Associations between day of admission, admission hyponatremia and hospital outcomes in medical patients: A retrospective multicenter cohort study

Abstract

Introduction We investigated the relationship between admission day, serum sodium concentration, and outcomes in medical inpatients. Hyponatremia, the most common electrolyte abnormality in hospitalized adults, is associated with prolonged hospital length of stay (LOS) and higher mortality. Weekend admissions are also linked to worse outcomes, but the magnitude of this “weekend effect” may vary with diagnoses and particular day of admission. Design Retrospective multicenter cohort study. Setting Four Ministry of National Guard – Health Affairs hospitals in Saudi Arabia (January 1, 2016 – May 9, 2022). Participants 43,361 adult medical admissions. Patients with hypernatremia (n = 1,892) or LOS > 30 days (n = 2,988) were excluded. Interventions Admissions were categorized by admission day (Sunday – Saturday) and serum sodium concentration (normonatremia and hyponatremia with severity sub-classification). Main outcome measures LOS, intensive therapy unit (ITU) admission, and mortality. Statistical analyses Multivariable gamma generalized linear models (GLM) were used to evaluate LOS, and logistic regression for ITU admission and mortality, adjusting for age, sex, and hospital site. Sensitivity analyses examined associations with the COVID-19 pandemic. Results Hyponatremia on admission (42.4%) was associated with longer LOS than normonatremia, on all admission days (except Fridays). Association strength between admission day and LOS decreased as the severity of hyponatremia increased. In Saudi Arabia the statutory weekend is Friday-Saturday. Friday admissions had the longest LOS, while Mondays had the shortest (reference = Sunday). Tuesday admissions showed an unexpectedly prolonged LOS, potentially reflecting internal medicine resident half day teaching activities. Mortality was not significantly associated with admission day or serum sodium concentration in adjusted analyses, but the study was underpowered to detect modest mortality differences (58% power for 10% relative risk reduction). Sensitivity analyses demonstrated that admission day associations with LOS and ITU admission remained consistent across the pre-COVID-19 and COVID-19 periods. Admission during the COVID-19 pandemic was associated with increased mortality (OR 1.41) in the primary analysis. Sensitivity analyses demonstrated that this excess mortality probably reflected an additive main effect of the COVID-19 period rather than a differential effect by day of admission, confirming that the primary findings on admission day associations are robust. Conclusion This study reveals associations between admission day, serum sodium, and hospital outcomes. LOS varied with both admission day and hyponatremia severity. Friday admissions had the longest LOS, but admission day associations with LOS weakened with increasing hyponatremia severity. Institutional scheduling and staffing patterns may create day-specific variations that extend beyond the traditional weekend-weekday dichotomy. These temporal variations highlight opportunities for targeted workflow adjustments and resource allocation that could improve care delivery, and potentially reduce morbidity and mortality.