Column Editor
Weill Cornell Medicine
New York City
Executive Director
Surgical Infection Society
Foundation for Research and Education
Diagnostic Testing for Intraabdominal Infection
Childs DD, Lalwani N, Craven C, et al. A meta-analysis of the performance of ultrasound, hepatobiliary scintigraphy, CT and MRI in the diagnosis of acute cholecystitis. Abdom Radiol (NY). 2023 Nov 20. doi:10.1007/s00261-023-04059-w. Online ahead of print.
Arruzza E, Milanese S, Li LSK, et al. Diagnostic accuracy of computed tomography and ultrasound for the diagnosis of acute appendicitis: a systematic review and meta-analysis. Radiography (Lond). 2022;28(4):1127-1141. doi:10.1016/j.radi.2022.08.012.
Summary: Childs et al. evaluated the diagnostic accuracy of ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), and radionuclide cholescintigraphy (RC) for the diagnosis of acute cholecystitis (AC). Of 6,121 screened articles (2000-2021), data from 22 were extracted and scored with a methodological quality tool (Quality Assessment of Diagnostic Accuracy Studies [QUADAS]). Pooled estimates of sensitivity and specificity were calculated, and head-to-head comparisons (US vs. CT, US vs. RC) were made. The prevalence of AC varied widely (9%-98%). Sensitivity and specificity estimates were 0.69 (95% CI, 0.62-0.76) and 0.79 (CI, 0.71-0.86) for US, 0.91 (CI, 0.86-0.94) and 0.63 (CI, 0.51-0.74) for RS, 0.78 (CI, 0.69-0.84) and 0.81 (CI, 0.71-0.88) for CT, and 0.91 (CI, 0.78-0.97) and 0.93 (CI, 0.70-0.99) for MRI. Head-to-head, the sensitivity of CT (0.88; CI, 0.70-0.96) was significantly higher than US (0.67; CI, 0.43-0.84), but specificity (US, 0.82; CI, 0.54-0.95; CT, 0.92; CI 0.67-0.99) was similar. The sensitivity of RC (0.87; CI, 0.76-0.94) was significantly greater than US (0.62; CI, 0.44-0.77), but the specificity of US (0.82; CI, 0.65-0.92) was not higher than RC (0.68; CI, 0.47-0.84). CT may have a higher sensitivity than US for diagnosing AC, with similar specificity. RC is sensitive, but has low specificity.
Arruza et al. determined the accuracy of CT and US for suspected acute appendicitis (AA) in adults, adhering to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines for diagnostic test accuracy. Screening, full-text retrieval, data extraction, and methodological quality assessment (QUADAS) were performed. Meta-analyses were performed for relevant subgroups, and sensitivity analysis evaluated for outliers. Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) methodology assessed the certainty of findings. Studies evaluating CT (n=31), US (n=10), and both (n=6) were included. Pooled sensitivity and specificity for CT were 0.97 (CI, 0.96-0.98) and 0.96 (CI, 0.94-0.97) respectively, and 0.82 (CI,0.74-0.88) and 0.86 (CI, 0.73-0.93) for US, respectively. When stratified by mode of contrast enhancement, sensitivity and specificity were highest for CT with intravenous (IV) and oral contrast (0.99 [CI, 0.97-0.99], 0.97 [CI, 0.94-0.99]), significantly higher than CT with only IV contrast (0.96 [CI, 0.92-0.97], 0.94 [CI, 0.92-0.96]). Low-dose CT produced comparable values (0.94 [CI, 0.89-0.96], 0.94 [CI, 0.91-0.96]) relative to these subgroups and standard dose non-contrast CT (0.88 [CI, 0.77-0.94], 0.91 [CI, 0.83-0.96]). The authors recommended CT with IV plus oral contrast for diagnosis of adult AA; either standard- or low-dose CT is appropriate.
Commentary: Have you heard of “diagnostic stewardship”? The terminology, derived conceptually from antibiotic stewardship, is new but the concept is not: Order only the most appropriate tests for the diagnosis being sought, considering test performance, morbidity, and cost, and do not over-test. The data above are not surprising, except perhaps that CT may supplant US for the diagnosis of AC based on accuracy (there is no cost analysis included in the data). Rectal contrast for the diagnosis of AA seems forgotten, despite being how the CT diagnosis of AA was described initially. More importantly, sufficient data have accumulated such that institutions and payors may become even more prescriptive regarding test ordering, rendering clinical judgment and individualized care increasingly irrelevant despite the trend toward personalized medicine.
