Introduction To correctly interpret plasma- or serum-sphingosine-1-phosphate (S1P) concentrations measured in clinical studies it is critical to understand all major determinants in healthy controls. Methods Serum- and plasma-S1P from 174 healthy blood donors was measured by liquid chromatography-tandem mass spectrometry and correlated to clinical laboratory data. Selected plasma samples, 10 with high and 10 with low S1P concentrations, were fractionated into very low-density lipoprotein (VLDL)-, low density lipoprotein (LDL)-, high density lipoprotein (HDL)-, and lipoprotein-free fractions. S1P was then measured in each fraction to determine its distribution. Results The mean S1P concentration in serum (1.04 ± 0.24 nmol/mL) was found 39% higher compared with plasma (0.75 ± 0.16 nmol/mL) and overall was not different between men and women. Only when stratified for age and gender, older women were found to exhibit higher circulatory S1P levels than men. In plasma, S1P levels correlate to red blood cell (RBC) counts but not to platelet counts. Conversely, serum-S1P correlates to platelet counts but not to RBC counts. In addition, eosinophil counts are strongly associated with serum-S1P concentrations. Both serum- and plasma-S1P correlate to total cholesterol but not to HDL-C. The distribution of S1P between VLDL-, LDL-, HDL-, and lipoprotein-free fractions is independent of total plasma-S1P concentrations. S1P concentrations in HDL but not in LDL are highly variable. Conclusion These data indicate S1P concentrations in plasma and serum to be differentially associated with cell counts and S1P carrier proteins. Besides platelets, eosinophil counts are identified as a novel determinant for serum-S1P concentrations further suggesting a role for S1P in eosinophil pathologies.

Background and Purpose: The aim of this study was to examine whether sphingosine-1-phosphate (S1P) levels in patients with acute stroke are associated with stroke severity and outcome. Methods: In a prospective stroke cohort (MARK-STROKE), 374 patients with acute ischemic stroke or transient ischemic attack were enrolled (mean age: 67.9±13.0 years, sex: 64.7% male), and serum-S1P at admission was analyzed with tandem mass spectrometry. In addition to cross-sectional analyses, 79 adverse events (death, stroke, myocardial infarction, rehospitalization) were recorded in 270 patients during follow-up. Regression analyses were adjusted for age, sex, low-density lipoprotein cholesterol, and vascular risk factors. Results were validated in an independent stroke cohort with 219 patients with acute ischemic stroke (CIRCULAS). Results: Low serum-S1P was associated with higher National Institutes of Health Stroke Scale score at admission and with anterior circulation nonlacunar infarcts determined by multivariate regression analyses. During a follow-up of 294±170 days, patients with S1P in the lowest tertile (<1.33 µmol/L) had more adverse events (Kaplan-Meier analysis, P =0.048 for trend). In adjusted Cox regression analysis, the lowest S1P tertile was associated with a worse outcome after stroke (hazard ratio, HR 0.51 [95% confidence interval 0.28–0.92]). Results were confirmed in an independent cohort, ie, low S1P levels were associated with higher National Institutes of Health Stroke Scale, larger infarct volumes and worse outcome after 90 days (β-coefficient: –0.03, P =0.026; β-coefficient: −0.099, P =0.009 and odds ratio 0.52 [0.28–0.96], respectively). Conclusions: Our findings imply a detrimental role of low S1P levels in acute stroke and therefore underpin the therapeutic potential of S1P-mimics.