Introduction: Nitric oxide (NO) likely plays a pivotal role in the pathogenesis of sepsis. Arginine is a substrate for NO, whereas the methylated arginines—asymmetric dimethylarginine (ADMA) and symmetric dimethylarginine (SDMA)—are endogenous by-products of proteolysis that inhibit NO production. We investigated if high-plasma levels of ADMA, SDMA, and arginine/ADMA ratio were associated with 90-day mortality in patients with severe sepsis or septic shock. Methods: We included 267 adult patients admitted to intensive care unit with severe sepsis or septic shock. The patients had previously been included in the randomized controlled trial “Scandinavian Starch for Severe Sepsis and Septic Shock (6S).” ADMA, SDMA, and arginine/ADMA ratio were measured in plasma. The risk of death within 90 days was estimated in multivariate Cox regression analyses adjusted for gender, age >=65 years, major cardiovascular disease, diabetes, hypertension, respiratory failure, vasopressor treatment, highest quartile of creatinine and bilirubin, and lowest quartile of platelet count. In the regression analyses missing values were estimated using multiple imputation. Results: Twenty-five patients had missing data in one or more of the baseline variables and 44 patients had missing methylarginine values. Both ADMA and SDMA were independently associated with 90-day mortality (ADMA: hazard ratio 1.54; 95% CI, 1.00–2.38; P = 0.046, and SDMA: hazard ratio 1.78; 95% CI, 1.14–2.72; P = 0.011). Arginine/ADMA ratio was not associated with 90-day mortality neither in univariate nor in multivariate analyses. The difference in mortality between patients with high and low ADMA was most pronounced in the first week after inclusion. Conclusions: High levels of ADMA and SDMA in plasma were associated with increased 90-day mortality in patients with severe sepsis or septic shock. Interfering with the methylarginine-NO systems may be a novel target in these patients.