Abstract
Background and Aims
Hyponatremia is associated with aggravation of inflammation in COVID-19 patients. However, to the best of our knowledge, no study has used longitudinal data and investigated the association between hyponatremia and COVID-19. Therefore, we analyzed COVID-19 patients’ changes in serum sodium (Na) levels from admission to discharge.
Method
We conducted a retrospective, single-center, observational cohort study, involving adult COVID-19 patients who were admitted to Japan Self-Defense Forces Central Hospital between April 1 and May 31, 2020. Serum Na level of < 135 mEq/L was defined as hyponatremia, 135–145 mEq/L as normonatremia, and > 145 mEq/L as hypernatremia. Inflammation was assessed by serum C-reactive protein (CRP) levels. Univariate logistic regression analyses were used to assess associations between hyponatremia at admission and need for oxygen or death during hospitalization. A comparison of serum Na levels at admission and discharge was tested using a paired t-test. Cross-sectional associations between serum Na and CRP levels at admission or days from onset to admission were analyzed using multivariate linear regression analyses. A restricted cubic spline (RCS) curve incorporated in one of these multivariate linear regression analyses was used to identify when serum Na levels were the lowest. In addition, we employed a mixed-effect model to examine the longitudinal association between changes in serum Na and CRP levels during hospitalization.
Results
Ninety-eight patients were enrolled, of whom 53 (54%) were male and 39 (40%) had a smoking history. Mean (SD) or median (IQR) of age, eGFR, body mass index (BMI), serum Na, and CRP at admission were 50 (17) years, 82 (20) mL/min/1.73m2, 23.3 (5.5), 138 (3.7) mEq/L, and 1.8 (0.2–6.0) mg/dL, respectively. It took an mean of 8.9 (3.7) days from onset to admission due to social disruption. According to government policies, during observation period, patients diagnosed with COVID-19 have to be hospitalized, even if they have minor or improving clinical symptoms.
At admission, hyponatremia was observed in 11 (11.2%) patients; the rest of the patients had normonatremia. Twenty-seven (27.6%) patients received oxygen, and 4 (4.1%) died during hospitalization. Hyponatremia at admission was significantly associated with the need for oxygen (odds ratio: 41.2; 95% CI: 4.9–344; P=0.001) and death (odds ratio: 32.3; 95% CI: 3.0-347; P=0.004). Irrespective of hyponatremia at admission, the serum Na levels at discharge were significantly higher than those at admission (Fig. 1).
In both cross-sectional and longitudinal analyses, serum Na levels were negatively associated with serum CRP levels after adjustment for age, sex, eGFR, BMI, and smoking history (P<0.001), which suggests that serum Na levels may reflect the activity of COVID-19. Furthermore, the serum Na levels at admission were positively associated with days from onset to admission after adjustment for age, sex, eGFR, BMI, smoking history, and serum CRP levels (P=0.035), and the RCS curve showed that the serum Na levels tended to be the lowest around the 7th day after onset (Fig. 2).
Conclusion
Hyponatremia in COVID-19 may occur secondarily, and a condition called “COVID-19-induced hyponatremia” might exist.
Being Conscious of Water Intake Positively Associated with Sufficient Non-Alcohol Drink Intake Regardless of Seasons and Reasons in Healthy Japanese; the KOBE Study: A Cross Sectional Study