Skip Navigation

In the Literature

  1. Stan Deresinski, Section Editor

Would Sparing the Anaerobes Prevent Graft-Versus-Host Disease–Related Mortality?

Shono Y, Docampo MD, Peled JU, et al. Increased GVHD-related mortality with broad-spectrum antibiotic use after allogeneic hematopoietic stem cell transplantation in human patients and mice. Sci Transl Med 2016; 8:339ra71. doi:10.1126/scitranslmed.aaf2311.

Narrowed diversity of the intestinal microbiome is associated with an increased risk of mortality in allogeneic hematopoietic stem cell transplant (allo-HSCT) recipients [1]. Holler and colleagues previously found that major shifts in the makeup of the microbiome, including loss of diversity, occurs early in the course of transplantation in association with antibiotic treatment, often given for febrile neutropenia [2]. Among the causes of the increased mortality is graft-vs-host disease (GVHD). Unfortunately, allo-HSCT recipients cannot escape exposure to broad-spectrum antibiotics, many of which drastically affect the anaerobic intestinal bacterial flora. Shono and colleagues examined the hypothesis that use of an antibiotic that relatively spares anaerobic commensals is associated with a reduced risk of severe GVHD.

The investigators evaluated 857 adult allo-HSCT recipients (recipients of T-cell–depleted grafts were excluded) seen over 13 years at their center. The overall 5-year GVHD-related mortality was 14.9%. They identified the 12 most frequently administered antibiotics given from 7 days before through 28 days after transplantation. They found that the 2 antibiotics with the greatest association with GVHD-related mortality when comparing recipients to nonrecipients of each antibiotic were piperacillin-tazobactam (P = .007) and imipenem-cilastatin (P = .025). No other antibiotic was significantly associated with this complication, and 2 antibiotics that are frequently used in empiric therapy in these patients and which have limited antianaerobe activity, cefepime and aztreonam, were not associated with GVHD-related mortality.

Treatment with either piperacillin-tazobactam or imipenem-cilastatin was significantly associated with an increased incidence of grade 2–4 GVHD as well as with upper gastrointestinal and lower gastrointestinal GVHD (the latter is the type that most frequently proves to be life-threatening). A direct comparison of the subset of the 306 patients who received piperacillin-tazobactam or imipenem-cilastatin to the 77 who received cefepime or aztreonam found that the former group had increased GVHD-related mortality. Multivariate analysis confirmed that exposure to either cefepime or aztreonam was associated with reduced GVHD-related mortality after adjustment for the intensity of the conditioning regimen. Examination of the fecal microbiota found that, when compared to treatment with cefepime or aztreonam, piperacillin-tazobactam exposure was associated with a similar alteration in the abundance of Enterococcus, Akkermansia, and Erysipelotrichia, but with a significantly greater reduction in the numbers of Bacteroides and Lactobacillus.

Using a murine model, the investigators went on to demonstrate that imipenem-cilastatin administration was associated with loss of protective mucus, compromised intestinal barrier function, and increased pathological evidence of colonic GVHD. Finally, when compared to aztreonam, administration of piperacillin-tazobactam or of imipenem-cilastatin was associated with worsened GVHD mortality. Of note is that there was an increased abundance of fecal Akkermansia muciniphila, a commensal of mice and humans that degrades mucin, using it as a source of carbon and nitrogen.

Loss of anaerobic intestinal flora has long been associated with reduced resistance to colonization with potential pathogens, including ones with significant antibiotic resistance. The evidence discussed here points to another adverse effect of this loss—increased mortality related to intestinal GVHD in allo-HSCT recipients. The mechanism for this has been hypothesized to be due to the loss of anti-inflammatory properties of some anaerobes. Cefepime and, to a greater extent, aztreonam have limited antianaerobe activity, especially compared with piperacillin-tazobactam and imipenem-cilastatin. All these agents are frequently used in the management of febrile neutropenia, making overall restriction of antibiotic use problematic. Furthermore, elimination of 2 of the agents from routine use would have a marked impact on the heterogeneity of antibiotic use, an approach that appears to be optimal for slowing the emergence of resistance.

