Critically ill patients show an acute phase characterized by systemic spread of the inflammatory response, irrespectively of the cause of intensive care unit (ICU) admission, and late sequelae, including ICU acquired muscle weakness (ICUAW) and neurocognitive impairment. Mechanisms driving these late sequelae are unknown and there are no effective therapies to date. PreMed4PICS hypothesis is that skeletal muscle pathogenetic phenotypes and long-term sequelae in survivors to critical illness can be predicted at ICU admission in peripheral blood samples by transcriptomic profiling of the acute systemic response. Our main objective is to identify pathogenesis-dependent predictive signatures of muscle injury and clinical outcomes such as ICUAW or cognitive impairment. A multicentric prospective observational study will be conducted including adult patients admitted to the ICU and followed up until 12 months after ICU discharge. This will allow for clinical subphenotyping, sample acquisition and histopathological studies. To identify subphenotype-specific molecular pathways involved in skeletal muscle recovery, single-nuclei RNAseq will be performed. Massive sequencing of whole blood RNA and circulating microRNA at ICU admission will be performed to identify transcriptomic signatures that result in quantitative scores predictive of the outcomes of interest. All the findings will be confirmed in two validation cohorts. Collectively, this project aims to characterize the molecular mechanisms leading to ICUAW development and recovery, identifying therapeutic targets. The potential of a quantitative approach to the acute inflammatory response to predict long-term sequelae in survivors of critical illness will be validated.

The main objective of this project is to demonstrate that transcriptomic profiling of the acute systemic inflammatory response in patients admitted to the ICU predicts the main molecular pathways activated in the skeletal muscle at the onset of critical illness and during recovery, and the long term post-ICU sequelae. To achieve this objective the following specific aims have been established and organized into four work packages (WP) with specific objectives:

WP1: Clinical subphenotying and sampling in a large cohort of critically ill patients.

WP1.1. To characterize clinical features of the acute response to critical illness in two independents cohorts.

WP1.2. To evaluate long-term post-ICU sequelae after ICU discharge. WP1.3. To create a biobanked collection of samples from critically-ill patients.

WP2: Pathogenetic subphenotyping of skeletal muscle injury and recovery in critically ill patients.

WP2.1. To perform sn-RNAseq in muscle samples obtained at ICU admission and during the recovery phase from patients with and without persistent postICU muscle weakness.

WP2.2. To identify local pathogenetic subphenotypes according to the mechanisms involved in ICUAW development and recovery.

WP3: Transcriptomic profiling of peripheral whole blood during the acute response to critical illness.

WP3.1. To massively sequence gene expression and c-miRNA in peripheral blood obtained at ICU admission (discovery cohort).

WP3.2. To identify transcriptomic signatures in peripheral blood predictive of pathogenetic muscular phenotypes and long-term post-ICU sequelae.

WP3.3. To characterize the acute inflammatory profiles in response to critical illness according to the transcriptomic signatures.

Inclusion Criteria:

• Patients admitted to ICU

Exclusion Criteria:

• Neurocritical patients
• Previous neurocognitive or muscular impairments
• Immunosuppression
• Known neoplasia
• Terminal illness

Patients admitted to the ICU of three different tertiary university hospitals

Data Management Plan will be implemented and followed-up during the project respecting the legal data privacy requirements in accordance with the General Data Protection Regulation 2016/679/EU (GDPR). Ethical principles will be rigorously respected.

Genomic data will be shared in agreement with the Genomics Declaration. PreMed4PICS investigators will strive to make the data generated "FAIR", Findable, Accessible, Interoperable, and Reusable:

• Clinical data will be collected in RedCap, under the current legislation for this sensitive information. All recordings will be anonimized, and a master file linking anonymized data with patient IDs will be kept inside the hospital network.
• The code used to analyze will be shared at the public repository GitHub. All the code will be programmed in R language (open source).

The data will become available after public release of results, for at least 3 years after sharing.