Do machine learning methods lead to similar individualized treatment rules? A comparison study on real data.

Identifying patients who benefit from a treatment is a key aspect of personalized medicine, which allows the development of individualized treatment rules (ITRs). Many machine learning methods have been proposed to create such rules. However, to what extent the methods lead to similar ITRs, that is, recommending the same treatment for the same individuals is unclear. In this work, we compared 22 of the most common approaches in two randomized control trials. Two classes of methods can be distinguished. The first class of methods relies on predicting individualized treatment effects from which an ITR is derived by recommending the treatment evaluated to the individuals with a predicted benefit. In the second class, methods directly estimate the ITR without estimating individualized treatment effects. For each trial, the performance of ITRs was assessed by various metrics, and the pairwise agreement between all ITRs was also calculated. Results showed that the ITRs obtained via the different methods generally had considerable disagreements regarding the patients to be treated. A better concordance was found among akin methods. Overall, when evaluating the performance of ITRs in a validation sample, all methods produced ITRs with limited performance, suggesting a high potential for optimism. For non-parametric methods, this optimism was likely due to overfitting. The different methods do not lead to similar ITRs and are therefore not interchangeable. The choice of the method strongly influences for which patients a certain treatment is recommended, drawing some concerns about their practical use.

Estimating individualized treatment effects using an individual participant data meta-analysis.

BACKGROUND: One key aspect of personalized medicine is to identify individuals who benefit from an intervention. Some approaches have been developed to estimate individualized treatment effects (ITE) with a single randomized control trial (RCT) or observational data, but they are often underpowered for the ITE estimation. Using individual participant data meta-analyses (IPD-MA) might solve this problem. Few studies have investigated how to develop risk prediction models with IPD-MA, and it remains unclear how to combine those methods with approaches used for ITE estimation. In this article, we compared different approaches using both simulated and real data with binary and time-to-event outcomes to estimate the individualized treatment effects from an IPD-MA in a one-stage approach. METHODS: We compared five one-stage models: naive model (NA), random intercept (RI), stratified intercept (SI), rank-1 (R1), and fully stratified (FS), built with two different strategies, the S-learner and the T-learner constructed with a Monte Carlo simulation study in which we explored different scenarios with a binary or a time-to-event outcome. To evaluate the performance of the models, we used the c-statistic for benefit, the calibration of predictions, and the mean squared error. The different models were also used on the INDANA IPD-MA, comparing an anti-hypertensive treatment to no treatment or placebo ( N = 40 237 , 836 events). RESULTS: Simulation results showed that using the S-learner led to better ITE estimation performances for both binary and time-to-event outcomes. None of the risk models stand out and had significantly better results. For the INDANA dataset with a binary outcome, the naive and the random intercept models had the best performances. CONCLUSIONS: For the choice of the strategy, using interactions with treatment (the S-learner) is preferable. For the choice of the method, no approach is better than the other.

Personalizing renal replacement therapy initiation in the intensive care unit: a reinforcement learning-based strategy with external validation on the AKIKI randomized controlled trials.

OBJECTIVE: The timely initiation of renal replacement therapy (RRT) for acute kidney injury (AKI) requires sequential decision-making tailored to individuals' evolving characteristics. To learn and validate optimal strategies for RRT initiation, we used reinforcement learning on clinical data from routine care and randomized controlled trials. MATERIALS AND METHODS: We used the MIMIC-III database for development and AKIKI trials for validation. Participants were adult ICU patients with severe AKI receiving mechanical ventilation or catecholamine infusion. We used a doubly robust estimator to learn when to start RRT after the occurrence of severe AKI for three days in a row. We developed a "crude strategy" maximizing the population-level hospital-free days at day 60 (HFD60) and a "stringent strategy" recommending RRT when there is significant evidence of benefit for an individual. For validation, we evaluated the causal effects of implementing our learned strategies versus following current best practices on HFD60. RESULTS: We included 3748 patients in the development set and 1068 in the validation set. Through external validation, the crude and stringent strategies yielded an average difference of 13.7 [95% CI -5.3 to 35.7] and 14.9 [95% CI -3.2 to 39.2] HFD60, respectively, compared to current best practices. The stringent strategy led to initiating RRT within 3 days in 14% of patients versus 38% under best practices. DISCUSSION: Implementing our strategies could improve the average number of days that ICU patients spend alive and outside the hospital while sparing RRT for many. CONCLUSION: We developed and validated a practical and interpretable dynamic decision support system for RRT initiation in the ICU.

