Antimicrobial agents: October 2018

When treating TB meningitis, should you sub in a fluoroquinolone for one of the components of standard 4-drug antitubercular therapy? This meta-analysis says no. I was taught that TB meningitis should be routinely treated with a quinolone in place of ethambutol, because the CNS penetration of the former agent is great, whereas the penetration of the latter agent is poor. Makes sense, right? But this meta-analysis of published RTCs on the subject (five studies; total N = 1115) leaves the question open. The authors found no reduction in mortality with either substituting a quinolone for a component of standard 4-drug therapy (RR 0.68; 95% CI 0.3-1.5) or adding a quinolone to standard 4-drug therapy (RR 0.97; 95% CI 0.8-1.2). A 30% reduction in mortality sure would be meaningful if the association was significant, but it wasn’t – that’s why I say the question is open. Looking more granularly, neither quinolone substitution for ethambutol nor quinolone substitution for rifampin offered benefit, and substitution for ethambutol suggested the possibility of increased mortality (RR 1.17; 95% CI 0.7-2.1). While overall adverse event rates were similar between groups, quinolone-containing regimens were associated with significantly increased risk of seizures (RR 2.1; 95% CI 1.0-2.4) and vision loss (RR 3.5; 95% CI 1.2-10.6) and a nearly significant increased risk of jaundice (RR 1.7; 95% CI 0.95-3).

Pending more data, I think I’ll stick to good old rifampin, isoniazid, pyrazinamide, and ethambutol for TB meningitis. 30017610

Crash and PhagoBurn: a bacteriophage cocktail failed to improve Pseudomonas aeruginosa burn wound infections in a trial hampered by technically difficulties. Bacteriophages have a long history of use as therapeutic “antimicrobials” in the former Soviet Union and its successor states, but only after a few recent and dramatic case reports (See The Perfect Predator by Dr. Steffanie Strathdee – or the closing plenary of IDWeek 2018 - for details) have they come to the attention of the West. Phages seem like the perfect answer to antimicrobial resistance – wouldn’t it be great to have an antimicrobial that can evolve right back? On the other hand, challenges to the widespread adoption of phages include concerns about safety, the narrow host range of most phages, and the potential for development of (human) host immunity to the phage, which might limit their efficacy in systemic infections with repeated use.

On to the study. The authors conducted a randomized controlled trial to compare a cocktail of twelve lytic Pseudomonas aeruginosa phages prepared as a 10^9 pfu/ml solution in buffered saline (“PP1131,” which was applied as a liquid to wound dressings) to standard care (1% sulfadiazine silver cream) for the treatment of P. aeruginosa burn wound infections. The primary outcome was median time to a sustained reduction in bacterial burden of wound swabs by at least two quadrants in a four-quadrant agar plate streaking method. The authors enrolled 27 patients from nine burn centers in France and Belgium over a two-year period; the trial was stopped early for futility. The primary outcome was achieved in a median of 144 hours in the PP1131 group versus 47 hours in the standard care group (HR 0.27; p=0.02). Half of the PP1131 recipients had no reduction in bacterial burden versus only 15% of sulfadiazine silver recipients, and failure of PP1131 to reduce the bacterial burden was associated with isolation of P. aeruginosa strains that were poorly susceptible or resistant to the phage cocktail.

So is this study proof phage therapy is bogus? Nah. The authors had intended to apply the phage preparation at an MOI of 10 (ten phage particles for every bacteria in the wound), which is a standard dose in the virology world if you want be sure every last cell gets infected in the first round of replication. However, it turned out that the phage preparation was unstable, to the degree that the phage dose patients received was 1,000-10,000 fold lower than intended. Interestingly, despite not clearing their bacterial burdens as quickly or well as sulfadiazine recipients, the PP1131 recipients had half the rates of septic shock and infectious complication. The authors hypothesize that this may have been due to systemic absorption and circulation of the phages, and/or their immunomodulatory and antinflammatory effects on the host. I want to suggest a totally different hypothesis – maybe the Pseudomonas communities’ physiologic responses to phage infection involved downregulation of virulence factors, making them less pathogenic. Either way, halving rates of gram-negative septicemia in burn patients seems like a worthy goal and sufficient grounds for another trial.

Also, my favorite phage cocktail is the Gin & T3. 30292481

The ACTIVE trial asked: which is better for invasive Candida infections, isavuconazole or caspofungin? Previous studies have suggested the superiority of echinocandins over fluconazole for candidemia and other invasive infections, but what about isavuconazole? This trial randomized adults with candidemia or invasive candidiasis to receive isavuconazole or caspofungin (both initially IV), with the option to switch to oral isavuconazole (or voriconazole for the caspofungin group) at day 10 for up to a total 56 days of treatment. The primary endpoint was treatment success at the end of IV therapy among patients with proven infection, and secondary outcomes included treatment success at 2 weeks after completing all antifungals, mortality at days 14 and 56, and drug safety.

