Can Machine Learning Personalize Cardiovascular Therapy in Sepsis?

Large randomized trials in sepsis have generally failed to find effective novel treatments. This is increasingly attributed to patient heterogeneity, including heterogeneous cardiovascular changes in septic shock. We discuss the potential for machine learning systems to personalize cardiovascular resuscitation in sepsis. While the literature is replete with proofs of concept, the technological readiness of current systems is low, with a paucity of clinical trials and proven patient benefit. Systems may be vulnerable to confounding and poor generalization to new patient populations or contemporary patterns of care. Typical electronic health records do not capture rich enough data, at sufficient temporal resolution, to produce systems that make actionable treatment suggestions. To resolve these issues, we recommend a simultaneous focus on technical challenges and removing barriers to translation. This will involve improving data quality, adopting causally grounded models, prioritizing safety assessment and integration into healthcare workflows, conducting randomized clinical trials and aligning with regulatory requirements.

Assessing the safety of physical rehabilitation in critically ill patients: a Delphi study.

BACKGROUND: Physical rehabilitation of critically ill patients is implemented to improve physical outcomes from an intensive care stay. However, before rehabilitation is implemented, a risk assessment is essential, based on robust safety data. To develop this information, a uniform definition of relevant adverse events is required. The assessment of cardiovascular stability is particularly relevant before physical activity as there is uncertainty over when it is safe to start rehabilitation with patients receiving vasoactive drugs. METHODS: A three-stage Delphi study was carried out to (a) define adverse events for a general ICU cohort, and (b) to define which risks should be assessed before physical rehabilitation of patients receiving vasoactive drugs. An international group of intensive care clinicians and clinician researchers took part. Former ICU patients and their family members/carers were involved in generating consensus for the definition of adverse events. Round one was an open round where participants gave their suggestions of what to include. In round two, participants rated their agreements with these suggestions using a five-point Likert scale; a 70% consensus agreement threshold was used. Round three was used to re-rate suggestions that had not reached consensus, whilst viewing anonymous feedback of participant ratings from round two. RESULTS: Twenty-four multi-professional ICU clinicians and clinician researchers from 10 countries across five continents were recruited. Average duration of ICU experience was 18 years (standard deviation 8) and 61% had publications related to ICU rehabilitation. For the adverse event definition, five former ICU patients and one patient relative were recruited. The Delphi process had a 97% response rate. Firstly, 54 adverse events reached consensus; an adverse event tool was created and informed by these events. Secondly, 50 risk factors requiring assessment before physical rehabilitation of patients receiving vasoactive drugs reached consensus. A second tool was created, informed by these suggestions. CONCLUSIONS: The adverse event tool can be used in studies of physical rehabilitation to ensure uniform measurement of safety. The risk assessment tool can be used to inform clinical practise when risk assessing when to start rehabilitation with patients receiving vasoactive drugs. Trial registration This study protocol was retrospectively registered on https://www.researchregistry.com/ (researchregistry2991).

From ICU Syndromes to ICU Subphenotypes: Consensus Report and Recommendations For Developing Precision Medicine in ICU.

Critical care uses syndromic definitions to describe patient groups for clinical practice and research. There is growing recognition that a "precision medicine" approach is required and that integrated biologic and physiologic data identify reproducible subpopulations that may respond differently to treatment. This article reviews the current state of the field and considers how to successfully transition to a precision medicine approach. In order to impact clinical care, identified subpopulations must do more than differentiate prognosis. They must differentiate response to treatment, ideally by defining subgroups with distinct functional or pathobiological mechanisms (endotypes). There are now multiple examples of reproducible subpopulations of sepsis, acute respiratory distress syndrome, and acute kidney or brain injury described using clinical, physiological, and/or biological data. Many of these subpopulations have demonstrated the potential to define differential treatment response, largely in retrospective studies, and that the same treatment-responsive subpopulations may cross multiple clinical syndromes (treatable traits). To bring about a change in clinical practice, a precision medicine approach must be evaluated in prospective clinical studies requiring novel adaptive trial designs. Several such studies are underway but there are multiple challenges to be tackled. Such subpopulations must be readily identifiable and be applicable to all critically ill populations around the world. Subdividing clinical syndromes into subpopulations will require large patient numbers. Global collaboration of investigators, clinicians, industry and patients over many years will therefore be required to transition to a precision medicine approach and ultimately realize treatment advances seen in other medical fields. This article is open access and distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives License 4.0 (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Adaptive multi-interventional trial platform to improve patient care for fibrotic interstitial lung diseases.