Protective effect of height on long-term survival of resectable lung cancer: a new feature of the lung cancer paradox.

INTRODUCTION: Unlike most malignancies, higher body mass index (BMI) is associated with a reduced risk of lung cancer and improved prognosis after surgery. However, it remains controversial whether height, one of determinants of BMI, is associated with survival independently of BMI and other confounders. METHODS: We extracted data on all consecutive patients with resectable non-small cell lung cancer included in Epithor, the French Society of Thoracic and Cardiovascular Surgery database, over a 16-year period. Height was analysed as a continuous variable, and then categorised into four or three categories, according to sex-specific quantiles. Cox proportional hazards regression was used to estimate the association of height with survival, adjusted for age, tobacco consumption, forced expiratory volume in one second (FEV1), WHO performance status (WHO PS), American Society of Anesthesiologists (ASA) score, extent of resection, histological type, stage of disease and centre as a random effect, as well as BMI in a further analysis. RESULTS: The study included 61 379 patients. Higher height was significantly associated with better long-term survival after adjustment for other variables (adjusted HR 0.97 per 10 cm higher height, 95% CI 0.95 to 0.99); additional adjustment for BMI resulted in an identical HR. The prognostic impact of height was further confirmed by stratifying by age, ASA class, WHO PS and histological type. When stratifying by BMI class, there was no evidence of a differential association (p=0.93). When stratifying by stage of disease, the prognostic significance of height was maintained for all stages except IIIB-IV. CONCLUSIONS: Our study shows that height is an independent prognostic factor of resectable lung cancer.

Immunomodulators for immunocompromised patients hospitalized for COVID-19: a meta-analysis of randomized controlled trials.

Background: Although immunomodulators have established benefit against the new coronavirus disease (COVID-19) in general, it is uncertain whether such agents improve outcomes without increasing the risk of secondary infections in the specific subgroup of previously immunocompromised patients. We assessed the effect of immunomodulators on outcomes of immunocompromised patients hospitalized for COVID-19. Methods: The protocol was prospectively registered with PROSPERO (CRD42022335397). MEDLINE, Cochrane Central Register of Controlled Trials and references of relevant articles were searched up to 01-06-2022. Authors of potentially eligible randomized controlled trials were contacted to provide data on immunocompromised patients randomized to immunomodulators vs control (i.e., placebo or standard-of-care). Findings: Eleven randomized controlled trials involving 397 immunocompromised patients hospitalized for COVID-19 were included. Ten trials had low risk of bias. There was no difference between immunocompromised patients randomized to immunomodulators vs control regarding mortality [30/182 (16.5%) vs 41/215 (19.1%); RR 0.93, 95% CI 0.61-1.41; p = 0.74], secondary infections (RR 1.00, 95% CI 0.64-1.58; p = 0.99) and change in World Health Organization ordinal scale from baseline to day 15 (weighed mean difference 0.27, 95% CI -0.09-0.63; p = 0.15). In subgroup analyses including only patients with hematologic malignancy, only trials with low risk of bias, only trials administering IL-6 inhibitors, or only trials administering immunosuppressants, there was no difference between comparators regarding mortality. Interpretation: Immunomodulators, compared to control, were not associated with harmful or beneficial outcomes, including mortality, secondary infections, and change in ordinal scale, when administered to immunocompromised patients hospitalized for COVID-19. Funding: Hellenic Foundation for Research and Innovation.

Prognostic value of automated assessment of interstitial lung disease on CT in systemic sclerosis.

