Andrew Sumarsono1,2, Jiexin Wang1, Luyu Xie3,4, Giuliana Cerro Chiang4, Thomas Tielleman5, Sarah E Messiah3,4, Amit G Singal5, Arjmand Mufti5, Catherine Chen6, Matthew Leveno6. 1. Department of Internal Medicine, UT Southwestern Medical Center, Dallas, TX. 2. Division of Hospital Medicine, Parkland Memorial Hospital, Dallas, TX. 3. University of Texas Health Science Center at Houston, School of Public Health, Dallas Regional Campus, Dallas, TX. 4. Center for Pediatric Population Health, Children's Health System of Texas and University of Texas Health Science Center, Dallas, TX. 5. Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Women's Guild Lung Institute, Cedar Sinai Medical Center, Los Angeles, CA. 6. Division of Digestive and Liver Diseases, Department of Internal Medicine, UT Southwestern Medical Center, Dallas, TX.
Abstract
OBJECTIVES: Cirrhosis is frequently complicated by electrolyte disturbances, with prior studies primarily focused on the importance of hyponatremia. Emerging evidence on patients with chronic heart failure and chronic kidney disease has identified hypochloremia as an independent predictor for mortality. This study aimed to investigate the prognostic value of serum chloride and its association with mortality in cirrhotic patients. DESIGN: Retrospective cohort study. SETTING: The medical ICU at Parkland Memorial Hospital, a tertiary care public health system in Dallas, Texas. PATIENTS: Adult patients with confirmed diagnosis of decompensated cirrhosis who were admitted to the ICU between March 2015 and March 2017. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Kaplan-Meier analysis and multivariable Cox proportional hazard ratio models were performed to determine the impact of hypochloremia on 180-day mortality. Of the 389 enrolled patients, 133 (34.2%) died within 180 days of ICU admission. Patients with hypochloremia had higher 180-day mortality than those with normochloremia (45.2% vs 26.7%; p < 0.0001). Cumulative survival via the Kaplan-Meier method was significantly lower in the hypochloremic group. Serum chloride was independently associated with 180-day mortality with multivariable adjustment (hazard ratio, 0.95; 95% CI, 0.93-0.98; p = 0.001) or after adjusting for Model for End-stage Liver Disease or Sequential Organ Failure Assessment. Contrarily, the inverse association between serum sodium and mortality no longer existed in all multivariable models. CONCLUSIONS: Serum chloride is independently and inversely associated with short-term mortality in critically ill cirrhotic patients. Hypochloremia, but not hyponatremia, remained associated with mortality with multivariable analyses, suggesting that hypochloremia may account for the mortality risk previously attributed to hyponatremia. These findings signify the prognostic value of serum chloride and potential inclusion of chloride into future cirrhosis prognostic scores.
OBJECTIVES:Cirrhosis is frequently complicated by electrolyte disturbances, with prior studies primarily focused on the importance of hyponatremia. Emerging evidence on patients with chronic heart failure and chronic kidney disease has identified hypochloremia as an independent predictor for mortality. This study aimed to investigate the prognostic value of serum chloride and its association with mortality in cirrhotic patients. DESIGN: Retrospective cohort study. SETTING: The medical ICU at Parkland Memorial Hospital, a tertiary care public health system in Dallas, Texas. PATIENTS: Adult patients with confirmed diagnosis of decompensated cirrhosis who were admitted to the ICU between March 2015 and March 2017. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Kaplan-Meier analysis and multivariable Cox proportional hazard ratio models were performed to determine the impact of hypochloremia on 180-day mortality. Of the 389 enrolled patients, 133 (34.2%) died within 180 days of ICU admission. Patients with hypochloremia had higher 180-day mortality than those with normochloremia (45.2% vs 26.7%; p < 0.0001). Cumulative survival via the Kaplan-Meier method was significantly lower in the hypochloremic group. Serum chloride was independently associated with 180-day mortality with multivariable adjustment (hazard ratio, 0.95; 95% CI, 0.93-0.98; p = 0.001) or after adjusting for Model for End-stage Liver Disease or Sequential Organ Failure Assessment. Contrarily, the inverse association between serum sodium and mortality no longer existed in all multivariable models. CONCLUSIONS: Serum chloride is independently and inversely associated with short-term mortality in critically ill cirrhoticpatients. Hypochloremia, but not hyponatremia, remained associated with mortality with multivariable analyses, suggesting that hypochloremia may account for the mortality risk previously attributed to hyponatremia. These findings signify the prognostic value of serum chloride and potential inclusion of chloride into future cirrhosis prognostic scores.