BACKGROUND: Fibrotic interstitial lung diseases (fILDs) are a heterogeneous group of lung diseases associated with significant morbidity and mortality. Despite a large increase in the number of clinical trials in the last 10 years, current regulatory-approved management approaches are limited to two therapies that prevent the progression of fibrosis. The drug development pipeline is long and there is an urgent need to accelerate this process. This manuscript introduces the concept and design of an innovative research approach to drug development in fILD: a global Randomised Embedded Multifactorial Adaptive Platform in fILD (REMAP-ILD). METHODS: Description of the REMAP-ILD concept and design: the specific terminology, design characteristics (multifactorial, adaptive features, statistical approach), target population, interventions, outcomes, mission and values, and organisational structure. RESULTS: The target population will be adult patients with fILD, and the primary outcome will be a disease progression model incorporating forced vital capacity and mortality over 12 months. Responsive adaptive randomisation, prespecified thresholds for success and futility will be used to assess the effectiveness and safety of interventions. REMAP-ILD embraces the core values of diversity, equity, and inclusion for patients and researchers, and prioritises an open-science approach to data sharing and dissemination of results. CONCLUSION: By using an innovative and efficient adaptive multi-interventional trial platform design, we aim to accelerate and improve care for patients with fILD. Through worldwide collaboration, novel analytical methodology and pragmatic trial delivery, REMAP-ILD aims to overcome major limitations associated with conventional randomised controlled trial approaches to rapidly improve the care of people living with fILD.

Developing Guidance for Donor Intervention Randomized Controlled Trials: Initial Discussions From the Canada-United Kingdom 2022 Workshop.

BACKGROUND: Donor interventions, including medications, protocols, and medical devices administered to donors, can enhance transplantable organ quality and quantity and maximize transplantation success. However, there is paucity of high-quality evidence about their effectiveness, in part because of ethical, practical, and regulatory challenges, and lack of guidance about conduct of donor intervention randomized controlled trials (RCTs). METHODS: With the vision to develop authoritative guidance for conduct of donor intervention RCTs, we convened a workshop of Canadian-United Kingdom experts in organ donation and transplantation ethics, research, and policy to identify stakeholders, explore unique challenges, and develop research agenda to inform future work in this promising field. RESULTS: Donor intervention trials should consider perspectives of broad group of stakeholders including donors, transplant recipients, and their families; researchers in donation and transplantation; research ethics boards; and healthcare providers and administrators involved in donation and transplantation. Unique challenges include (1) research ethics (living versus deceased status of the donor at the time of intervention, intervention versus outcomes assessment in different individuals, harm-benefit analysis in donors versus recipients, consent, and impact on research bystanders); (2) outcome data standardization and linkage; and (3) regulatory and governance considerations. CONCLUSIONS: Donor intervention RCTs hold potential to benefit organ transplantation outcomes but face unique research ethics, outcome data, and regulatory challenges. By developing research agenda to address these challenges, our workshop was an important first step toward developing Canada-United Kingdom guidance for donor intervention RCTs that are poised to improve the quality and availability of transplantable organs.

Patient stratification using plasma cytokines and their regulators in sepsis: relationship to outcomes, treatment effect and leucocyte transcriptomic subphenotypes.