OBJECTIVE: Stratifying the risk of death in SSc-related interstitial lung disease (SSc-ILD) is a challenging issue. The extent of lung fibrosis on high-resolution CT (HRCT) is often assessed by a visual semiquantitative method that lacks reliability. We aimed to assess the potential prognostic value of a deep-learning-based algorithm enabling automated quantification of ILD on HRCT in patients with SSc. METHODS: We correlated the extent of ILD with the occurrence of death during follow-up, and evaluated the additional value of ILD extent in predicting death based on a prognostic model including well-known risk factors in SSc. RESULTS: We included 318 patients with SSc, among whom 196 had ILD; the median follow-up was 94 months (interquartile range 73-111). The mortality rate was 1.6% at 2 years and 26.3% at 10 years. For each 1% increase in the baseline ILD extent (up to 30% of the lung), the risk of death at 10 years was increased by 4% (hazard ratio 1.04, 95% CI 1.01, 1.07, P = 0.004). We constructed a risk prediction model that showed good discrimination for 10-year mortality (c index 0.789). Adding the automated quantification of ILD significantly improved the model for 10-year survival prediction (P = 0.007). Its discrimination was only marginally improved, but it improved prediction of 2-year mortality (difference in time-dependent area under the curve 0.043, 95% CI 0.002, 0.084, P = 0.040). CONCLUSION: The deep-learning-based, computer-aided quantification of ILD extent on HRCT provides an effective tool for risk stratification in SSc. It might help identify patients at short-term risk of death.

Risk factors for thromboembolic events in patients hospitalized for COVID-19 pneumonia in a general ward and requiring treatment with oxygen.

PURPOSE: To assess risk factors for arterial and venous thromboses (AVT) in patients hospitalized in general wards for COVID-19 pneumonia and requiring oxygen therapy. METHODS: Our study was based on three randomized studies conducted as part of the CORIMUNO-19 platform in France between 27 March and 26 April 2020. Adult inpatients with COVID-19 pneumonia requiring at least 3 l/min of oxygen but not ventilation were randomized to receive standard care alone or standard care plus biologics. Patients were followed up for 3 months, and adverse events were documented. Risk factor for AVT and bleeding was identified by analyzing clinical, laboratory, and treatment data at baseline among the 315 patients with complete datasets. A Fine and Gray model was used to take account of competing events. RESULTS: During the 3-month follow-up period, 39 AVT occurred in 38 (10%) of the 388 patients: 26 deep vein thromboses and/or pulmonary embolisms in 25 (6%) patients, and 14 arterial thrombotic events in 13 (3%) patients. A history of diabetes at inclusion [sHR (95% CI) = 2.65 (1.19-5.91), P = .017] and the C-reactive protein (CRP) level (sHR = 1 [1-1.01], P = .049) were significantly associated with an elevated risk of thrombosis. Obesity was not associated with a higher risk of thrombosis (sHR = 1.01 [0.4-2.57], P = .98). The CRP level and diabetes were not risk factors for hemorrhage. CONCLUSION: Among patients hospitalized in general wards for COVID-19 pneumonia during the first wave of the epidemic, diabetes (but not obesity) and a high CRP level were risk factors for AVT. The use of higher doses of anticoagulant in these high-risk patients could be considered.

Rituximab as maintenance therapy for ANCA-associated vasculitides: pooled analysis and long-term outcome of 277 patients included in the MAINRITSAN trials.