The modified intention to treat population included 400 patients from the 450 who were randomized. The groups were well-balanced. Treatment success at the end of IV therapy was 71% in caspofungin arm versus 60% in the isavuconazole arm (difference in efficacy 11%; 95% CI 20% to 2%). However, no differences between groups were observed in any of the secondary endpoints (e.g. clinical success by the end of all therapy and mortality), and time to blood culture clearance in cases of candidemia were similar. So, it was kind of a wash. 30289478

Speaking of candidemia, a retrospective study in EJCMID reported outcomes in 302 patients with candidemia stratified by appropriateness of treatment based on the CLSI and EUCAST antifungal susceptibility breakpoints. The bad news was that 90-day mortality in the cohort was a whopping 55%. The good news is that appropriate definitive therapy had a statistically significant association with reduced 90-day mortality for both the 2012 CLSI (OR 0.3; 95% CI 0.1-0.8) and EUCAST (OR 0.4; 95% CI 0.2-0.8) definitions of antifungal susceptibilities. You may say “of course patients do better when they get antifungals active against their infections,” but the reason this is worth noting is that it backs up the relevance of the CLSI and EUCAST antifungal breakpoints with clinical data. 30284179

Egypt can treat 2,100 people for HCV infection for cost of one round of treatment in the United States. This study of 971 patients treated in the Egyptian national health program compared the efficacy of hepatitis C treatment using brand-name versus generic sofosbuvir-daclatasvir or sofosbuvir-ledipasvir (which they’ve managed to get down to $84/person for a course of treatment). The groups were similar in baseline characteristics, and the overall sustained virologic response rate was 98%, with no differences in outcome between recipients of the generic and brand-name drugs. UpToDate lists US pricing for daclatasvir and sofosbuvir at $2,100 a day, for a total cost of $176,400 for one 12-week course of treatment. I wonder if United Healthcare would let you take a 3-month vacation to Cairo to get treated and then pocket the difference? 30077791

Actually, it turns out there’s a sad but interesting story behind Egypt’s interest in improving access to HCV treatment. In the 1950s the Egyptian Ministry of Health started a mass drug administration campaign to eradicate schistosomiasis, then the country’s leading cause of liver disease. The treatment of choice for schistosomiasis, tartar emetic, was injectable, and needles ended up being reused – can you can see where this is going? In an ironic twist of fate, the Ministry of Health ended up causing a massive outbreak of HCV infection, which was only discovered with the advent of HCV serologies in the 1990s. By that time, HCV had replaced schistosomiasis as the major driver of liver disease in Egypt, with an estimated 6 million people infected. Credit to the Egyptian public health system for trying to make it right. This 2006 article in Hepatology has more details: 16628669

Amoxicillin-ceftibuten-clavulanate for the treatment of ESBL pyelonephritis??? I can see how the authors got to the point of thinking of this regimen. Where I practice, we just don’t have a reliable outpatient regimen for pyelonephritis; the incidence of fluoroquinolone resistance among E. coli in the community exceeds 30%, and resistance to trimethoprim-sulfa is almost as bad. But I’d be very nervous about using oral beta-lactams alone for pyelonephritis - they’re already associated with greater rates of clinical failure and relapse in uncomplicated cystitis! In my book, pyelonephritis due to an ESBL-producing organism gets a quinolone, trimethoprim-sulfa, an aminoglycoside, or a carbapenem.

Still, this group demonstrated that the combination of amox-clav and ceftibuten for 14 days was effective for ESBL-producing E. coli and K. pnuemoniae pyelonephritis in ten patients. All patients experienced clinical cure, and all eight who had follow-up urine cultures within the next three months had eradicated the original organism. The authors took the urine isolates and performed disc diffusion testing, showing that placing an amox-clav disc next to a ceftibuten disc extended the zone of inhibition of the ceftibuten disc, demonstrating a synergistic interaction. This isn’t enough data to make me comfortable with the regimen, but it is enough to make me interested in seeing the results of a study with a bigger sample size. 30117050

Early switch to oral linezolid yields outcomes similar to standard IV therapy in patients with low-risk Staphylococcus aureus bacteremia (SAB). This was a retrospective cohort study of patients treated at a single academic center over a 4-year period. The authors compared patients with uncomplicated SAB who were treated with a standard 14-day course of IV antibiotics versus those who were switched to linezolid between day 3 and 9 of therapy. Propensity matching with a 2:1 ratio was used to form the comparison groups.

For what it’s worth, using propensity score matching to generate a balanced cohort is methodologically problematic; I confess I don’t have the statistical chops to properly explain why, but folks smarter than me say so (if you can explain this to me in plain English I will buy you a beer at IDWeek 2019) .

In any case, the authors ended up with 135 cases of SAB (standard care group, n=90; oral linezolid group, n=45). Half of the cases of SAB were related to a central venous catheter, and most of the remainder were either due to skin and soft tissue infections or no had no readily identifiable source. Fewer patients in the linezolid group had CKD (4% vs 22%) but otherwise the groups were similar; with regards to the patients’ treatment, rates of source control were comparable, and ID consultation was obtained in 100% of patients switched to linezolid. Switch to oral linezolid was associated with a shorter duration of IV therapy (7 versus 15 days; p < 0.01) and shorter length of stay (8 versus 19 days; p <0.01). Rates of 90-day relapse were similar between the linezolid and standard care groups (2% vs 4%; p=0.87), and there was a trend toward lower 30-day mortality in the linezolid group (2% vs 13%; p=0.08). The authors conclude, “Treatment of SAB in selected low-risk patients with an oral switch to linezolid from day 3 to 9 of treatment until completion yielded similar clinical outcomes as [continued IV therapy], allowing earlier discharge from the hospital.”

If you read this month’s header you understand why this study makes me nervous. It’s not that I doubt the potency oral antibiotics, particularly when we’re talking about a completely bioavailable drug like linezolid. Rather, I fear a nonspecialist reading this study and interpreting it as “it’s OK to switch to oral linezolid and discharge on day three so long as the patient doesn’t have endocarditis.” Because ultimately the optimal treatment of SAB is mostly about identifying and controlling foci of infection, not whether you give oral or parenteral therapy, and that requires the sort of careful workup that’s easily undercut by a hasty discharge. Had I written the abstract’s conclusion, it would have read something like “Early switch to oral linezolid in patients with uncomplicated SAB who have undergone a week of inpatient diagnostic evaluation for metastatic infection, including an infectious diseases consultation, was associated with similar clinical outcomes and shorter length of stay than continued IV therapy.” 30351401