RATIONALE: Heterogeneity of the host response within sepsis, acute respiratory distress syndrome (ARDS) and more widely critical illness, limits discovery and targeting of immunomodulatory therapies. Clustering approaches using clinical and circulating biomarkers have defined hyper-inflammatory and hypo-inflammatory subphenotypes in ARDS associated with differential treatment response. It is unknown if similar subphenotypes exist in sepsis populations where leucocyte transcriptomic-defined subphenotypes have been reported. OBJECTIVES: We investigated whether inflammatory clusters based on cytokine protein abundance were seen in sepsis, and the relationships with previously described transcriptomic subphenotypes. METHODS: Hierarchical cluster and latent class analysis were applied to an observational study (UK Genomic Advances in Sepsis (GAinS)) (n=124 patients) and two clinical trial datasets (VANISH, n=155 and LeoPARDS, n=484) in which the plasma protein abundance of 65, 21, 11 circulating cytokines, cytokine receptors and regulators were quantified. Clinical features, outcomes, response to trial treatments and assignment to transcriptomic subphenotypes were compared between inflammatory clusters. MEASUREMENTS AND MAIN RESULTS: We identified two (UK GAinS, VANISH) or three (LeoPARDS) inflammatory clusters. A group with high levels of pro-inflammatory and anti-inflammatory cytokines was seen that was associated with worse organ dysfunction and survival. No interaction between inflammatory clusters and trial treatment response was found. We found variable overlap of inflammatory clusters and leucocyte transcriptomic subphenotypes. CONCLUSIONS: These findings demonstrate that differences in response at the level of cytokine biology show clustering related to severity, but not treatment response, and may provide complementary information to transcriptomic sepsis subphenotypes. TRIAL REGISTRATION NUMBER: ISRCTN20769191, ISRCTN12776039.

A Transcriptomic Approach to Understand Patient Susceptibility to Pneumonia After Abdominal Surgery.

OBJECTIVE: To describe immune pathways and gene networks altered following major abdominal surgery and to identify transcriptomic patterns associated with postoperative pneumonia. BACKGROUND: Nosocomial infections are a major healthcare challenge, developing in over 20% of patients aged 45 or over undergoing major abdominal surgery, with postoperative pneumonia associated with an almost 5-fold increase in 30-day mortality. METHODS: From a prospective consecutive cohort (n=150) undergoing major abdominal surgery, whole-blood RNA was collected preoperatively and at 3 time-points postoperatively (2-6, 24, and 48 h). Twelve patients diagnosed with postoperative pneumonia and 27 matched patients remaining infection-free were identified for analysis with RNA-sequencing. RESULTS: Compared to preoperative sampling, 3639 genes were upregulated and 5043 downregulated at 2 to 6 hours. Pathway analysis demonstrated innate-immune activation with neutrophil degranulation and Toll-like-receptor signaling upregulation alongside adaptive-immune suppression. Cell-type deconvolution of preoperative RNA-sequencing revealed elevated S100A8/9-high neutrophils alongside reduced naïve CD4 T-cells in those later developing pneumonia. Preoperatively, a gene-signature characteristic of neutrophil degranulation was associated with postoperative pneumonia acquisition ( P =0.00092). A previously reported Sepsis Response Signature (SRSq) score, reflecting neutrophil dysfunction and a more dysregulated host response, at 48 hours postoperatively, differed between patients subsequently developing pneumonia and those remaining infection-free ( P =0.045). Analysis of the novel neutrophil gene-signature and SRSq scores in independent major abdominal surgery and polytrauma cohorts indicated good predictive performance in identifying patients suffering later infection. CONCLUSIONS: Major abdominal surgery acutely upregulates innate-immune pathways while simultaneously suppressing adaptive-immune pathways. This is more prominent in patients developing postoperative pneumonia. Preoperative transcriptomic signatures characteristic of neutrophil degranulation and postoperative SRSq scores may be useful predictors of subsequent pneumonia risk.

Alpha 2 agonists for sedation to produce better outcomes from critical illness (A2B Trial): protocol for a multicentre phase 3 pragmatic clinical and cost-effectiveness randomised trial in the UK.

INTRODUCTION: Almost all patients receiving mechanical ventilation (MV) in intensive care units (ICUs) require analgesia and sedation. The most widely used sedative drug is propofol, but there is uncertainty whether alpha2-agonists are superior. The alpha 2 agonists for sedation to produce better outcomes from critical illness (A2B) trial aims to determine whether clonidine or dexmedetomidine (or both) are clinically and cost-effective in MV ICU patients compared with usual care. METHODS AND ANALYSIS: Adult ICU patients within 48 hours of starting MV, expected to require at least 24 hours further MV, are randomised in an open-label three arm trial to receive propofol (usual care) or clonidine or dexmedetomidine as primary sedative, plus analgesia according to local practice. Exclusions include patients with primary brain injury; postcardiac arrest; other neurological conditions; or bradycardia. Unless clinically contraindicated, sedation is titrated using weight-based dosing guidance to achieve a Richmond-Agitation-Sedation score of -2 or greater as early as considered safe by clinicians. The primary outcome is time to successful extubation. Secondary ICU outcomes include delirium and coma incidence/duration, sedation quality, predefined adverse events, mortality and ICU length of stay. Post-ICU outcomes include mortality, anxiety and depression, post-traumatic stress, cognitive function and health-related quality of life at 6-month follow-up. A process evaluation and health economic evaluation are embedded in the trial.The analytic framework uses a hierarchical approach to maximise efficiency and control type I error. Stage 1 tests whether each alpha2-agonist is superior to propofol. If either/both interventions are superior, stages 2 and 3 testing explores which alpha2-agonist is more effective. To detect a mean difference of 2 days in MV duration, we aim to recruit 1437 patients (479 per group) in 40-50 UK ICUs. ETHICS AND DISSEMINATION: The Scotland A REC approved the trial (18/SS/0085). We use a surrogate decision-maker or deferred consent model consistent with UK law. Dissemination will be via publications, presentations and updated guidelines. TRIAL REGISTRATION NUMBER: ClinicalTrials.gov NCT03653832.