OBJECTIVE: To compare the long-term efficacy and safety of azathioprine (AZA), 18-month fixed-schedule rituximab (RTX), 18-month tailored RTX and 36-month RTX in preventing relapses in patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis who achieved a complete remission after induction therapy. Patients treated with 36-month RTX received either a fixed or a tailored regimen for the first 18 months and a fixed regimen for the last 18 months (36-month fixed/fixed RTX and 36-month tailored/fixed RTX, respectively). METHODS: The Maintenance of Remission using Rituximab in Systemic ANCA-associated Vasculitis (MAINRITSAN) trials sequentially compared: 18-month fixed-schedule RTX versus AZA (MAINRITSAN); 18-month fixed-schedule RTX versus 18-month tailored-RTX (MAINRITSAN2); and extended therapy to 36 months with four additional RTX infusions after MAINRITSAN2 versus placebo (MAINRITSAN3). Patients were then followed prospectively through month 84 and their data were pooled to analyse relapses and adverse events. The primary endpoint was relapse-free survival at month 84. RESULTS: 277 patients were enrolled and divided in 5 groups: AZA (n=58), 18-month fixed-schedule RTX (n=97), 18-month tailored-RTX (n=40), 36-month tailored/fixed RTX (n=42), 36-month fixed/fixed RTX (n=41). After adjustment for prognostic factors, 18-month fixed-schedule RTX was superior to AZA in preventing major relapses at month 84 (HR 0.38, 95% CI 0.20 to 0.71). The 18-month tailored-RTX regimen was associated with an increased risk of major relapse compared with fixed-schedule regimen (HR 2.92, 95% CI 1.43 to 5.96). The risk of major relapse was similar between 36-month fixed/fixed and 18-month fixed-RTX (HR 0.69, 95% CI 0.38 to 1.25). CONCLUSIONS: According to these results, it appears that the 84-month remission rate is higher with an 18-month fixed RTX regimen compared with AZA and 18-month tailored RTX. Also, extending RTX to 36 months does not appear to reduce the long-term relapse rate compared with the 18-month fixed RTX regimen. However, as this study was underpowered to make this comparison, further prospective studies are needed to determine the potential long-term benefits of extending treatment in these patients.

Association of Lack of Fear of Dying With New Organ Failure: Results of a Multicenter Prospective Cohort Study.

BACKGROUND: Patients' anxiety on intensive care unit (ICU) admission is associated with subsequent deterioration. OBJECTIVE: To assess whether patients' fears/anxiety are predictive of new organ failure within 7 days of ICU admission. METHODS: In a prospective 3-center cohort study of non-comatose patients without delirium or invasive mechanical ventilation, 9 specific fears were evaluated through yes/no questions. Illness severity was assessed using the Simplified Acute Physiology Score II (SAPS II) and the Sequential Organ Failure Assessment (SOFA). Intensity of acute and chronic anxiety was assessed with the state and trait components of the State-Trait Anxiety Inventory (STAI). Patients were followed up for 7 days. RESULTS: From April 2014 to December 2017, 373 patients (median [IQR] age, 63 [48-74] years; 152 [40.8%] women; median (IQR) SAPS II, 27 [19-37]) were included. Feelings of vulnerability and fear of dying were reported by 203 (54.4%) and 172 (46.1%) patients, respectively. The STAI-State score was 40 or greater in 192 patients (51.5%). Ninety-four patients (25.2%) had new organ failure. Feelings of vulnerability (odds ratio, 1.96 [95% CI, 1.12-3.43]; P=.02) and absence of fear of dying (odds ratio, 2.38 [95% CI, 1.37-4.17]; P=.002) were associated with new organ failure after adjustment for STAI-State score (≥40), SAPS II, and SOFA score. CONCLUSION: Absence of fear of dying is associated with new organ failure within the first 7 days after ICU admission. Fear of dying may protect against subsequent deterioration by mobilizing patients' homeostatic resources. ClinicalTrials.gov Identifier: NCT02355626.

Serum and Tissue Biomarkers Associated With Composite of Relevant Endpoints for Sjögren Syndrome (CRESS) and Sjögren Tool for Assessing Response (STAR) to B Cell-Targeted Therapy in the Trial of Anti-B Cell Therapy in Patients With Primary Sjögren Syndrome (TRACTISS).

