Emmanuelle Gras

Les infections fongiques invasives (IFI) constituent une complication majeure en oncohématologie, responsable d'une morbi-mortalité élevée malgré les avancées récentes. Leur diagnostic reste complexe, en raison de signes cliniques peu spécifiques et des prélèvements mycologiques difficiles, entraînant un recours fréquent aux traitements antifongiques empiriques, au risque de surtraitement. L'identification précoce des patients les plus à risque apparaît essentielle pour optimiser leur prise en charge. Cette thèse de sciences avait un double objectif : (1) réaliser une revue systématique avec méta-analyse d'études observationnelles (cohortes, cas-témoins) afin de synthétiser les facteurs de risque d'IFI chez l'adulte atteint d'hémopathie maligne et/ou greffé de cellules souches hématopoïétiques, et (2) développer un modèle prédictif dédié aux infections à champignons filamenteux. La revue systématique, incluant 121 études (près de 200 000 patients), a identifié des facteurs de risque : hémopathie réfractaire ou en rechute, greffe allogénique à haut risque, maladie du greffon contre l'hôte, neutropénie, corticothérapie, lymphopénie et infection à CMV. La prophylaxie antifongique démontrait un effet protecteur. À partir d'une cohorte monocentrique de 977 patients (1820 admissions, 43 IFI prouvées/probables), un modèle prédictif a été développé par apprentissage automatique supervisé (Gradient Boosting) et validé par validation croisée. Le modèle optimal incluait 8 variables (dont âge, données de transfusions, durée de lymphopénie, paramètres vitaux, corticothérapie, délai depuis dernière admission). Ces travaux proposent ainsi un outil d'aide à la décision individualisé, visant à affiner le dépistage et la stratégie thérapeutique des IFI à champignons filamenteux en hématologie adulte.

Improving prevention and treatment strategies for invasive fungal infections (IFI) is essential to reduce associated morbidity and mortality. We performed a systematic review and meta-analyses to assess factors associated with IFI in adult patients with hematological malignancies and/or hematopoietic stem cell transplantation (HSCT). Data sources. We searched PubMed, Embase, CENTRAL and the grey literature from 01/01/2002 to 08/02/2024. Study eligibility criteria. Eligible studies were case-control or cohort studies including adult patients with hematological malignancies and/or HSCT and reporting risk factors for IFI. Participants. Adult patients with hematological malignancies and/or HSCT. Assessment of risk of bias. Risk of bias assessment was assessed independently using the Critical Appraisal Skills Programme checklists for cohort studies and case-control studies. Methods of data synthesis. Study selection and data extraction were done independently. Adjusted estimates were pooled using random effects models. Among 12 624 references identified, 69 studies (reporting 2917 IFI) were included in the systematic review, 20 of which were included in meta-analyses. Factors independently associated with IFI included previous allo-HSCT (pooled aHR 3.21 [95%CI 1.54–6.71]), especially with a haploidentical donor (pooled aHR 2.41 [95%CI 1.27–4.57]), acute graft vs. host disease (aGvHD) ≥ 2 (pooled aHR 2.59 [95%CI 1.36–4.90]), corticosteroids (pooled aOR 2.84 [95%CI 1.42–5.70]) or T-cell depleting agents (pooled aOR 2.73 [95%CI 1.61–4.64]). Antifungal prophylaxis was a protective factor for IFI (pooled aOR 0.20 [95%CI 0.13–0.28]). The identification of factors independently associated with IFI may help to stratify IFI risk among hematological patients. Registration: PROSPERO, CRD42023429103.

Objectives: To describe the management of haematological patients experiencing prolonged SARS-CoV-2 viral shedding, as the optimal management strategy for this condition remains undetermined. Methods: We conducted a retrospective evaluation of our prospectively followed cohort of haematological patients treated with remdesivir for more than 10 days. Starting January 2023, upon COVID-19 diagnosis, the treatment strategy was based on symptoms and PCR cycle threshold (Ct) as follows: (i) when Ct was 25 or less or if the patient had symptoms, a course of remdesivir for at least 10 days, nirmatrelvir/ritonavir for 5 days (whenever possible) and convalescent plasma was administered; and (ii) when the patient was asymptomatic and had a PCR Ct of more than 25, when possible, a course of 5 days of nirmatrelvir/ritonavir was administered. The patient was considered to have achieved viral clearance and, thus, remdesivir was stopped, in either of these cases: (i) PCR negativity, or (ii) subgenomic RNA negativity. Results: From January to November 2023, 18 patients benefited from a safe extended remdesivir administration, resulting in detection of SARS-CoV-2 viral clearance in a median time of 3.5 weeks (IQR 2.6–3.9) (min–max 1.6–8.0). No clinical or biological side effects were detected. No patient died or needed further treatment for their COVID-19 episode. Conclusions: The extended course of remdesivir, combined with other active therapies for COVID-19 infection, was well tolerated. Cure and virus negativity were obtained in all these high-risk patients.