Clinician and patient perspectives on the barriers and facilitators to physical rehabilitation in intensive care: a qualitative interview study.

OBJECTIVES: The objective of this study is to explore patient, relative/carer and clinician perceptions of barriers to early physical rehabilitation in intensive care units (ICUs) within an associated group of hospitals in the UK and how they can be overcome. DESIGN: Qualitative study using semi-structured interviews and thematic framework analysis. SETTING: Four ICUs over three hospital sites in London, UK. PARTICIPANTS: Former ICU patients or their relatives/carers with personal experience of ICU rehabilitation. ICU clinicians, including doctors, nurses, physiotherapists and occupational therapists, involved in the delivery of physical rehabilitation or decisions over its initiation. PRIMARY AND SECONDARY OUTCOMES MEASURES: Views and experiences on the barriers and facilitators to ICU physical rehabilitation. RESULTS: Interviews were carried out with 11 former patients, 3 family members and 16 clinicians. The themes generated related to: safety and physiological concerns, patient participation and engagement, clinician experience and knowledge, teamwork, equipment and environment and risks and benefits of rehabilitation in intensive care. The overarching theme for overcoming barriers was a change in working model from ICU clinicians having separate responsibilities (a multidisciplinary approach) to one where all parties have a shared aim of providing patient-centred ICU physical rehabilitation (an interdisciplinary approach). CONCLUSIONS: The results have revealed barriers that can be modified to improve rehabilitation delivery in an ICU. Interdisciplinary working could overcome many of these barriers to optimise recovery from critical illness.

Quantifying the impact of AI recommendations with explanations on prescription decision making.

The influence of AI recommendations on physician behaviour remains poorly characterised. We assess how clinicians' decisions may be influenced by additional information more broadly, and how this influence can be modified by either the source of the information (human peers or AI) and the presence or absence of an AI explanation (XAI, here using simple feature importance). We used a modified between-subjects design where intensive care doctors (N = 86) were presented on a computer for each of 16 trials with a patient case and prompted to prescribe continuous values for two drugs. We used a multi-factorial experimental design with four arms, where each clinician experienced all four arms on different subsets of our 24 patients. The four arms were (i) baseline (control), (ii) peer human clinician scenario showing what doses had been prescribed by other doctors, (iii) AI suggestion and (iv) XAI suggestion. We found that additional information (peer, AI or XAI) had a strong influence on prescriptions (significantly for AI, not so for peers) but simple XAI did not have higher influence than AI alone. There was no correlation between attitudes to AI or clinical experience on the AI-supported decisions and nor was there correlation between what doctors self-reported about how useful they found the XAI and whether the XAI actually influenced their prescriptions. Our findings suggest that the marginal impact of simple XAI was low in this setting and we also cast doubt on the utility of self-reports as a valid metric for assessing XAI in clinical experts.

biomArker-guided Duration of Antibiotic treatment in hospitalised Patients with suspecTed Sepsis (ADAPT-Sepsis): A protocol for a multicentre randomised controlled trial.