OBJECTIVE: This study aimed to identify peripheral and salivary gland (SG) biomarkers of response/resistance to B cell depletion based on the novel concise Composite of Relevant Endpoints for Sjögren Syndrome (cCRESS) and candidate Sjögren Tool for Assessing Response (STAR) composite endpoints. METHODS: Longitudinal analysis of peripheral blood and SG biopsies was performed pre- and post-treatment from the Trial of Anti-B Cell Therapy in Patients With Primary Sjögren Syndrome (TRACTISS) combining flow cytometry immunophenotyping, serum cytokines, and SG bulk RNA sequencing. RESULTS: Rituximab treatment prevented the worsening of SG inflammation observed in the placebo arm, by inhibiting the accumulation of class-switched memory B cells within the SG. Furthermore, rituximab significantly down-regulated genes involved in immune-cell recruitment, lymphoid organization alongside antigen presentation, and T cell co-stimulatory pathways. In the peripheral compartment, rituximab down-regulated immunoglobulins  and auto-antibodies together with pro-inflammatory cytokines and chemokines. Interestingly, patients classified as responders  according to STAR displayed significantly higher baseline levels of C-X-C motif chemokine ligand-13 (CXCL13), interleukin (IL)-22, IL-17A, IL-17F, and tumor necrosis factor-α (TNF-α), whereas a longitudinal analysis of serum T cell-related cytokines showed a selective reduction in both STAR and cCRESS responder patients. Conversely, cCRESS response was better associated with biomarkers of SG immunopathology, with cCRESS-responders showing a significant decrease in SG B cell infiltration and reduced expression of transcriptional gene modules related to T cell costimulation, complement activation, and Fcγ-receptor engagement. Finally, cCRESS and STAR response were associated with a significant improvement in SG exocrine function linked to transcriptional evidence of SG epithelial and metabolic restoration. CONCLUSION: Rituximab modulates both peripheral and SG inflammation, preventing the deterioration of exocrine function with functional and metabolic restoration of the glandular epithelium. Response assessed by newly developed cCRESS and STAR criteria was associated with differential modulation of peripheral and SG biomarkers, emerging as novel tools for patient stratification.

More rapid blood interferon α2 decline in fatal versus surviving COVID-19 patients.

Background: The clinical outcome of COVID-19 pneumonia is highly variable. Few biological predictive factors have been identified. Genetic and immunological studies suggest that type 1 interferons (IFN) are essential to control SARS-CoV-2 infection. Objective: To study the link between change in blood IFN-α2 level and plasma SARS-Cov2 viral load over time and subsequent death in patients with severe and critical COVID-19. Methods: One hundred and forty patients from the CORIMUNO-19 cohort hospitalized with severe or critical COVID-19 pneumonia, all requiring oxygen or ventilation, were prospectively studied. Blood IFN-α2 was evaluated using the Single Molecule Array technology. Anti-IFN-α2 auto-Abs were determined with a reporter luciferase activity. Plasma SARS-Cov2 viral load was measured using droplet digital PCR targeting the Nucleocapsid gene of the SARS-CoV-2 positive-strand RNA genome. Results: Although the percentage of plasmacytoid dendritic cells was low, the blood IFN-α2 level was higher in patients than in healthy controls and was correlated to SARS-CoV-2 plasma viral load at entry. Neutralizing anti-IFN-α2 auto-antibodies were detected in 5% of patients, associated with a lower baseline level of blood IFN-α2. A longitudinal analysis found that a more rapid decline of blood IFN-α2 was observed in fatal versus surviving patients: mortality HR=3.15 (95% CI 1.14-8.66) in rapid versus slow decliners. Likewise, a high level of plasma SARS-CoV-2 RNA was associated with death risk in patients with severe COVID-19. Conclusion: These findings could suggest an interest in evaluating type 1 IFN treatment in patients with severe COVID-19 and type 1 IFN decline, eventually combined with anti-inflammatory drugs. Clinical trial registration: https://clinicaltrials.gov, identifiers NCT04324073, NCT04331808, NCT04341584.

Use of covid-19 convalescent plasma to treat patients admitted to hospital for covid-19 with or without underlying immunodeficiency: open label, randomised clinical trial.

