This study looks at how different ways of giving magnesium (a nutrient) affect the risk of death within 28 days for adults with low magnesium levels who are in the intensive care unit (ICU) for non-heart-related reasons. We will use data from past ICU patients (2008-2019) to compare three approaches:

No magnesium treatment (even if levels are low); Daily magnesium treatment (regardless of daily levels); Dynamic treatment (only giving magnesium when daily tests show low levels). The goal is to find out which strategy is safest and most effective for this group of patients.

This is a target trial emulation (a type of observational study that mimics a clinical trial) using data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database (2008-2019).

Inclusion criteria: Adults (≥18 years) admitted to the ICU, with serum magnesium <1.7 mg/dL within 72 hours of admission, and no history of heart disease or cardiac surgery.

Sample size: 12,887 patients. Interventions (simulated): Three magnesium replacement strategies (Never Treat, Always Treat, Dynamic Treat).

Outcome: 28-day all-cause mortality (death from any cause). Analysis methods: We will use the G-formula (to handle time-varying factors that could affect results) and sensitivity analyses to check if findings are reliable.

Inclusion Criteria:

1. Age ≥ 18 years.
2. Admitted to the Intensive Care Unit (ICU).
3. Developed hypomagnesemia (Serum Magnesium < 1.7 mg/dL) within the first 72 hours of ICU admission.

Exclusion Criteria:

1. Admitted due to cardiac or cardiosurgical causes (e.g., Cardiac Surgery Recovery Unit, CCU, CVICU).
2. Underwent cardiac surgery during hospitalization (identified by ICD-9/10 procedure codes).
3. Primary discharge diagnosis of major cardiac conditions (e.g., myocardial infarction, heart failure, arrhythmias).

This study includes adult (≥18 years) non-cardiac critically ill patients with hypomagnesemia (serum magnesium < 1.7 mg/dL) within 72 hours of ICU admission, identified from the MIMIC-IV (version 3.1) database (2008-2019). Patients with cardiac-related admission or missing key data are excluded. The population represents a diverse cohort of non-cardiac ICU patients (e.g., respiratory, gastrointestinal, neurological etiologies) to evaluate the association between dynamic magnesium replacement strategies and 28-day mortality.

Individual participant data (IPD) cannot be shared because the study uses data from the MIMIC-IV database, which requires researchers to complete mandatory training and obtain independent authorization from the MIMIC-IV Data Access Committee. IPD is not owned by the study team; qualified researchers may access the data directly via the official MIMIC-IV application process (https://mimic.mit.edu/).