Aim: To describe the protocol for a multi-centre randomised controlled trial to determine whether treatment protocols monitoring daily CRP (C-reactive protein) or PCT (procalcitonin) safely allow a reduction in duration of antibiotic therapy in hospitalised adult patients with sepsis. Design: Multicentre three-arm randomised controlled trial. Setting: UK NHS hospitals. Target population: Hospitalised critically ill adults who have been commenced on intravenous antibiotics for sepsis. Health technology: Three protocols for guiding antibiotic discontinuation will be compared: (a) standard care; (b) standard care + daily CRP monitoring; (c) standard care + daily PCT monitoring. Standard care will be based on routine sepsis management and antibiotic stewardship. Measurement of outcomes and costs. Outcomes will be assessed to 28 days. The primary outcomes are total duration of antibiotics and safety outcome of all-cause mortality. Secondary outcomes include: escalation of care/re-admission; infection re-lapse/recurrence; antibiotic dose; length and level of critical care stay and length of hospital stay. Ninety-day all-cause mortality rates will also be collected. An assessment of cost effectiveness will be performed. Conclusion: In the setting of routine NHS care, if this trial finds that a treatment protocol based on monitoring CRP or PCT safely allows a reduction in duration of antibiotic therapy, and is cost effective, then this has the potential to change clinical practice for critically ill patients with sepsis. Moreover, if a biomarker-guided protocol is not found to be effective, then it will be important to avoid its use in sepsis and prevent ineffective technology becoming widely adopted in clinical practice.

Landiolol and Organ Failure in Patients With Septic Shock: The STRESS-L Randomized Clinical Trial.

Importance: Patients with septic shock undergo adrenergic stress, which affects cardiac, immune, inflammatory, and metabolic pathways. ß-Blockade may attenuate the adverse effects of catecholamine exposure and has been associated with reduced mortality. Objectives: To assess the efficacy and safety of landiolol in patients with tachycardia and established septic shock requiring prolonged (>24 hours) vasopressor support. Design, Setting, and Participants: An open-label, multicenter, randomized trial involving 126 adults (≥18 years) with tachycardia (heart rate ≥95/min) and established septic shock treated for at least 24 hours with continuous norepinephrine (≥0.1 µg/kg/min) in 40 UK National Health Service intensive care units. The trial ran from April 2018 to December 2021, with early termination in December 2021 due to a signal of possible harm. Intervention: Sixty-three patients were randomized to receive standard care and 63 to receive landiolol infusion. Main Outcomes and Measures: The primary outcome was the mean Sequential Organ Failure Assessment (SOFA) score from randomization through 14 days. Secondary outcomes included mortality at days 28 and 90 and the number of adverse events in each group. Results: The trial was stopped prematurely on the advice of the independent data monitoring committee because it was unlikely to demonstrate benefit and because of possible harm. Of a planned 340 participants, 126 (37%) were enrolled (mean age, 55.6 years [95% CI, 52.7 to 58.5 years]; 58.7% male). The mean (SD) SOFA score in the landiolol group was 8.8 (3.9) compared with 8.1 (3.2) in the standard care group (mean difference [MD], 0.75 [95% CI, -0.49 to 2.0]; P = .24). Mortality at day 28 after randomization in the landiolol group was 37.1% (23 of 62) and 25.4% (16 of 63) in the standard care group (absolute difference, 11.7% [95% CI, -4.4% to 27.8%]; P = .16). Mortality at day 90 after randomization was 43.5% (27 of 62) in the landiolol group and 28.6% (18 of 63) in the standard care group (absolute difference, 15% [95% CI, -1.7% to 31.6%]; P = .08). There were no differences in the number of patients having at least one adverse event. Conclusion and Relevance: Among patients with septic shock with tachycardia and treated with norepinephrine for more than 24 hours, an infusion of landiolol did not reduce organ failure measured by the SOFA score over 14 days from randomization. These results do not support the use of landiolol for managing tachycardia among patients treated with norepinephrine for established septic shock. Trial Registration: EU Clinical Trials Register Eudra CT: 2017-001785-14; isrctn.org Identifier: ISRCTN12600919.

Intravenous Vitamin C for Patients Hospitalized With COVID-19: Two Harmonized Randomized Clinical Trials.

