IDJournal.Club - The ID Literature In Review

ID consultation is associated with better quality of care and lower mortality in patients with candidemia. Disclosure: I know the senior author of this paper, twitter colleague Andrej Spec at Washington University in St. Louis. There’s a good body of literature supporting the value of ID consultation in S. aureus bacteremia, but less evidence of for other conditions. Here, Mejia-Chew and colleagues provide data extending the proof of our utility to another life-threatening bloodstream infection, candidemia.

The authors performed a retrospective single-center study of adults with candida bloodstream infection treated between 2002 and 2015. In addition to demographics and clinical risk factors, they recorder whether the patients had central lines removed, had workup for endopthalmitis and endocarditis, what type of antifungal therapy they received, and their 90-day all-cause mortality. They stratified the cohort based on whether or not the patietns received an ID consultation. In the survival analysis, they used inverse weighting by propensity score to assess the effects of ID consultation on measures of quality clinical management and mortality. Some important notes on the methods: First, in cases where no antifungal therapy was given for a candidemia, the authors reviewed the charts to sort the cases into three groups: the positive blood culture was treated as a contaminant, the treating clinicians were unaware of the positive culture, or the patient left AMA before the culture could be treated. Second, they only analyzed the first episode of candida bloodstream infection for each patient and they excluded patients who died within 24 hours of the positive blood culture being collected, as well as those who were not treated for their candidemia because they opted for palliative care.

The authors identified a total 1794 patients with candida bloodstream infection, of whom 103 were excluded from the analysis (94 due to death within 24 hours, 6 due to opting for palliative care, 2 for having recurrent candida infections [i.e. they’d had candidemias in the last few months of 2001], and 1 for missing data). Of the remaining 1691 patients, the mean age was 58, just over half were men, and their most common predisposing factors for candidemia were solid or hematologic malignancy (33-36% and 14-20%, respectively), receipt of TPN in the preceding 30 days (29-31%), and receipt of steroids in the preceding 90 days (26-28%). Only 46% of patients received an ID consultation.

Patients who received an ID consult were just as likely to get azole versus echinocandin or amphotericin therapy, but they received longer durations of therapy (mean 18 versus 14 days; p<0.001), and their positive blood cultures were less likely to have gone untreated (2% vs 14%; p<0.001). Patients who received an ID consultation also received more followup blood cultures (9 vs 7; p<0.0001), resulting in more positive blood cultures (2 vs 1; p<0.0001), which may explain the extended course of antifungal therapy; those with ID consults were also more likely to get an echocardiogram (57% vs 33%; p<0.0001) and a diagnosis of endocarditis (4% vs 1%; p<0.0001), an ophthalmology consult (53% vs 13%; p<0.0001) and a diagnosis of endophthalmitis (3% vs 1%; p<0.0001), and removal of a central line (76% vs 59%; p<0.0001). Finally, the patients who received an ID consult had lower 90-day mortality (29% vs 51%; p<0.001), with a calculated NNT of about 5 to prevent one death with ID consultation.

So, add candidemia to S.aureus bacteremia on the list of diseases that should mandate an ID consult. 31562024

Marabavir appears about as effective as valganciclovir for suppression of CMV viremia in hematopoietic stem cell and/or solid organ transplant recipients. Marabivir is a novel CMV antiviral that inhibits the UL97 protein kinase. It’s in late stage clinical trials but has had a rocky development course; it looked great as a prophylactic in phase II trials, but then failed to show benefit versus placebo at preventing CMV viremia in a phase III trial in 2009 (later, critics pointed out that the study may have used an inadequate dose). Anyway, now it’s back in clinical trials, this time being used for preemptive treatment of CMV.

The authors conducted an open-label but dose-blinded trial randomizing adult HSCT and/or SOT recipients with a CMV viral titer of 1,000-100,000 copies/ml to receive either standard therapy with valganciclovir, or 400mg, 800mg, or 1200mg of marabivir twice daily for up to 12 weeks. Their primary endpoint was response to treatment, defined as an undetectable CMV plasma viral load, by 3 and 6 weeks after initiation of treatment; they also had a safety endpoint, defined as the incidence of new or worsening adverse events during treatment. Importantly, the sample size was not chosen based on power calculations for the efficacy endpoint, but was “based on the feasibility of assessing the safety of the trial drugs and the antiviral activity of maribavir.” I don’t know what means.

Anyway, a total 161 patients were randomized and included in the ITT population (40 in each arm except the valganciclvoir group, which had 41 patients). Of the folks who had postbaseline CMV viral load data available, 62% of those who received marabavir had a response within 3 weeks vs 56% of those who received valganciclovir (RR 1.1 with 95% CI 0.8-1.5); among the marabivir recipients, response rates were similar across doses. Outcomes were similar at 6 weeks (79% response in the marabivir group vs 67% in the valganciclovir group; RR 1.2 with 95% CI 1.0-1.5); interestingly, 6 week outcomes seemed to more clearly favor marabivir vs valganciclovir among the HSCT recipients (75% vs 48%). The percentages of patients who had recurrent CMV infection during the trial was similar between groups (22% in the marabivir group vs 18% in the valganciclovir group), as were median times from suppression of CMV viremia on treatment to recurrent CMV infection (72 days in the marabivir group vs 80 days in the valganciclovir group). Two patients who initially cleared their viremia on marabivir had later relapses with emergene of marabivir resistance via mutation T409M in the UL97 protein kinase.

