Comparison of Incidence of Hyponatremia between Linezolid and Vancomycin by Propensity Score Matching Analysis

抄録

Several cases of severe hyponatremia induced by linezolid (LZD) were reported. However, severe infections could also cause hyponatremia by increasing vasopressin secretion. To prove that hyponatremia is associated with LZD rather than infection, we compared the incidence and risk of developing hyponatremia between patients receiving LZD and those receiving vancomycin (VCM). A retrospective, single-center, observational cohort study was conducted in patients aged 18 years or older who received intravenous LZD or VCM for 7 d or longer. Hyponatremia was defined as serum sodium level lower than 134 mEq/L and more than 5% decrease from baseline after treatment initiation. The incidence and risk of developing hyponatremia were analyzed between LZD and VCM groups using chi-square test. Four hundred and fifty patients who satisfied the selection criteria were divided into LZD (n = 97) and VCM groups (n = 353). Significant differences in patient characteristics between LZD and VCM groups were observed before propensity score matching, but no significant differences were found after matching. LZD group showed a significantly higher incidence and risk of developing hyponatremia compared to VCM group both before (LZD: 16.5%, VCM: 5.4%; p < 0.001, odds ratio 3.472 [95% confidence interval (CI) 1.711–7.048]) and after (LZD: 17.8%, VCM: 5.5%; p = 0.020, odds ratio 3.738 [95% CI 1.157–12.076]) propensity score matching. In conclusion, propensity score analyses suggest that the risk of hyponatremia associated with LZD is approximately 3.7-fold higher than that associated with VCM, regardless of patient background.