Importance: The efficacy of vitamin C for hospitalized patients with COVID-19 is uncertain. Objective: To determine whether vitamin C improves outcomes for patients with COVID-19. Design, Setting, and Participants: Two prospectively harmonized randomized clinical trials enrolled critically ill patients receiving organ support in intensive care units (90 sites) and patients who were not critically ill (40 sites) between July 23, 2020, and July 15, 2022, on 4 continents. Interventions: Patients were randomized to receive vitamin C administered intravenously or control (placebo or no vitamin C) every 6 hours for 96 hours (maximum of 16 doses). Main Outcomes and Measures: The primary outcome was a composite of organ support-free days defined as days alive and free of respiratory and cardiovascular organ support in the intensive care unit up to day 21 and survival to hospital discharge. Values ranged from -1 organ support-free days for patients experiencing in-hospital death to 22 organ support-free days for those who survived without needing organ support. The primary analysis used a bayesian cumulative logistic model. An odds ratio (OR) greater than 1 represented efficacy (improved survival, more organ support-free days, or both), an OR less than 1 represented harm, and an OR less than 1.2 represented futility. Results: Enrollment was terminated after statistical triggers for harm and futility were met. The trials had primary outcome data for 1568 critically ill patients (1037 in the vitamin C group and 531 in the control group; median age, 60 years [IQR, 50-70 years]; 35.9% were female) and 1022 patients who were not critically ill (456 in the vitamin C group and 566 in the control group; median age, 62 years [IQR, 51-72 years]; 39.6% were female). Among critically ill patients, the median number of organ support-free days was 7 (IQR, -1 to 17 days) for the vitamin C group vs 10 (IQR, -1 to 17 days) for the control group (adjusted proportional OR, 0.88 [95% credible interval {CrI}, 0.73 to 1.06]) and the posterior probabilities were 8.6% (efficacy), 91.4% (harm), and 99.9% (futility). Among patients who were not critically ill, the median number of organ support-free days was 22 (IQR, 18 to 22 days) for the vitamin C group vs 22 (IQR, 21 to 22 days) for the control group (adjusted proportional OR, 0.80 [95% CrI, 0.60 to 1.01]) and the posterior probabilities were 2.9% (efficacy), 97.1% (harm), and greater than 99.9% (futility). Among critically ill patients, survival to hospital discharge was 61.9% (642/1037) for the vitamin C group vs 64.6% (343/531) for the control group (adjusted OR, 0.92 [95% CrI, 0.73 to 1.17]) and the posterior probability was 24.0% for efficacy. Among patients who were not critically ill, survival to hospital discharge was 85.1% (388/456) for the vitamin C group vs 86.6% (490/566) for the control group (adjusted OR, 0.86 [95% CrI, 0.61 to 1.17]) and the posterior probability was 17.8% for efficacy. Conclusions and Relevance: In hospitalized patients with COVID-19, vitamin C had low probability of improving the primary composite outcome of organ support-free days and hospital survival. Trial Registration: ClinicalTrials.gov Identifiers: NCT04401150 (LOVIT-COVID) and NCT02735707 (REMAP-CAP).

Heterogeneous Treatment Effects of Therapeutic-Dose Heparin in Patients Hospitalized for COVID-19.