With regards to safety, 67% of marabivir recipients and 22% of valganciclovir recipients had adverse events judged likely to be due to trial medication. Most of these were mild to moderate; among those who got marabivir, the most common side effects were dysgeusia (40%, with no difference between doses), and GI side effects (nausea, 23%; vomiting, 20%, and diarrhea, 20%, all worse with escalating marabivir doses). The most common severe adverse events (reported in 3% each)were GVHD, severe diarrhea, renal failure, and UTI; overall, the incidence of severe AEs was lower than in those who got valganciclovir. Rates of death were similar across arms (cumulatively 5% for the marabivir recipients and 8% for the valganciclovir recipients).

So, marabivir seems to be as effective as valganciclvir at clearing CMV viremia and preventing recurrent CMV infection in HSCT and SOT patients, when used with a preemptive treatment strategy, in this small (maybe not adequately powered?) trial. On the flip side, it seems to be much less well tolerated, though severe AEs are no more common that with valganciclovir. Transplant isn’t my area of focus, but I guess I’d be surprised to see marabivir replacing valganciclvoir any time soon – it looks to be no better than the current standard of care, and that huge incidence of dysguesia (an AE patients particularly despise – just ask anyone on metronidazole) and GI side effects will be very unwelcome. I’d certainly rather gamble my neutrophils with valganciclovir if I was the patient. 31532960

Chlorhexidine bathing was associated with a reduced rate of central-line associated blood stream infection in a hematology ward. There’s some recent data to support this practice in the ICU / critical care literature. I guess it’s not too surprising to see benefit in another group of patients who are at particularly high risk of CLABSI, though it is a little given the traditional dogma of neutropenic fever in cancer/chemotherapy patients being due to bacterial translocation from breakdown of the GI and oral mucosa. Anyway, these authors took patients getting chemo for hematologic malignancies and offered them daily baths with 2% chlorhexidine. The primary outcome was rates of CLABSI or bacteremia with gram-positive organisms (i.e. potential skin flora). A total 893 patients were included in the study, and the incidence of the primary outcome was modestly lower in the chlorhexidine group (3.4 vs 8.4 infections per 1000 patient-days; p=0.02); this association persisted in multivariate analysis (HR 0.4 with p<0.001). The bathes were well-tolerated. Seems like a safe and reasonable intervention, albeit one that targets the minority of the infections this patient population develops. 31504341

The cobas6800/8800 system’s CMV PCR looks to be more sensitive than the traditional cobasAmpliPrep/TaqMan assay. The Mayo transplant group evaluated this assay against the standard quantitative PCR test in 36 SOT and 28 HSCT recipients. The new assay yielded higher viral loads and longer durations of detected viremia – 44% of patients with a negative CMV viral load by the traditional assay had detectable viremia with the newer assay. Most importantly, clearance of viremia with the new assay at the end of therapy (vs just a negative traditional PCR) was associated with lower odds of future relapse (OR 0.26; 95% CI 0.04-0.99). But does this study tell us who will genuinely benefit from more antivirals, or just who is doomed to relapse once they’re withdrawn? 31504626

T-cell infusions may be an important new approach for control of chronic viral infections in severely immunocompromised patients. This was a translational proof-of-concept paper with no real clinical component, but I think the methods make it worth a share. The authors collected peripheral WBCs from donors with detectable antibodies against norovirus (recall that not everyone is susceptible to every genotype of norovirus, because noroviruses require the host to express different histo-blood group antigens to establish infection in a genotype-dependent manner), then stimulated those WBCs with short overlapping peptides synthesized from the coding regions of the genome of GII.4, the most common norovirus genotype. They were able to generate polyclonal T-cells with specificity for various norovirus proteins that were broadly cross-reactive against other norovirus strains (reportedly; the paper is a bit light on details here). T-cell immunotherapy has been FDA-approved for a couple of types of lymphoma, and has been used with some success in a clinical trial / investigational capacity for PML and CMV infections. So, perhaps this approach could be generalizable for a variety of viruses; if they actually impact the disease course when put in patients, these ex-vivo generated T cells could be a nice answer to the longstanding problem of most viruses not having a clinically effective antiviral. 31562500

Oral fosfomycin is a reasonable choice for cystitis among kidney transplant recipients. I’m not sure why we’d be worried otherwise, but hey, now there’s data. This was a review of 133 kidney recipients treated with fosfomycin for cystitis in Spain between 2005 and 2017. Most were due to GNRs (though only 67% were E.coli), and about half were multidrug resistant. The patients received a median daily dose of 1.5g for a median 7 days (note that there are two formulations of fosfomycin in Spain: fosfo trometamol, the 3g sachets given every 72 hours that we’re familiar with in the US, and fosfo calcium, which is a 500mg capsule taken 2-3 times daily). Patients achieved a clinical cure in 84% of cases, and microbiological cure in 74% of the cases who had followup urine cultures a week later. Treatment failure was more common among men than women, and among those who did vs did not have a percutaneous nephrostomy. I did think these numbers might be off: specifically, the rates of clinical failure reported among the men and women are both higher than the total failure rate of the cohort, which doesn’t seem possible. 31550408