Electrolyte Variability During Acute Decompensated Heart Failure Hospitalization
Jaskeerat S. Gulati,Michael Kimbrell,Lin Zhong,Yiqing Wang,Donghan Yang,Bethany Roehm,Susan S. Hedayati,Mary Jane Farr,Nicholas Hendren,Mark Drazner,Wai Hong Tang,Justin Grodin
DOI: https://doi.org/10.1016/j.cardfail.2023.10.299
IF: 6.592
2024-01-01
Journal of Cardiac Failure
Abstract:Introduction Abnormal electrolytes are associated with greater risk for adverse events during treatment of acute decompensated heart failure (ADHF). Variability in electrolyte levels can be quantified by calculating average successive variability (ASV) - defined as the average absolute difference between successive values. The clinical implications of the ASV of serially measured inpatient electrolytes for ADHF are unknown. Hypothesis Greater variability in serum sodium (Na), chloride (Cl), potassium (K), and bicarbonate (HCO3), as assessed by the ASV over 96-hours, will be associated with greater 96-h urine output (UOP) and 96-h change in serum creatinine. Methods Data from patient encounters with a discharge diagnosis of decompensated heart failure were abstracted from the Texas Health Resources EHR, a 26-hospital health system in North Texas from 2008-2018 (Epic, Verona, WI). Cohort eligibility criteria were: 1) home furosemide PO equivalent dose of ≥20 mg/d; 2) IV loop diuretic use within 24 h of hospitalization; 3) initial serum creatinine ≤3.0 mmol/L; 4) systolic blood pressure ≥90 mm Hg throughout first 24 h of hospitalization; and 5) length of hospitalization ≥96 h. The ASV was calculated for over 4 values of each electrolyte for 477 encounters. Spearman correlation coefficients between ASV of each electrolyte and creatinine were calculated and multivariable linear regression models (adjusted for age, race, sex, ethnicity, BMI, BNP, Na, K, CI, creatinine, hemoglobin, albumin, home loop diuretic dose, ACE/ARB/ARNI, beta-blocker, and MRA use) tested the association of each electrolyte ASV to 96-h UOP and 96-h creatinine change. Results In the study cohort (51.9% female, median [IQR] age 74 [65-83] years, LVEF 50% [33-62], BNP 1091 [546-1994] pg/mL, and creatinine 1.5 [1.1-1.9] mg/dL), the baseline value and 96-h ASV for each electrolyte was: Na 138 [135-141] meq/L, ASV 1.5 mmol/L; Cl 102 [98-105] mmol/L, ASV 1.7 mmol/L; K 4.2 [3.8-4.5] mmol/L, ASV 0.3 mmol/L; and HCO3 25 [23-29] mmol/L, ASV 2.0 mmol/L. All 4 electrolytes mildly correlated with the ASV of creatinine: Na (r=0.25), Cl (r=0.20), K (r=0.16), and HCO3 (r=0.11, P<0.05 for each). After multivariable adjustment, only the variability in Cl was associated with 96-h UOP (beta=892 mL per ASV mmol/L, P=0.03, Figure 1 panel A-D) but not Na, K, or HCO3. No electrolyte ASV was associated with 96-h change in creatinine (Figure 1 panel E-H). Conclusion Unlike Na, K, or HCO3, variability in Cl during the treatment of ADHF is independently associated with increased 96-h UOP. No electrolyte variabilities were associated with 96-h serum creatinine change. These observations represent a link between Cl variability and active decongestion and support the hypothesis that Cl variability may represent diuretic responsiveness.
cardiac & cardiovascular systems