Allo-HSCT patients are enormously complex and, as a consequence, teasing out valid associations is complicated by the huge number of potentially confounding variables. In this case, however, murine experiments provide added weight to the investigators' clinical observations. In contrast, however some previous studies have come to what appears to be a conclusion opposite to that reached by Shono et al. Vossen and colleagues reported that gastrointestinal decontamination by oral administration of nonabsorbable antimicrobials (which in a proportion of cases included piperacillin-tazobactam) prevented acute GVHD in allo-HSCT recipients [3]. Furthermore, Shono and colleagues did not detect an association with the use of metronidazole, a drug with potent antianaerobe activity, and GVHD-related mortality. Thus, much sorting out of this complex clinical problem remains.

Transplant recipients are often excluded from antimicrobial stewardship efforts, but such exclusion is an enormous mistake. In addition to application of general stewardship principles (ie, optimization of antimicrobial therapy), we may need to consider the need for preferential use of microbiota-sparing antibiotic regimens, especially those that spare the anaerobic component.


Tropheryma whipplei Bacteremia in Febrile Patients in West Africa

Bassene H, Mediannikov O, Socolovschi C, et al. Tropheryma whipplei as a cause of epidemic fever, Senegal, 2010–2012. Emerg Infect Dis 2016; 22:1229–34.

Among the organisms recently identified in patients with fever of unknown origin in West Africa, particularly Senegal, are Borrelia crocidurae, Rickettsia felis, Rickettsia conorii, and Coxiella burnetii [1]. An additional organism, Tropheryma whipplei, was also unexpectedly identified in a prospective 2008–2009 survey of febrile patients in 2 Senegalese villages. In that study, Fenollar and colleagues identified T. whipplei DNA in the blood of 13 of the 204 (6.4%) subjects tested, with children more often being affected [1]. The most frequent associated symptom, in addition to fever, was cough. Tropheryma whipplei had previously been shown to be present in feces of approximately one-third of Senegalese (and almost three-fourths of children <2 years of age), but there was no correlation of stool carriage with the presence of bacteremia. These studies have now been extended by Bassene et al to include a larger number of subjects and a local control group.

Tropheryma whipplei DNA was detected by quantitative polymerase chain reaction (PRC) in blood samples from 36 of 786 (4.6%) febrile subjects and 1 of 385 (0.25%) healthy control subjects from 2 rural villages in Senegal. The single positive PCR from a control subject demonstrated a low concentration of the organism's DNA as determined by a cycle threshold that was higher than that of any febrile patient with a positive result. The distribution of positive results in the febrile subjects did not significantly vary by age. In addition to fever, the most frequently reported symptoms in the 36 febrile patients were headache (68.9%), cough (36.1%), rhinorrhea (22.2%), nausea (13.9%), vomiting (11.1%), and diarrhea (8.3%).

All positive results occurred during the June–October rainy season; 3 in 2011 and 33 in 2010. The occurrence of an outbreak was suggested by the fact that 28 of the positives in 2010 occurred in August when T. whipplei DNA was detected in 30% of 93 febrile patients. Familial clustering was identified in some cases.

Despite the high rates of stool and, to a lesser extent, saliva, asymptomatic carriage of T. whipplei, classic Whipple's disease is “virtually unknown” in Africa [1]. Stool carriage does not appear to be a predictor of bacteremia in febrile patients or in those who are asymptomatic. Marth and colleagues found that 0 of 18 asymptomatic French carriers had evidence of the organism in their bloodstream, suggesting that asymptomatic bacteremia is uncommon, a finding confirmed in the study reviewed here [2].

Tropheryma whipplei infection appears to be limited to humans [2]. The high frequency of fecal carriage in regions with limited access to toilet facilities suggests transmission by a fecal–oral route. These observations suggest that infection with this organism may be controlled by implementation of simple public health interventions.


| Table of Contents

Published on behalf of

cid cover HIV Medicine Association Logo

Society Members: For your free access to this journal, log in via the IDSA members area.

Impact Factor: 8.886

5-Yr impact factor: 9.206

For Reviewers

Looking for your next opportunity?

Looking for jobs...

For the Media