Increasing evidence suggests that dysnatremia at intensive care unit (ICU) admission may predict mortality. Little information is available, however, on the potential effect of dysnatremia correction. This is an observational multicenter cohort study in patients admitted between 2005 and 2012 to 18 French ICUs. Hyponatremia and hypernatremia were defined as serum sodium concentration less than 135 and more than 145 mmol/L, respectively. We assessed the influence on day 28 mortality of dysnatremia correction by day 3 and of the dysnatremia correction rate. Of 7,067 included patients, 1,830 (25.9%) had hyponatremia and 634 (9.0%) had hypernatremia at ICU admission (day 1). By day 3, hyponatremia had been corrected in 1,019 (1,019/1,830; 55.7%) and hypernatremia in 393 (393/634; 62.0%) patients. After adjustment for confounders, persistent hyponatremia or hypernatremia on day 3 was independently associated with higher day 28 mortality (odds ratio [OR], 1.31; 95% confidence interval [95% CI], 1.06 - 1.61; and OR, 1.86; 95% CI, 1.37 - 2.54; respectively). Hyponatremia corrected by day 3, hypernatremia corrected by day 3, and ICU-acquired hyponatremia were not associated with day 28 mortality. Median correction rate from days 1 to 3 was 2.58 mmol/L per day (interquartile range, 0.67 - 4.55). Higher natremia correction rate was associated with lower crude and adjusted day 28 mortality rates (OR per mmol/L per day, 0.97; 95% CI, 0.94 - 1.00; P = 0.04; and OR per mmol/L per day, 0.93; 95% CI, 0.90 - 0.97; P = 0.0003, respectively). Our results indicate that dysnatremia correction is independently associated with survival, with the effect being greater with faster correction rates of up to 12 mmol/L per day.