Objective: To evaluate the efficacy of covid-19 convalescent plasma to treat patients admitted to hospital for moderate covid-19 disease with or without underlying immunodeficiency (CORIPLASM trial). Design: Open label, randomised clinical trial. Setting: CORIMUNO-19 cohort (publicly supported platform of open label, randomised controlled trials of immune modulatory drugs in patients admitted to hospital with moderate or severe covid-19 disease) based on 19 university and general hospitals across France, from 16 April 2020 to 21 April 2021. Participants: 120 adults (n=60 in the covid-19 convalescent plasma group, n=60 in the usual care group) admitted to hospital with a positive SARS-CoV2 test result, duration of symptoms <9 days, and World Health Organization score of 4 or 5. 49 patients (n=22, n=27) had underlying immunosuppression. Interventions: Open label randomisation to usual care or four units (200-220 mL/unit, 2 units/day over two consecutive days) of covid-19 convalescent plasma with a seroneutralisation titre >40. Main outcome measures: Primary outcomes were proportion of patients with a WHO Clinical Progression Scale score of ≥6 on the 10 point scale on day 4 (higher values indicate a worse outcome), and survival without assisted ventilation or additional immunomodulatory treatment by day 14. Secondary outcomes were changes in WHO Clinical Progression Scale scores, overall survival, time to discharge, and time to end of dependence on oxygen supply. Predefined subgroups analyses included immunosuppression status, duration of symptoms before randomisation, and use of steroids. Results: 120 patients were recruited and assigned to covid-19 convalescent plasma (n=60) or usual care (n=60), including 22 (covid-19 convalescent plasma) and 27 (usual care) patients who were immunocompromised. 13 (22%) patients who received convalescent plasma had a WHO Clinical Progression Scale score of ≥6 at day 4 versus eight (13%) patients who received usual care (adjusted odds ratio 1.88, 95% credible interval 0.71 to 5.24). By day 14, 19 (31.6%) patients in the convalescent plasma group and 20 (33.3%) patients in the usual care group needed ventilation, additional immunomodulatory treatment, or had died. For cumulative incidence of death, three (5%) patients in the convalescent plasma group and eight (13%) in the usual care group died by day 14 (adjusted hazard ratio 0.40, 95% confidence interval 0.10 to 1.53), and seven (12%) patients in the convalescent plasma group and 12 (20%) in the usual care group by day 28 (adjusted hazard ratio 0.51, 0.20 to 1.32). In a subgroup analysis performed in patients who were immunocompromised, transfusion of covid-19 convalescent plasma was associated with mortality (hazard ratio 0.39, 95% confidence interval 0.14 to 1.10). Conclusions: In this study, covid-19 convalescent plasma did not improve early outcomes in patients with moderate covid-19 disease. The efficacy of convalescent plasma in patients who are immunocompromised should be investigated further. Trial registration: ClinicalTrials.gov NCT04345991.

The COVID-19 and chloroquine infodemic: Cross-sectional observational study of content analysis on YouTube.

The objective of this study is to evaluate the association between quality and features related to internet users of the most viewed YouTube videos about COVID-19 and chloroquine, during the first wave of the pandemic. We conducted a cross-sectional observational study focusing on the most viewed YouTube videos on COVID-19 and chloroquine treatment, in French and English. The primary outcome was the association between video quality as assessed by DISCERN and modified JAMA scores, and video features related to internet users as assessed by number of viewership and likes. By June 2020, 168 videos accumulating more than 57 million views and nearly 2 million reactions from Internet users, were included. Most of the videos did not support or oppose the treatment and came from news channels (N = 100; 60%). Videos taking sides were mostly pro-chloroquine (N = 69; 89%). The number of non-factual videos analyzed was very low (N = 3; 2%). The quality of the videos was average (mean DISCERN score = 2.4 (DS 1.0) and mean modified JAMA score = 2.6 (DS 0.6)) and declined over time. The best quality videos were those published by health care professionals or those from educational channels. Most experts interviewed were men (N = 136; 81%). More than 1 in 5 videos featured a political figure (N = 35; 21%), and these were mostly pro-chloroquine with lower DISCERN or modified JAMA scores (p<0.001). We found an association between the number of likes and the quality of the videos evaluated by the DISCERN score (rho = 0.29; p<0.001) and the modified JAMA score (rho = 0.30; p<0.001). The association observed between the quality of the videos and the number of likes suggests a better health literacy and critical thinking of lay internet users. Although YouTube has become a major player in the dissemination of medical information, more involvement of health professionals and governmental organizations is needed.