Importance: Randomized clinical trials (RCTs) of therapeutic-dose heparin in patients hospitalized with COVID-19 produced conflicting results, possibly due to heterogeneity of treatment effect (HTE) across individuals. Better understanding of HTE could facilitate individualized clinical decision-making. Objective: To evaluate HTE of therapeutic-dose heparin for patients hospitalized for COVID-19 and to compare approaches to assessing HTE. Design, Setting, and Participants: Exploratory analysis of a multiplatform adaptive RCT of therapeutic-dose heparin vs usual care pharmacologic thromboprophylaxis in 3320 patients hospitalized for COVID-19 enrolled in North America, South America, Europe, Asia, and Australia between April 2020 and January 2021. Heterogeneity of treatment effect was assessed 3 ways: using (1) conventional subgroup analyses of baseline characteristics, (2) a multivariable outcome prediction model (risk-based approach), and (3) a multivariable causal forest model (effect-based approach). Analyses primarily used bayesian statistics, consistent with the original trial. Exposures: Participants were randomized to therapeutic-dose heparin or usual care pharmacologic thromboprophylaxis. Main Outcomes and Measures: Organ support-free days, assigning a value of -1 to those who died in the hospital and the number of days free of cardiovascular or respiratory organ support up to day 21 for those who survived to hospital discharge; and hospital survival. Results: Baseline demographic characteristics were similar between patients randomized to therapeutic-dose heparin or usual care (median age, 60 years; 38% female; 32% known non-White race; 45% Hispanic). In the overall multiplatform RCT population, therapeutic-dose heparin was not associated with an increase in organ support-free days (median value for the posterior distribution of the OR, 1.05; 95% credible interval, 0.91-1.22). In conventional subgroup analyses, the effect of therapeutic-dose heparin on organ support-free days differed between patients requiring organ support at baseline or not (median OR, 0.85 vs 1.30; posterior probability of difference in OR, 99.8%), between females and males (median OR, 0.87 vs 1.16; posterior probability of difference in OR, 96.4%), and between patients with lower body mass index (BMI <30) vs higher BMI groups (BMI ≥30; posterior probability of difference in ORs >90% for all comparisons). In risk-based analysis, patients at lowest risk of poor outcome had the highest propensity for benefit from heparin (lowest risk decile: posterior probability of OR >1, 92%) while those at highest risk were most likely to be harmed (highest risk decile: posterior probability of OR <1, 87%). In effect-based analysis, a subset of patients identified at high risk of harm (P = .05 for difference in treatment effect) tended to have high BMI and were more likely to require organ support at baseline. Conclusions and Relevance: Among patients hospitalized for COVID-19, the effect of therapeutic-dose heparin was heterogeneous. In all 3 approaches to assessing HTE, heparin was more likely to be beneficial in those who were less severely ill at presentation or had lower BMI and more likely to be harmful in sicker patients and those with higher BMI. The findings illustrate the importance of considering HTE in the design and analysis of RCTs. Trial Registration: ClinicalTrials.gov Identifiers: NCT02735707, NCT04505774, NCT04359277, NCT04372589.

Heterogeneity of surrogate outcome measures used in critical care studies: A systematic review.

BACKGROUND: The choice of outcome measure is a critical decision in the design of any clinical trial, but many Phase III clinical trials in critical care fail to detect a difference between the interventions being compared. This may be because the surrogate outcomes used to show beneficial effects in early phase trials (which informed the design of the subsequent Phase III trials) are not valid guides to the differences between the interventions for the main outcomes of the Phase III trials. We undertook a systematic review (1) to generate a list of outcome measures used in critical care trials, (2) to determine the variability in the outcome reporting in the respiratory subgroup and (3) to create a smaller list of potential early phase endpoints in the respiratory subgroup. METHODS: Data related to outcomes were extracted from studies published in the six top-ranked critical care journals between 2010 and 2020. Outcomes were classified into subcategories and categories. A subset of early phase endpoints relevant to the respiratory subgroup was selected for further investigation. The variability of the outcomes and the variability in reporting was investigated. RESULTS: A total of 6905 references were retrieved and a total of 294 separate outcomes were identified from 58 studies. The outcomes were then classified into 11 categories and 66 subcategories. A subset of 22 outcomes relevant for the respiratory group were identified as potential early phase outcomes. The summary statistics, time points and definitions show the outcomes are analysed and reported in different ways. CONCLUSION: The outcome measures were defined, analysed and reported in a variety of ways. This creates difficulties for synthesising data in systematic reviews and planning definitive trials. This review once again highlights an urgent need for standardisation and validation of surrogate outcomes reported in critical care trials. Future work should aim to validate and develop a core outcome set for surrogate outcomes in critical care trials.

An immune dysfunction score for stratification of patients with acute infection based on whole-blood gene expression.

Dysregulated host responses to infection can lead to organ dysfunction and sepsis, causing millions of global deaths each year. To alleviate this burden, improved prognostication and biomarkers of response are urgently needed. We investigated the use of whole-blood transcriptomics for stratification of patients with severe infection by integrating data from 3149 samples from patients with sepsis due to community-acquired pneumonia or fecal peritonitis admitted to intensive care and healthy individuals into a gene expression reference map. We used this map to derive a quantitative sepsis response signature (SRSq) score reflective of immune dysfunction and predictive of clinical outcomes, which can be estimated using a 7- or 12-gene signature. Last, we built a machine learning framework, SepstratifieR, to deploy SRSq in adult and pediatric bacterial and viral sepsis, H1N1 influenza, and COVID-19, demonstrating clinically relevant stratification across diseases and revealing some of the physiological alterations linking immune dysregulation to mortality. Our method enables early identification of individuals with dysfunctional immune profiles, bringing us closer to precision medicine in infection.