The 2018 ABIM infectious diseases subspecialty certification exam in is in two weeks, which means I need to go cram trivia like which new parasite causes anemia in Bolivian schoolchildren and which causes meningitis in people who eat raw centipedes. I’ve been trying to ignore it, because every time I remember the boards are coming, this is how I feel.
Anyway, October marked my last clinical rotation of fellowship, so I thought I’d discuss a difficult case we encountered and some of the literature surrounding the topic. The problem was ID bread and butter: MRSA in the blood, the hunt for an elusive source, and what to do when bacteremia persists.
If you aren’t interested in this discussion, skip to the bottom of this post for links to this month’s literature reviews.
The patient is a young woman with cystic fibrosis and its typical complications: pancreatic insufficiency, brittle diabetes with an A1c just shy of double digits, malnutrition managed with supplemental feeds by PEG tube, and a right chest wall port placed for difficult venous access after a lifetime of daily morning phlebotomy.
A couple months prior she undergoes bilateral lung transplantation with basiliximab induction. In the immediate postoperative period, she develops acute respiratory failure and hypotension requiring ECMO and vasopressor support. Blood cultures grow methicillin-resistant Staphylococcus aureus in one of four bottles with the following susceptibilities:
Vancomycin: Susceptible (MIC 0.5)
During her initial decompensation, the transplant team starts empiric meropenem and linezolid (not because she had suspected MRSA pneumonia – we are not adherents of ZEPHyR – but because the patient has a history of true anaphylactic reaction to vancomycin). After the results of the cultures become available, her antimicrobial regimen is changed to ceftaroline. Repeat blood cultures drawn 48 hours after the initial set are sterile and ceftaroline is continued for two weeks. The patient improves, weans from ECMO and pressors, has an uncomplicated further hospital course, and goes home a week later.
OK, so hindsight is 20/20, but since you already know the story doesn’t end here, a few questions should be coming to mind:
Why treat MRSA bacteremia with linezolid or ceftaroline rather than daptomycin? The 2011 IDSA MRSA treatment guidelines recommend either vancomycin or daptomycin for bacteremia. I couldn’t find explicit reasoning in the chart, but my guess is that these drugs were chosen over daptomycin because of the acute respiratory failure (and the typical “pulmonary opacities, cannot rule out infection” report on the initial chest X-ray). Daptomycin is inhibited by surfactant, so its use in respiratory infections is limited to septic pulmonary emboli.
It wrong to treat MRSA bacteremia with linezolid or ceftaroline? I’ve been unable to find any high-quality clinical data to suggest inferior outcomes with either drug when compared to vancomycin or daptomycin for MRSA bacteremia. A single small retrospective study (n=122) found higher mortality in patients with MRSA bacteremia who received linezolid rather than vancomycin or daptomycin (40% vs 14%; p=0.019), but with only 15 patients in the linezolid arm I take this finding with a grain of salt. A multicenter observational study of ceftaroline for MRSA bacteremia reported a clinical success rate of 68%, which sounds low but is comparable rates reported for vancomycin. Finally, this recent prospective study out of Japan examined MRSA infection outcomes stratified by antibiotic strategy, and found a trend toward increased 30-day mortality with use of vancomycin versus daptomycin, linezolid, or teicoplanin (HR 2.47; p=0.067) and a statistically significant increase in mortality in the subgroup with lung infections (HR 4.85; p=0.034). I couldn’t find any papers comparing outcomes of MRSA bacteremia treated with ceftaroline versus daptomycin.
Was an echocardiogram and/or a longer course of therapy indicated? Let’s go back to those 2011 MRSA guidelines. To be categorized as an uncomplicated bacteremia, which is adequately treated with two weeks of IV therapy, a MRSA bloodstream infection should meet the following criteria:
There are no implanted prosthetic devices (check - once she came off ECMO)
Repeat blood cultures are sterile within 2-4 days (check)
Fever has resolved within 72hr of antibiotics (she never had a fever – but on ECMO, who does?)
There is no evidence of metastatic sites of infection (check)
Endocarditis has been excluded by echocardiography, preferably transesophageal (not done)
Researchers have tried to characterize risk factors for endocarditis in Staphylococus aureus bacteremia, attempting to triage who really needs a TEE versus a simple transthoracic echocardiogram. This 2016 prospective multicenter cohort study out of France (n=2,008) produced a 10-item risk scoring system, by which our patient would have had been in the low risk category, with a predicted endocarditis incidence of 1.1%. JAMA’s 2014 review of the topic cites 5 studies (n=3,333) that enumerate endocarditis risk factors without which a patient might forgo TEE: these included the presence of a permanent intracardiac device (i.e. valve, pacemaker, or ICD), prolonged bacteremia, dialysis dependence, community-acquired infection, metastatic foci of infection, and embolic or immunologic phenomena, none of which our patient had.
