July 2019: ID diagnostics

CRP can help PCPs curb unnecessary antibiotic use in outpatients seeking care for acute COPD exacerbation.  In one study, over 20% of severe COPD exacerbations had noninfectious etiologies (16484677), yet in many practices nearly all patients presenting with acute COPD exacerbation receive an antibiotic prescription.  Over the past decade there have been a whole slew of papers about procalcitonin (PCT) establishing with reasonable certainty that PCT reliably identifies patients with suspected respiratory infections who are likely vs unlikely to benefit from antibiotics.  So then – could other biomarkers, like CRP, be used for the same ends?

This open-label RTC enrolled patients with COPD who presented to one of 86 clinics across England and Wales with acute exacerbation and at least one of the Anthonisen criteria for antibiotic prescription (increased dyspnea, increased sputum production, or increased sputum purulence).  They excluded anyone who needed hospital admission, had suspected pneumonia or infection at another site, were immunocompromised or pregnant, or had a past history of respiratory failure needing mechanical ventilation.  The researchers randomized patients to receive either usual care or prescribing guided by a point of care CRP; in the latter arm, clinicians received guidance that antibiotics were unlikely to benefit patients with a CRP <20, could be useful in those with a CRP 20-40, and were likely to be useful in patients with a CRP >40.  They assessed two outcomes: first, patient-reported antibiotic use in the 4 weeks following randomization, and second, patient-reported COPD health status using a standardized questionnaire.

Over a two year period, the authors enrolled a total 653 patients (n=325, usual care; n=324 CRP-guided care; 4 patients were excluded due to randomization error & withdrawal of consent).  Within the CRP-guided arm, 97% of patients received a CRP and the median value was 6mg/L (IQR 5-19); in total, 76% had a CRP value under 20 and only 12% had a value over 40mg/L.  Antibiotic use was 20% lower in the CRP-guided arm (57% vs 77%; aOR 0.31 with 95% CI 0.2-0.47) and the mean difference in the COPD health score was 0.19 points on a 6=point scale, favoring patients in the CRP group, with a 90% CI of -0.33 to -0.05 (for reference, the accepted least meaningful difference in scores for this questionnaire is 0.4).  In the six months following randomization, there were no important differences between the CRP-guided and usual care arms in the rates of hospitalization (9% vs 9%), diagnosis of pneumonia (3% vs 4%), or death (0% vs 1%), with all numerical counts favoring the CRP-guided arm.

So, in this RTC, even though clinicians’ adherence to the CRP-guided antibiotic recommendations was low (76% of patients in the CRP-guided arm had a CRP value <20mg/L, yet 57% received an antibiotic prescription anyway), antibiotic usage was curbed by 20% without any apparent patient harm.  CRP may be helpful as a stewardship tool to reassure outpatient clinicians & patients and give them the confidence not to start antibiotics in a low-risk setting. 31291514

Culturing periprosthetic tissue in blood culture bottles speeds detection of microorganisms in PJI. Previous studies have shown that use of blood culture bottle systems can improve the microbiologic yield of synovial fluid and sonication fluid from explanted prostheses.  This study examines the same question, but for periprosthetic tissue.  Over 11 months, the researchers took tissue samples submitted from suspected PJI of the hip, knee, elbow, ankle, and shoulder and processed them via routine microbiology lab protocol, then placed samples with excess tissue (158 samples from 62 patients) into aerobic and anaerobic BacTec blood culture bottles and incubated for up to 12 days.  The paper also included a set of “spiking” experiments, wherein native femoral head and surrounding joint tissue was inoculated with bacteria and then homogenized and placed in the blood culture bottles; this was done simply to demonstrate that the blood culture bottles could grow microorganisms in the presence of homogenized tissue/bone, which they did, so I won’t mention them further.

A quarter of the blood culture bottles with suspected PJI tissue turned positive – 44 aerobic and 36 anaerobic bottles – accounting for 29% of patients submitting clinical samples.  The majority of organisms were detected in both bottles (74%), followed by from the aerobic bottle only (22%) and from the anaerobic bottle only (4%).  All positive cultures resulted within 48 hours, and 89% of the positive aerobic and 97% of the positive anaerobic cultures did so within 24 hours.  Compared to blood culture bottles, culturing periprosthetic tissue via standard methods was similarly sensitive (44 versus 46 positives among 158 samples, sensitivity of 76% for standard workup vs 79% for blood culture bottles), but took much longer, with 5 days incubation to yield 90% of all positive cultures, and 8 days of incubation to yield 100% of positives. 

