Dataset: Systematic review and meta-analysis of extravascular lung water index (EVLWI) and mortality in critically ill adults
Description
Research Hypothesis This dataset is grounded in the hypothesis that there is a measurable prognostic association between the extravascular lung water index (EVLWI), measured by transpulmonary thermodilution, and mortality outcomes in critically ill adult patients. Data Description and Collection This file contains raw, study-level aggregate data compiled for an updated systematic review and meta-analysis. The data were systematically extracted from primary literature to evaluate EVLWI effect-size inputs. The dataset is structured into two main analytical domains: Continuous-effect dataset: Provides group sample sizes, EVLWI means, standard deviations, mean differences, standard errors, and 95% confidence intervals comparing EVLWI between non-survivors and survivors. Association dataset: Contains log odds ratios, standard errors, back-transformed odds ratios, and 95% confidence intervals for mortality based on higher versus lower EVLWI, or study-specific definitions of elevated EVLWI. Study identification: Includes essential variables such as record ID, author, year, and subgroup classifications to facilitate the exploration of clinical or methodological heterogeneity. What the Data Shows and Notable Findings: The dataset provides a comprehensive overview of the extracted effect sizes used in the quantitative synthesis. Notable summary metrics include: The continuous-effect synthesis encompasses 27 study-level comparisons , representing a total analytical sample of 2,423 patients (1,092 non-survivors and 1,331 survivors). The association-effect dataset includes records from 17 studies with their respective odds ratio estimates. Subgroup variables demonstrate clinical heterogeneity across multiple domains: mortality endpoints (12 records for 28-day mortality, 15 for ICU mortality) ; sepsis status (12 septic, 15 mixed populations) ; and ARDS status (8 ARDS, 19 mixed populations). The timing of EVLWI measurements is recorded across day 1 (18 records), day 2 (5 records), and day 3 (4 records). Data Interpretation and Usage This file is provided to improve methodological transparency, facilitate reproducibility, and allow independent verification of the forest plots and subgroup analyses. To appropriately interpret and utilize this dataset: The study-level aggregated data must be interpreted together with the eligibility criteria, extraction procedures, and statistical analysis plan detailed in the main manuscript and supplementary material. Because thresholds and definitions of elevated EVLWI vary across primary studies, the odds-ratio dataset should not be interpreted as establishing or supporting a single universal clinical cutoff. Variables retaining the terminology "ELWI" refer directly to the EVLWI construct evaluated throughout the manuscript. Both author-verified and forest-plot reproduction values are retained where applicable to preserve complete transparency in how the uploaded forest plots were generated.