Can efficacy and safety data from clinical trials of rituximab in RA be extrapolated? Insights from 1984 patients from the AIR-PR Registry.

OBJECTIVES: To investigate whether the efficacy and safety data from drug-registration trials can be extrapolated to real-life RA patients receiving RTX. METHODS: The AIR-PR registry is a French multicentre, prospective cohort of RA patients treated with RTX in a real-life setting. We compared treatment responses at 12 months and serious AEs between eligible and non-eligible patients, by retrieving the eligibility criteria of the three rituximab-registration trials. We determined critical eligibility criteria and modelled the benefit-risk ratio according to the number of fulfilled critical eligibility criteria. RESULTS: Among 1984 RA patients, only 9-12% fulfilled all eligibility criteria. Non-eligible patients had less EULAR response at 12 months (40.3% vs 46.9%, p= 0.044). Critical inclusion criteria included SJC ≥ 4, TJC ≥ 4, CRP ≥ 15 mg/l, and RF positivity. Critical exclusion criteria were age >80 years, RA-associated systemic diseases, ACR functional class IV, other DMARD than methotrexate, and prednisone > 10 mg/day. Only 20.8% fulfilled those critical eligibility criteria. During the first year, serious AEs occurred for 182 (9.2%) patients, (70.3% serious infections) and patients with ≥1 critical exclusion criterion were at higher risk (HR 3.03; 95%CI 2.25-4.06; for ≥ 3 criteria vs 0). The incremental risk-benefit ratio decreased with the number of unmet critical inclusion criteria and of fulfilled exclusion criteria. CONCLUSION: Few real-life RA patients were eligible for the drug-registration trials. Non-eligible patients had lower chance of response, and higher risk of serious AEs. Efficacy and safety data obtained from those trials may not be generalizable to RA patients receiving RTX in real-world clinical practice.

Improving Diagnosis of Pulmonary Mucormycosis: Leads From a Contemporary National Study of 114 Cases.

BACKGROUND: Pulmonary mucormycosis (PM) is a life-threatening invasive mold infection. Diagnosis of mucormycosis is challenging and often delayed, resulting in higher mortality. RESEARCH QUESTION: Are the disease presentation of PM and contribution of diagnosis tools influenced by the patient's underlying condition? STUDY DESIGN AND METHODS: All PM cases from six French teaching hospitals between 2008 and 2019 were retrospectively reviewed. Cases were defined according to updated European Organization for Research and Treatment of Cancer/Mycoses Study Group criteria with the addition of diabetes and trauma as host factors and positive serum or tissue PCR as mycologic evidence. Thoracic CT scans were reviewed centrally. RESULTS: A total of 114 cases of PM were recorded, including 40% with disseminated forms. Main underlying conditions were hematologic malignancy (49%), allogeneic hematopoietic stem cell transplantation (21%), and solid organ transplantation (17%). When disseminated, main dissemination sites were the liver (48%), spleen (48%), brain (44%), and kidneys (37%). Radiologic presentation included consolidation (58%), pleural effusion (52%), reversed halo sign (26%), halo sign (24%), vascular abnormalities (26%), and cavity (23%). Serum quantitative polymerase chain reaction (qPCR) was positive in 42 (79%) of 53 patients and BAL in 46 (50%) of 96 patients. Results of transthoracic lung biopsy were diagnostic in 8 (73%) of 11 patients with noncontributive BAL. Overall 90-day mortality was 59%. Patients with neutropenia more frequently displayed an angioinvasive presentation, including reversed halo sign and disseminated disease (P < .05). Serum qPCR was more contributive in patients with neutropenia (91% vs 62%; P = .02), and BAL was more contributive in patients without neutropenia (69% vs 41%; P = .02). Serum qPCR was more frequently positive in patients with a > 3 cm main lesion (91% vs 62%; P = .02). Overall, positive qPCR was associated with an early diagnosis (P = .03) and treatment onset (P = .01). INTERPRETATION: Neutropenia and radiologic findings influence disease presentation and contribution of diagnostic tools during PM. Serum qPCR is more contributive in patients with neutropenia and BAL examination in patients without neutropenia. Results of lung biopsies are highly contributive in cases of noncontributive BAL.