Nasal Decolonization to Prevent ICU Infections
Huang SS, Septimus EJ, Kleinman K, et al. Nasal iodophor antiseptic vs nasal mupirocin antibiotic in the setting of chlorhexidine bathing to prevent infections in adult ICUs: a randomized clinical trial. JAMA. 2023;330(14):1337-1347. doi:10.1001/jama.2023.17219.
Summary: To compare topical iodophor and mupirocin for universal ICU nasal decolonization in combination with chlorhexidine (CHG) bathing, a pragmatic, cluster-randomized noninferiority trial was conducted in U.S. community hospitals, all of which used mupirocin-CHG for universal decolonization in ICUs at baseline. ICU-attributable Staphylococcus aureus clinical cultures (primary outcome), methicillin-resistant S. aureus (MRSA) cultures, and all-cause blood stream infections were evaluated using proportional hazard models. The prespecified noninferiority margin for the primary outcome was 10%. Adult ICU patients in 137 randomized hospitals were included during baseline (May 2015-April 2017) and intervention periods (November 2017-April 2019).
Among 801,668 patients admitted to 233 ICUs, participants’ mean age was 63 years and 46% were female. The mean ICU length of stay was 4.8 days. S. aureus clinical isolates in the intervention versus baseline periods were 5.0 versus 4.3/1,000 ICU days (hazard ratio [HR], 1.17) for iodophor-CHG compared with mupirocin-CHG’s 4.1 versus 4.0/1,000 ICU days (HR, 0.99). The HR difference in differences was 18.4% lower (95% CI, 10.7%-26.6%) for mupirocin-CHG (P<0.001). For MRSA clinical cultures, HRs were 1.13 for iodophor-CHG (2.3 vs. 2.1/1,000 ICU days) and 0.99 for mupirocin-CHG (2.0 vs. 2.0/1,000 ICU days). The HR difference in differences was 14.1% lower (95% CI, 3.7%-25.5%) for mupirocin-CHG (P=0.007). For all-pathogen blood stream infections, HRs were nearly identical (difference in differences, –0.9% [95% CI, –9.0% to 8.0%]; P=0.84).
Commentary: Universal nasal mupirocin plus CHG bathing in ICUs has been demonstrated to prevent MRSA infections and all-cause blood stream infections. Emerging antibiotic resistance to mupirocin has raised the question of whether an antiseptic rather than an antibacterial agent could be advantageous for ICU decolonization. Nasal iodophor antiseptic combined with CHG bathing did not meet pre-specified noninferiority criteria compared with nasal mupirocin for the outcome of reducing S. aureus clinical cultures in adult ICU patients. In actuality, the results were consistent with nasal iodophor being inferior to nasal mupirocin.
Empiric Antibiotic Therapy for Acute Infection and Risk of Acute Kidney Injury
Qian ET, Casey JD, Wright A, et al. Cefepime vs. piperacillin-tazobactam in adults hospitalized with acute infection: the ACORN randomized clinical trial. JAMA. 2023;330(16):1557-1567. doi: 10.1001/jama.2023.20583.
Tong SYC, Venkatesh B, McCreary EK. Acute kidney injury with empirical antibiotics for sepsis. JAMA. 2023;330(16):1531-1533. doi: 10.1001/jama.2023.18591.
Alosaimy S, Rybak MJ, Sakoulas G. Understanding vancomycin nephrotoxicity augmented by ss-lactams: a synthesis of endosymbiosis, proximal renal tubule mitochondrial function, and ss-lactam chemistry. Lancet Infect Dis 2023:S1473-3099(23)00432-2. doi:10.1016/S1473-3099(23)00432-2. Online ahead of print.
