Dataset for "Berberine for prevention of contrast-induced acute kidney injury in patients with type 2 diabetes undergoing coronary angiography: a single-center randomized controlled trial"

DATASET README==============Title: Dataset for "Berberine for prevention of contrast-induced acute kidney injury in patientswith type 2 diabetes undergoing coronary angiography: a single-center randomized controlled trialwith secondary biomarker, prediction, urine-output and pharmacogenomic analyses" Study: Single-center, randomized, open-label, outcomes-assessor-blinded controlled trial ofperiprocedural oral berberine vs no berberine (N=769 randomized; N=706 in the ITT analysis set).Registration: ClinicalTrials.gov NCT02808351 (registered 21 June 2016, prospective).Ethics: Ethics Committee of Nanjing First Hospital, approval No. KY20160608-04; all participantsprovided written informed consent. FILES- berberine_ciaki_dataset.csv : de-identified analysis dataset (769 rows = all randomized participants; the ITT analysis set is the subset with dm==1 and non-missing ciaki_peak).- data_dictionary.csv : variable-level documentation (description, unit, coding). DE-IDENTIFICATIONNo names, hospital record numbers, dates, or free-text fields are included. The 'id' column is asequential anonymous row identifier (the original randomization number contained duplicate entriesand is not used in any analysis). Ages and routine clinical measurements are retained as collected. NOTES FOR REUSE1. Serum creatinine was scheduled pre-procedure and at 24/72 h only; a 48-h measurement was never part of the protocol design (cr_48h intentionally empty). The primary endpoint used peak in-hospital creatinine.2. Urinary NGAL values were flagged as unexpected and potentially unreliable (see manuscript); CYP2C19 genotype data failed Hardy-Weinberg quality control (P=0.011) and are provided for transparency only — they should not be analyzed.3. HbA1c is missing for 50.5% of participants.4. Analysis code reproducing all results accompanies the associated manuscript as Supporting Information (Python 3).