So, long story short – putting periprosthetic tissue into blood culture bottles doesn’t do much to improve the microbiologic yield over standard techniques, but it does lead to earlier identification of the recovered pathogens. 31291897

Bone biopsy in a patient with diabetic foot osteomyelitis: are you better off with a transcutaneous biopsy, or can you biopsy through the foot wound?  Transcutaneous bone biopsy is the gold standard for diagnosing osteomyelitis in DFI, but bone biopsy cultures obtained in wound care centers are often done through the foot wound.  This is potentially problematic because it provides an opportunity to collect microrganisms living peaceably in the wound on the way into the bone (and yes, the same is true of transcutaneous biopsies, but the intact skin is less densely populated with bacteria and more easily sanitized beforehand), and isolation of additional bystander organisms could lead to unnecessarily broad antimicrobial therapy.

Over a three year period, the authors prospectively enrolled patients seen at their diabetic foot care center who had positive probe-to-bone tests and both radiographic and clinical symptoms of active infection.  Consenting patients underwent simultaneous transcutaneous and through-wound bone biopsies; the latter was done following careful debridement of the wound and irrigation with iodine and then sterile saline.  Both samples were referred to the microbiology lab, and the researchers compared their yields.

The authors performed a total 46 paired bone biopsy cultures on 43 patients with confirmed diabetic foot osteomyelitis over the course of the study.  The patients had a mean age of 71; 63% had vascular disease involving the legs, and 24% had had a prior amputation.  Cultures were negative in 17% of transcutaneous bone biopsies but only 2% of through-wound bone biopsies (p=0.005).  Overall concordance of the positive culture results was poor (48%; totally different organisms were recovered from the two biopsies in 19% of cases, and the per-wound biopsies were missing one or more organisms cultured via transcutaneous biopsy in another 13% of cases). Agreement between the biopsy modalities varied by organism, with higher concordance when S. aureus was isolated (70%) and lower concordance for coagulase-negative staph (32%) and gram-negative bacilli (30%); the through-wound biopsies contained additional organisms in over a quarter of cases.

So, if you have choice, prefer transcutaneous to through-wound bone biopsy for DFI with osteomyelitis: not only does the latter technique more often recover likely colonizers, it fails to reliablely recover the presumptive genuine pathogens obtained from a transcutaneous biopsy. 30980264

When your GI pathogen multiplex PCR panel lights up with EPEC, maybe you shouldn’t get too excited.  There are now a bunch of these pathogen PCR panels in clinical use, and while they promised to revolutionize syndromic diagnosis of infection, the respiratory viral panel, the meningitis panel, and the GI panel all seem to add no data or muddy the waters as often as they bring clarity.  In this paper, the authors used a retrospective case-case study design to assess the clinical import of GI pathogen PCR panels that returned positive for enteropathogenic E.coli (EPEC) or Campylobacter sp.  Specifically, they compared those patients who were (1) PCR positive for EPEC only, (2) PCR positive for EPEC and any other pathogen, (3) PCR positive but culture negative for Campylobacter, and (4) both PCR and culture positive for Campylobacter.

A total 291 patients seen between 2013 and 2015 were included in the study, among whom 169 were PCR positive for Campylobacter and 122 were PCR positive for EPEC.  Of these, 27 had no documented diarrhea in the medical record – these cases represented 20% of the EPEC-only group, 5% of the EPEC co-infected group, and 5% of the Campylobacter group.  A non-infectious cause of diarrhea was identified in 73% of patients in whom only EPEC was detected on the PCR vs 25% of those with EPIC and a second pathogen and only 17% of those with Campylobacter detected.  Similarly, when patients with a positive PCR for Camplyobacter were stratified based on the result of stool cultures, the culture-negative patients were more likely to have ongoing antibiotic treatment (24% vs 3%), detection of a second pathogen (18% vs 2%), no diarrhea documented (9% vs 1%), and an identified non-infectious cause of the diarrhea (48% vs 4%).  When sequencing was used to identify the Campylobacter species in the culture-negative sample, two thirds were C.concisus, a colonizer of the oral cavity that seems to have a preference for colonizing the intestinal tracts of patients with IBD (reference: 19052183).

Take home points: (1) three out of four patients with isolated EPEC on GI pathoen PCR panel end up having a non-infectious cause of their diarrhea, and (2) PCR-positive but stool culture negative Campylobacter probably reflects colonization, not with C.jejuni but with a relative species that may be a marker for IBD. 31315581