Winkler · Critical care (London, England) 2018 · prospective cohort study · n=120

Symmetrical (SDMA) and asymmetrical dimethylarginine (ADMA) in sepsis: high plasma levels as combined risk markers for sepsis survival.

Cited 62 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective non-randomized observational cohort study assessing prognostic markers.

PubMed 30231905 · doi:10.1186/s13054-018-2090-1 · record verified 2026-08-30

What was done

A prospective single-center study enrolled 120 ICU patients diagnosed with sepsis. Plasma levels of symmetric dimethylarginine (SDMA) and asymmetric dimethylarginine (ADMA) were measured via mass spectrometry on days 1, 3, and 7. Disease severity was evaluated using the Sequential Organ Failure Assessment (SOFA) score, and survival was tracked through day 28. Associations between marker levels, SOFA scores, and 28-day mortality were analyzed using non-parametric tests, Kaplan-Meier curves, and a decision tree algorithm.

What was found

Median admission plasma concentrations of both markers were significantly higher in non-survivors than survivors: SDMA was 1.14 vs. 0.82 µmol/L (P = 0.002) and ADMA was 0.93 vs. 0.73 µmol/L (P = 0.016). Higher concentrations of both metabolites were significantly associated with higher SOFA scores. The 28-day mortality increased across percentiles: for SDMA, mortality was 12% in the lower, 25% in the intermediate, and 43% in the ≥75th percentile (P = 0.018); for ADMA, mortality was 18–20% in lower/intermediate groups and 48% in the ≥75th percentile (P = 0.006). Combining both markers identified a subset with SDMA > 1.34 µmol/L and ADMA > 0.97 µmol/L who experienced 61% mortality.

Why it matters

SDMA and ADMA, endogenous inhibitors of nitric oxide pathways, serve as measurable prognostic biomarkers that identify septic ICU patients at high risk of 28-day mortality.

Limits

The study was conducted at a single center with a modest sample size (n = 120). Nitric oxide bioavailability itself was not directly quantified, multivariable adjustment details are not reported in the abstract, and the decision tree cutoff thresholds require external validation.

Cited by