, and the median total duration of therapy was just 15 days.  There were no treatment failures. 31504582

In invasive meningococcal infection, does receiving a dose of oral antibiotics before you get to the hospital improve your odds of mortality?  This observational cohort study of 527 patients out of Spain (24% of whom got pre-hospital antibiotics) suggests just that.  Mortality was ten-fold higher in those who didn’t get pre-hospital antibiotics (0.8% vs 8% with p=0.003), and remained protective in multivariate analysis (OR 0.2 with 95% CI 0.01-0.88) but not after propensity score matching (OR 0.2 with 95% CI 0.02-1.42).  Still, these data are promising enough that I think I’d rather see this practice implemented widely and then studied rather than the reverse. 31140070