| Description: |
Background/Objectives: Hyponatremia has been associated with mortality and hospital readmissions. Although multimorbid older patients are particularly affected, specific data on this group are lacking. Methods: A prospective cohort was used based on the OPERAM (OPtimising thERapy to prevent Avoidable hospital admissions in the Multimorbid elderly) trial, a European multicenter, cluster-randomized trial among hospitalized patients aged ≥70 years with ≥3 chronic medical conditions taking ≥5 long-term medications, with documented sodium values at admission, excluding participants with hypernatremia (>145 mmol/L). The primary outcome was all-cause 1-year mortality, and secondary outcomes were 30-day mortality and readmission at 1 year and at 30 days. We examined the association between hyponatremia and mortality in comparison to normonatremia using a mixed-effects survival model, with adjustment for age, sex, comorbidities, study intervention arm, study site and cluster; and the association between hyponatremia and readmission using competing risk models with death as the competing risk. Subgroup analyses were performed across sodium hyponatremia categories (mild 134-130 mmol/L, moderate 129-125 mmol/L, severe < 125 mmol/L). Results: Of 2008 OPERAM participants, 1968 had a sodium value at admission, and 33 were excluded due to hypernatremia. In the 1935 participants, the mean age was 79.4 years (standard deviation 6.3), 866 (44.8%) were female, the median number of comorbidities was 11 (IQR 8-16), the median number of drugs was 10 (IQR 7-13), and 401 (20.7%) had hyponatremia at admission. The multivariate-adjusted hazard ratio (HR) for 1-year mortality with hyponatremia was 1.41 (95% confidence interval [CI] 1.11-1.78, 364 deaths) and for 30-day mortality was 1.20 (95%CI 0.74-1.94, 89 deaths). The adjusted sub-HR for 1-year readmission was 0.94 (95%CI 0.79-1.11), and that for 30-day readmission was 1.1 (95%CI 0.78-1.59). There was a linear increase in 1-year mortality across hyponatremia categories (HR from ... |