Medical Director, Infection Prevention, El Camino Hospital, Palo Alto Medical Foundation
SOURCE: Lin MY, Stein BD, Kothadia SM, et al. Impact of mandatory infectious disease specialist approval on hospital-onset Clostridioides difficile infection rates and testing appropriateness. Clin Infect Dis 2023;77:346-350.
Reduction of hospital-onset Clostridioides difficile infection (HOCD) is a major goal of most hospital-acquired infection (HAI) prevention programs, which often involves the use of multiple evidence-based interventions, including the introduction of antimicrobial stewardship programs, enhanced environmental cleaning measures, systems for monitoring and providing feedback to the environmental services employees, the use of specific sporicidal cleaning products, the use of ultraviolet-C radiation for C. difficile infection (CDI) rooms, hand hygiene campaigns, and improved use of personal protective equipment. Our facility has gone one step further and provides polymerase chain reaction (PCR) peri-rectal screens for toxigenic CD of high-risk patients on admission, so asymptomatically colonized individuals can be pre-emptively isolated (and tested and treated more proactively).
Just to quickly review, hospital-onset CDI is defined as occurring > day 3 of hospitalization. Notably, this is not a 72-hour window, but it could be as little as 49 hours, depending on the time of admission, and it is based on the date of admission. Thus, delays in recognizing diarrhea, placing orders, and waiting to collect the next stool all lead to delays in the diagnosis. It is estimated that 20% to 40% of CDI cases are pushed from day 1-3 of hospitalization to day 4-7 based on these simple delays. If cases could only be recognized sooner and stools submitted to the laboratory a little sooner, many HOCD cases could be avoided.
Other authors have tried a different tack — by restricting access to testing beyond day 3 of hospitalization, fewer cases of HOCD are diagnosed. These authors performed a retrospective analysis of HOCD rates during a baseline period of no intervention, a second period with a computer-generated alert with decision support, and a third period with mandatory approval by an infectious disease (ID) specialist for any CD test ordered beyond day 3. The computer algorithm recommended soft stops on hospital day 4 or later if there was use of a laxative in the previous 24 hours, there were fewer than three loose stools in the preceding 24 hours or nursing staff charted hard or formed stool, or if a negative CD test had been obtained in the previous seven days. Test approval beyond day 4 of hospitalization was handled by two ID specialists from 8 a.m. to 5 p.m. seven days per week, which coincided with availability of testing in the laboratory. Testing was performed using a single-step nucleic acid amplification test (NAAT).
Introduction of the practice alert had a marginal effect on HOCD rates. And over time, the alerts generally were disregarded based on what the authors described as “alert fatigue.”
In contrast, mandatory approval by an ID specialist decreased the rate of HOCD by 60% over a 15-month period (P < 0.001), and “test appropriateness” increased. The test positivity rate increased from 11.2% at baseline, to 13.4% using the alerts, and to 16.2% with mandatory ID approval of testing.
The ID specialists handled a total of 608 requests for testing, of which 493 (81%) were approved. A total of 115 tests were denied for the following reasons: prior laxative use (43%), not meeting criteria for diarrhea (38%), clinical syndrome not consistent with CDI (12%), prior testing for CD within seven days (3%), or a previous positive CD test (3%). To assess the potential for harm, the first 102 consecutive patients for whom testing was not initially approved were observed during their hospital stay. Forty-five patients were subsequently tested, of whom six were positive, indicating a potential delay in diagnosis in 1% of all test requests. However, these figures do not include those physicians who were too busy or too discouraged to not pursue test approval.
The ID specialists spent a median of three minutes per telephone call (range, 1-15 minutes). (I do not know about you, but it takes me longer than three minutes to walk to the phone, dial the number, hoping to reach the caller on my first try, and discuss the details of a case). Suppose each phone request was 10 minutes per call for 653 phone calls, totaling 109 hours of ID specialist time over 15 months (7.3 hours per month), with about two interruptions per day. Basically, one ID specialist spent an entire day per month providing test approvals.