What should be done and what should be avoided when comparing two treatments?

The preferred approach to compare two treatments is a randomized controlled trial (RCT). Indeed, randomization ensures that the groups compared are similar. Well-designed and well-conducted RCTs thus allow to draw causal conclusions on the relative efficacy and safety of treatments compared. However, it is not always possible to conduct RCTs for all clinical questions of interest, and observational data may also be used to infer on the relative effectiveness of treatments. In this review, we present different approaches that allow statistically valid comparisons of the effectiveness of treatments using observational data under some assumptions. Those are based on regression modelling or the propensity score. We also present the principles of target trial emulation.

Trajectories of the evolution of post-COVID-19 condition, up to two years after symptoms onset.

OBJECTIVES: We aimed to identify trajectories of the evolution of post-COVID-19 condition, up to 2 years after symptom onset. METHODS: The ComPaRe long COVID e-cohort is a prospective cohort of patients with symptoms lasting at least 2 months after SARS-CoV2 infection. We used trajectory modeling to identify different trajectories in the evolution of post-COVID-19 condition, based on symptoms collected every 60 days using the long COVID Symptom Tool. RESULTS: A total of 2197 patients were enrolled in the cohort between December 2020 and July 2022 when the Omicron variant was not dominant. Three trajectories of the evolution of post-COVID-19 condition were identified: "high persistent symptoms" (4%), "rapidly decreasing symptoms" (5%), and "slowly decreasing symptoms" (91%). Participants with highly persistent symptoms were older and more likely to report a history of systemic diseases. They often reported tachycardia, bradycardia, palpitations, and arrhythmia. Participants with rapidly decreasing symptoms were younger and more likely to report a confirmed infection. They often reported diarrhea and back pain. Participants with slowly decreasing symptoms were more likely to have a history of functional diseases. CONCLUSION: Most patients with post-COVID-19 condition improve slowly over time, while 5% have rapid improvement in the 2 years after symptom onset and 4% have a persistent condition.

Injectable long acting antiretroviral for HIV treatment and prevention: perspectives of potential users.

BACKGROUND: The use of long acting injectable (LAA) antiretroviral drugs may be an alternative option for HIV treatment and prevention. Our study focused on patient perspectives to understand which individuals, among people with HIV (PWH) and pre-exposure prophylaxis (PrEP) users, would constitute the preferential target for such treatments in terms of expectations, tolerability, adherence and quality of life. METHODS: The study consisted in one self-administrated questionnaire. Data collected included lifestyle issues, medical history, perceived benefits and inconveniences of LAA. Groups were compared using Wilcoxon rank tests or Fisher's exact test. RESULTS: In 2018, 100 PWH and 100 PrEP users were enrolled. Overall, 74% of PWH and 89% of PrEP users expressed interest for LAA with a significantly higher rate for PrEP users (p = 0.001). No characteristics were associated with acceptance of LAA in both groups in term of demographics, lifestyle or comorbidities. CONCLUSION: PWH and PrEP users expressed a high level of interest in LAA, since a large majority seems to be in favor of this new approach. Further studies should be conducted to better characterize targeted individuals.