The MRSA guidelines are clear that every case of Staphylococcus aureus bacteremia (SAB) should receive an echocardiogram. Without a negative echo you can’t make a diagnosis of uncomplicated SAB and ought to be treating for 4-6 weeks. That said, as she didn’t have evidence of infective endocarditis on her subsequent visits, I suspect the echo would have been negative, classifying her bacteremia as uncomplicated and her duration of antibiotic therapy as adequate.
For my part, I increasingly doubt whether uncomplicated SAB should even exist as a diagnostic entity. This recent RTC of staphylococcal bacteremia treatment in JAMA, which had so many risk factors for complicated bacteremia as exclusion criteria that the authors had to screen nearly 30,000 patients in order to reach their sample size of 500, still found that a third of SABs initially thought to be uncomplicated were later found to be complicated. If we’re that bad at misdiagnosing the condition, given that undertreatment of SAB frequently has dire consequences, maybe we should do away with the notion of “uncomplicated SAB” entirely.
But let’s get back to the story.
Our patient returns two weeks after her discharge with chest pain and dyspnea and is found to have a big pleural effusion. This is drained and grows – you guessed it – MRSA, also present in the blood. The pleural fluid is exudative but not loculated or frankly purulent, and rather than undergo VATS and decortication the team manages our patient’s empyema with catheter drainage and serial instillations of tPA and dornase, eventually leading to total resolution of the fluid collection. For their part, the ID team orders a TTE (negative), has her port removed as a potential nidus of infection, starts daptomycin 8mg/kg, and then sends her home on that via a PICC line after the blood cultures have sterilized with a plan to complete 4 weeks of antibiotics.
A week after discharge she comes into the transplant clinic with diffuse myalgias. Her serum creatine kinase is 5000. Daptomycin-induced rhabdomyolysis - crap. Her transplant docs pull the PICC line and send her home on oral linezolid to finish the remaining two weeks of her 4 weeks of treatment.
A couple days after finishing her course of antibiotics she comes into the transplant clinic again, now with malaise and low-grade temperatures at home but no localizing symptoms. She’s readmitted to the hospital, where blood cultures are positive for – you guessed it – MRSA. And this time the bacteremia refuses to go away.
On admission her physicians document that she has a normal physical exam. Now, normally if you said “complicated SAB” and “normal physical exam” to me in the same sentence, my eyes would roll out of my head. But I went back and looked and four different physicians all chart the same thing without using copy/paste, including at least one person who I trust to have done a proper, thorough, effortful examination. And when I examined her later, the only remarkable thing was how chipper she was for someone who’d been bacteremic for the past two weeks. So, I promise you there was no readily apparent source by exam.
Her bacteremia persists, as I said, for two weeks before I join the service. During this time she receives an extensive diagnostic workup and various combinations of antibiotics, which I have detailed below:
What to do now?
Well, I wanted an FDG-PET scan. The obvious source was the previous empyema, but neither the plain film nor the chest CT, which usually overcalls effusions, showed even a trace of fluid - and good luck convincing anyone to go in for thoracoscopy with negative imaging. The other obvious potential source was endovascular. There is, after all, only so much of the human body we can ultrasound. And it turns out that FDG-PET is a pretty darn good tool for locating the source(s) of persistent SAB. In a single-center retropective cohort study from the Netherlands (n=184), FDG-PET in the workup of staphylococcal bacteremia was the first study to identify a majority of the metastatic foci in recipients of the scan, including 75% of the endovascular infections, 75% of the pericardial and mediastinal infections, and 61% of the cases of endocarditis. In FDG-PET recipients, the scan led to treatment modification in 70% (including an additional source control procedure in 20%), and in multivariate analysis, receipt of FDG-PET was the only factor associated with 90-day survival (OR 0.204; p=0.005). But alas, her insurance was only willing to cover an FDG-PET performed as an outpatient.
Note: If you’re reading this and happen to be on the MRSA guidelines committee (who am I kidding?), do me a solid and make sure the next edition of the guidelines recommends FDG-PET for the evaluation of persistent SAB, so that insurance companies will actually cover it. Thanks!
So our plan became to clear the bacteremia, discharge her on at least six weeks of home IV antibiotics, and get an outpatient FDG-PET to see what source control procedures needed to be done. This lead to our next dilemma: what antibiotic regimen to choose?
Add rifampin? I asked; our PharmD laughed and then said no. Not only would it have wreaked havoc on the patient’s tacrolimus dosing, a recent RTC shows adding rifampin for SAB doesn’t help anyway.
Add bactrim? Limited data supports the use of combination therapy with ceftaroline and trimethoprim-sulfa for salvage of persistent SAB. When I say limited I mean “one case series with 29 patients.” Blood cultures sterilized quickly on this regimen, though short-term mortality was over 30%.
Add fosfomycin and imipenem? Don’t laugh, there’s data. A Spanish group reported their experience using fosfo/imipenem as salvage therapy in 16 patients with persistent SAB back in 2014. Again, cultures sterilized quickly, but about a third of the patients ended up dying anyway. This wasn’t an option for us on account of IV Fosfomycin having yet to be approved in the US.