Artificial intelligence (AI) and machine learning (ML) have the potential to revolutionize the management of febrile neutropenia (FN) and drive progress toward personalized medicine. In this review, we detail how the collection of a large number of high-quality data can be used to conduct precise mathematical studies with ML and AI. We explain the foundations of these techniques, covering the fundamentals of supervised and unsupervised learning, as well as the most important challenges, e.g. data quality, 'black box' model interpretation and overfitting. To conclude, we provide detailed examples of how AI and ML have been used to enhance predictions of chemotherapy-induced FN, detection of bloodstream infections (BSIs) and multidrug-resistant (MDR) bacteria, and anticipation of severe complications and mortality. There is promising potential of implementing accurate AI and ML models whilst managing FN. However, their integration as viable clinical tools poses challenges, including technical and implementation barriers. Improving global accessibility, fostering interdisciplinary collaboration, and addressing ethical and security considerations are essential. By overcoming these challenges, we could transform personalized care for patients with FN.

Purpose Major bleedings have been described with cefazolin. The objective was to determine the frequency of bleeding events in cefazolin-treated patients and to identify risk factors for these complications. Methods Monocenter prospective observational study of all consecutive cefazolin-treated patients. Patients benefited from a daily clinical assessment of bleedings and a twice-a-week blood sampling including hemostasis. Bleedings were classified according to the International Society on Thrombosis and Hemostasis classification: major, clinically relevant non-major bleedings (CRNMB) and minor bleedings. Results From September 2019 to July 2020, 120 patients were included, with a mean age of 59.4 (± 20.7) years; 70% of them (84/120) were men. At least 1 CRNMB or major bleeding were observed in 10% of the patients (12/120). Compared to patients with no or minor bleeding, patients with CRNMB or major bleeding were, upon start of cefazolin, more frequently hospitalized in an intensive care unit (7/12, 58.3%, vs. 12/108, 11.1%, P < 0.001, respectively) and receiving vitamin K antagonists (4/12, 33.3%, vs. 8/108, 7.4%, P = 0.019, respectively). After multivariate analysis, patients receiving vitamin K antagonists the day prior bleeding and/or treated for endocarditis were factors associated with an increased risk of CRNMB or major bleeding (odd ratio 1.36, confidence interval 95%, 1.06–1.76, P = 0.020 and 1.30, 1.06–1.61, P = 0.015, respectively). Conclusions Bleeding events associated with cefazolin treatment are frequent. Close clinical monitoring should be performed for patients treated for endocarditis and/or receiving vitamin K antagonists. Hemostasis work-up could be restricted to these patients.

Aims: To assess the incidence density of local complications of peripheral venous catheters in patients aged 70 years and older, to identify risk factors for local complications of peripheral venous catheters, to describe microbiological epidemiology and to assess the impact of complications on patient outcomes.Design: Prospective, observational, single‐centre study.Methods: Patients 70 years and older admitted to the geriatric department of a teaching hospital in France between December 2019 and May 2020 were considered for inclusion if they had a peripheral venous catheter during their stay. Nurses checked the catheter insertion site three times a day for local complications; physicians ensured the follow‐up of complications. The STROBE checklist was used in this prospective observational study.Results: A total of 322 patients were included, with 849 peripheral venous catheters; the median age was 88 years and 182 (56.5%) were women. The incidence density of local complications was 50.5/1000 peripheral venous catheter‐days. Risk factors for local complications on multivariate analysis were dressing replacement (OR 1.18), furosemide (OR 1.11) and vancomycin (OR 1.60) infusion, urinary continence (OR 1.09) and hematoma at the catheter insertion site (OR 1.15). Thirteen cellulitis and three abscesses were diagnosed. Occurrence of a local complication was associated with a 3‐day increased duration of hospital stay (17 vs. 14 days).Conclusion: Risk factors for peripheral venous catheter local complications include urinary continence, furosemide or vancomycin infusion, hematoma at the peripheral venous catheter insertion site or dressing replacement.Implication for the Patient Care: Closer clinical monitoring may help reduce the occurrence of local peripheral venous catheters complication in patients 70 years and older.Relevance to Clinical Practice: Patients at greater risk of peripheral venous catheter local complications deserve closer clinical monitoring or improved preventive measures, which may be beneficial to reduce the length of hospital stay.No Patient or Public Contribution: The study was designed to describe risk factors for local complications of peripheral venous catheters in order to reinforce surveillance in this specific population by nurses and medical staffs. Patients had their peripheral venous catheter insertion site checked thrice a day by the nurse in charge as part of usual care. They, as service users, caregivers or members of the public, were not solicited for data collection, analysis, interpretation or preparation of the manuscript.