The issue with this and other attempts to restrict testing beyond day 4 is the obvious dichotomous approach to testing for days 1-3 vs. day 4 and beyond, which amounts to a manipulation of testing protocols to achieve a desired outcome. In fact, one could argue that in the absence of other interventions that might reduce the usual occurrence of CDI in these hospitals, they have successfully achieved a 60% under-diagnosis of HOCD. If in fact no harm occurred to such patients going forward — and I imagine a longer period of observation would be required to assess harm — one has to wonder if our usual approach to testing in hospital is uncovering “too many” patients with positive toxin tests. In other words, does a subset of toxin-positive minimally symptomatic patients exist who would recover without recognition of infection or specific treatment?
The use of PCR in this study also raises questions of over-diagnosis of patients diagnosed based on mere colonization. Many patients with colonization nonetheless can experience antibiotic-associated diarrhea without having acute CDI. Since CDI is defined by the presence of toxin production (enzyme immunoassay) and not toxigenic genes (PCR), these authors likely could have achieved a similar reduction in HOCD infection rates by switching to a toxin-based assay.
SOURCE: Ando N, Mizushima D, Omata K, et al. Combination of amoxicillin 3000 mg and probenecid versus 1500 mg amoxicillin monotherapy for treating syphilis in patients with human immunodeficiency virus: An open-label, randomized, controlled, non-inferiority trial. Clin Infect Dis 2023;77:779-787.
These authors at the AIDS Clinical Center in Tokyo performed a randomized trial of low-dose amoxicillin vs. higher-dose amoxicillin plus probenecid in the treatment of syphilis. A total of 115 participants with newly diagnosed syphilis were randomized in a 1:1 fashion to either amoxicillin 500 mg three times per day × 14 days for early syphilis or × 28 days for late syphilis vs. higher-dose amoxicillin 1,000 mg three times per day plus probenecid 250 mg three times per day × 14 days for early syphilis or × 28 days for late syphilis. Early syphilis was defined as primary, secondary, or early latent infection, whereas late syphilis was defined as late latent or syphilis of unknown duration. Serological “cure” was defined as a four-fold or greater decrease in rapid plasma reagin (RPR) titer or negative conversion at 12 months.
Of the 115 participants, 105 were human immunodeficiency virus-positive (HIV+) and on antiretroviral therapy, and their median CD4 count was 525 cells/microliter. One hundred participants (87%) had early syphilis and 15 (13%) had late syphilis. However, the median RPR was 1:64, which may be of concern when evaluating patients with HIV disease.
Overall serologic cure rates at 12 months were 90.6% for low-dose amoxicillin therapy vs. 94.4% for high-dose combination therapy (P = NS). Specifically, for those participants with early syphilis, low-dose amoxicillin performed well compared with combination therapy at 12 months, with serologic cure rates of 93.5% vs. 97.9%, respectively. However, at every time point of the investigation, serologic responses were lower for the lower dose of amoxicillin compared with the combination therapy. Specifically at three months post-treatment, response rates for early syphilis were 65.9% vs. 86.7% (P = 0.021), and overall cure rates at three months also were lower (62.7% vs. 78.8%, respectively). Is this difference at three months post-treatment clinically meaningful?
The number of persons with late syphilis in this study was small (n = 15), and serological responses at 12 months were achieved in only 71.4% for both the low-dose amoxicillin and the combination therapy groups. It is true that the longer a person has had syphilis, the longer duration of time to response, and some of these patients still may have achieved serologic response at 24 months. Actual failures, as opposed to serofast results, were not provided. Further, RPRs were not repeated at the time of treatment — and recent data found that 12% of patients can have a significant four-fold or greater increase in titer from the time of diagnosis to the point of treatment.
Low-dose amoxicillin was noninferior to higher-dose amoxicillin plus probenecid at 12 months post-treatment in this smaller study of largely HIV+ people. However, the predominance of HIV+ persons, coupled with the higher baseline RPR titers, argues for some circumspection as to the risk of central nervous system (CNS) involvement. It is not known how these results would extrapolate to non-HIV-infected persons.