Summary: Piperacillin-tazobactam (P-T) has been hypothesized to cause acute kidney injury (AKI) and cefepime (C) has been hypothesized to cause neurological dysfunction, but their comparative safety has not been evaluated in a randomized trial. The ACORN (Antibiotic Choice On ReNal outcomes) trial randomized adults (1:1 to C vs. P-T) who were started on an anti-pseudomonal antibiotic within 12 hours of presentation at a single U.S. academic medical center. The primary outcome was the highest stage of AKI or death by day 14, measured on a 5-level ordinal scale. Secondary outcomes included the incidence of major adverse kidney events at day 14 and the number of days alive and delirium/coma-free during the study. A total of 2,511 patients included in the primary analysis had a median age of 58 years; 42.7% were female; 94.7% were enrolled in the emergency department; and 77.2% also received vancomycin at enrollment. The highest stage of AKI or death was not different; there were 85 patients (n=1,214; 7.0%) in the C group with stage 3 AKI and 92 (7.6%) who died versus 97 patients (n=1,297; 7.5%) in the P-T group with stage 3 AKI and 78 (6.0%) who died (odds ratio [OR], 0.95; CI, 0.80-1.13; P=0.56). Major adverse kidney events at day 14 did not differ between groups. Patients in the C group experienced fewer days alive and delirium/coma-free by 14 days (OR, 0.79; CI, 0.65-0.95), but the median difference was a mere 0.3 day.
Commentary: C and P-T are administered commonly to hospitalized adults for empiric treatment of infection. Both are associated with risk, in particular P-T. The purported increased OR of AKI from the vancomycin/P-T combination (based on meta-analysis of observational studies) is 3.6, with an absolute incidence of AKI of nearly 23% compared with 12% for comparators, as noted in the accompanying editorial cited above. Based thereon, the Food and Drug Administration has warned that vancomycin/P-T co-administration may increase AKI risk, despite the questions raised by observational studies, such as reliance on increased serum creatinine concentration (which may inhibit tubular reabsorption of creatinine) rather than estimates of kidney function, and unmeasured confounders. Alternatively, hydrophobic b-lactams such as P-T may have increased uptake into renal tubular cells compared with hydrophilic agents (most cephalosporins and carbapenems), increasing oxidative stress on the tubules. There are numerous limitations to this study, including that it was a single-center trial. The dose of P-T was lower than is usually recommended, and C was given by rapid IV push rather than infusion, which can enhance toxicity. The duration of therapy was brief (median, 3 days) and nearly 50% of enrollees did not have sepsis. There was substantial crossover in antibiotic received (C, 19%; P-T, 17%), and thus the intervention groups are not “clean.” The question of nephrotoxicity from vancomycin/P-T co-administration is still unsettled, but kudos to the investigators for trying. Clinicians are advised to consider less nephrotoxic antibiotic regimens, particularly for at-risk patients (e.g., septic shock, chronic kidney disease).
Carbapenems for Empiric Therapy of Sepsis
Umemura Y, Yamakawa K, Tanaka Y, et al. Efficacy of carbapenems compared with non-carbapenem broad-spectrum beta-lactam antibiotics as initial antibiotic therapy against sepsis: a nationwide observational study. Crit Care Med. 2023;51(9):1210-1221. doi:10.1097/CCM.0000000000005932.
Summary: To reduce indiscriminate carbapenem use, this observational study conducted in Japanese tertiary hospitals evaluated the survival effect associated with empiric carbapenem therapy for adult sepsis compared with alternative regimens such as P-T or a fourth-generation cephalosporin. Patients were stratified by carbapenem versus non-carbapenem broad-spectrum beta-lactam agent as initial treatment. In-hospital mortality was estimated overall and in several subgroups by logistic regression adjusted by weighting using propensity scores. Among 7,392 patients with sepsis, 3,547 (48%) received a carbapenem. There was no association between carbapenem therapy and lower mortality (adjusted OR, 0.88; P=0.108). Subgroup analyses suggested survival benefits associated with carbapenem therapy in patients with septic shock, in ICUs, or on mechanical ventilation (P for effect: <0.001, 0.014, and 0.105, respectively).
Commentary: Empiric carbapenem therapy for sepsis was not associated with lower mortality, except in sick patients, essentially. However, the prevalence of extended-spectrum beta-lactamase (ESBL)-producing Enterobacterales is relatively low in Japan. Even though retrospective, these findings are consistent with other recent observations which suggest that early (within 1 hour) empiric therapy for sepsis is non-beneficial compared with source control, specimen collection and analysis, and directed therapy for surgical patients who are not in shock. Carbapenem sparing is one of the main tenets of antimicrobial stewardship programs, of which all surgeons should take heed.
This article is from the February 2024 print issue.
Please log in to post a comment