Add oritavancin? Oritavancin, like daptomycin, has evidence for synergy with beta-lactams (reference here). This makes sense: beta-lactam antibiotics damage gram-positive bacterial cell walls, providing better access to the cell membranes both daptomycin and oritavancin act to disrupt. However, I couldn’t find any clinical data to support the combination of oritavancin with ceftaroline or daptomycin (or all three) for persistent SAB.
Add dalbavancin (or desensitize to vancomycin)? We did not seriously consider either of these options. But for what it’s worth, yes, both vancomycin and dalbavancin have in vitro data for synergy with beta-lactams, and vancomycin has an RTC suggesting clinically meaningful synergism in the CAMERA trial.
I suppose we could have also considered linezolid or an aminoglycoside (though the MRSA guidelines advise against the latter approache), but what we actually did was sit tight and continue combination therapy with ceftaroline and daptomycin, the latter of which we’d restarted with CK montoring every 2-3 days. The patient’s blood cultures became persistently sterile starting on day 16 and her CK stayed normal after more than a week of daptomycin therapy. We eventually sent her home on daptomycin and ceftaroline with strict instructions to return with any myalgias, close clinic followup, and a referral for her outpatient FDG-PET.
I want to end by talking about the combination of ceftaroline and daptomycin. This regimen is wildly expensive, but there’s good reason to think it should be superior to daptomycin monotherapy. First, beta-lactams enhance binding of daptomycin to bacterial cell membranes. Second, penicillin binding protein 1 is important to the bacterial compensatory response to daptomycin-induced injury, resulting in a “seesaw effect” in which MICs to beta-lactams drop as MICs to daptomycin rise, and hence the organism is unable to simultaneously maintain resistance to both drugs. Evidence suggests all of this is likely a class effect of beta-lactams (or at least those with affinity to the right PBPs) rather than just ceftaroline. Ceftaroline seems like the obvious choice among the beta-lactams on account of having its own intrinsic activity against MRSA, but perhaps cheaper beta-lactams would work just as well.
What’s the clinical data to support combination daptomycin and ceftaroline for persistent SAB? Mostly case series, which again, show the blood cultures sterilize quickly but the patients do poorly. Here we’re seeing an inherent limitation of retrospective research, which I think is nicely illustrated by this IDWeek 2018 poster concluding that “patients receiving daptomycin/ceftaroline did not experience any significant improvement in clinical outcomes compared to patients who received vancomycin.” This sounds like a fair conclusion if you just read the endpoints given in Table 3; however, looking at Figure 1 we see that half the patients in the daptomycin/ceftaroline group were switched to that regimen due to persistent SAB on their prior antibiotics. Well, of course people who’ve been bacteremic for days and days despite effective antibiotic therapy are likely to have a poor outcome! I think we ought to have an RTC comparing one or more daptomycin/beta-lactam combinations to daptomycin monotherapy for SAB, and I’ve heard the SNAP trial being organized in Australia may do something similar.
And now for ID literature from the month of October. Here’s what I have for you:
Antimicrobial agents research included the role of quinolones in TB meningitis and phages in burn wounds, the efficacy of caspofungin versus isavuconazole for invasive candidiasis and generic versus brand name antivirals for HCV infection, and the use of double beta-lactams for ESBL pyelonephritis and oral linezolid for Staphylococcus aureus bacteremia.
ID diagnostics research this month covered the diagnostic accuracy of multiplex PCR viral respiratory panels, the emergence of secretly-linezolid-resistant VRE Enterococcus faecium, the potential value of urine PCR for Whipple’s disease, and procalcitonin for the early cessation of antibiotics in patients with bacteremia.
General ID research topics included using the sputum MTB PCR to rule out active TB and discontinue airborne isolation, short-course therapy for gram-negative bacteremia, a 3-day atovaquone-proguanil regimen for malaria prophylaxis, clinical validation of the definition of complicated S. aureus bacteremia (sort of), and increased fluid intake for the prevention of recurrent UTI in women.
HIV and STD publications included a case series of rectal STDs masquerading as inflammatory bowel disease, extragenital STDs in transgender patients, the superiority of seven versus one days of metronidazole for trichomoniasis, and a description of clinical presentations and prognostic indicators in a large series of patients with neurosyphilis.
Onc and transplant ID research this month included pretransplant CMV-specific T cell function as a marker of reactivation risk in solid organ transplant recipients, a report of letermovir resistance in a transplant patient with a novel UL56 mutation, Mucorales (and C. difficile!) on hospital linens, and the subsequent infection risk after an episode of febrile neutropenia.
Antimicrobial stewardship and infection control research included the failure of decontamination regimens to prevent MDR bloodstream infections in the ICU, antibiotic locks to prevent CLABSI in dialysis patients, the positive long-term impacts of restricting quinolone use, and the use of urine cultures for early antibiotic deescalation in bacteremic UTI.