Kevin McCarthy1, Richard Conway1, Declan Byrne1, Seán Cournane2, Deirdre O'Riordan1, Bernard Silke3. 1. Department of Internal Medicine, St James's Hospital, Dublin 8, Ireland. 2. Medical Physics and Bioengineering Department, St James's Hospital, Dublin 8, Ireland. 3. Department of Internal Medicine, St James's Hospital, Dublin 8, Ireland. Electronic address: bernardsilke@physicians.ie.
Abstract
BACKGROUND: Altered sodium balance at time of an emergency medical admission adversely impacts on outcome; whether hyponatraemia is independently associated with outcomes or a surrogate of acute illness severity has been debated. METHODS: All emergency medical admissions between 2002 and 2017 were studied and a risk score calculated. We compared univarate deciles of admission sodium using a multivariable model, adjusting for risk score. RESULTS: There were 106,586 admissions in 54,928 patients. Patients with lower sodium at admission were older at 66.7 years (IQR 46.7-79.5) compared with 63.3 years (IQR 42.9-78.2) with a longer length of stay (LOS) of 6.8 days (IQR 3.0-14.7) versus 4.9 days (IQR 1.8-10.9). They had a higher 30-day in-hospital mortality at 6.4% vs 4.4% (p < 0.001). Admission sodium predicted survival - OR 0.89 (95%CI 0.88-0.90). We adjusted the model with a Risk Score that is predictive and exponentially related to 30-day in-hospital mortality. When adjusted for Risk Score, the admission sodium value was less predictive - OR 0.95 (95%CI 0.92-0.97). The cumulative percentages within the lowest five deciles fell from 63.3% between 2002 and 2009 to 48.1% from 2010 to 2017. The slope of the prediction line relating admission sodium to mortality did not change over time but a lower mortality rate was predicted at any given sodium level. CONCLUSION: Hyponatraemia at the time of an emergency medical admission is predictive and probably a marker of Acuity Illness Severity and Case Complexity. Both the frequency of abnormality in admission sodium and mortality have fallen over time.
BACKGROUND: Altered sodium balance at time of an emergency medical admission adversely impacts on outcome; whether hyponatraemia is independently associated with outcomes or a surrogate of acute illness severity has been debated. METHODS: All emergency medical admissions between 2002 and 2017 were studied and a risk score calculated. We compared univarate deciles of admission sodium using a multivariable model, adjusting for risk score. RESULTS: There were 106,586 admissions in 54,928 patients. Patients with lower sodium at admission were older at 66.7 years (IQR 46.7-79.5) compared with 63.3 years (IQR 42.9-78.2) with a longer length of stay (LOS) of 6.8 days (IQR 3.0-14.7) versus 4.9 days (IQR 1.8-10.9). They had a higher 30-day in-hospital mortality at 6.4% vs 4.4% (p < 0.001). Admission sodium predicted survival - OR 0.89 (95%CI 0.88-0.90). We adjusted the model with a Risk Score that is predictive and exponentially related to 30-day in-hospital mortality. When adjusted for Risk Score, the admission sodium value was less predictive - OR 0.95 (95%CI 0.92-0.97). The cumulative percentages within the lowest five deciles fell from 63.3% between 2002 and 2009 to 48.1% from 2010 to 2017. The slope of the prediction line relating admission sodium to mortality did not change over time but a lower mortality rate was predicted at any given sodium level. CONCLUSION: Hyponatraemia at the time of an emergency medical admission is predictive and probably a marker of Acuity Illness Severity and Case Complexity. Both the frequency of abnormality in admission sodium and mortality have fallen over time.