BackgroundKeshan disease (KD) is strongly associated with selenium deficiency. Selenoprotein P (SELENOP) is a recognized molecular biomarker of selenoproteins and an important indicator of selenium nutrition. This study was aimed at providing geographically precisely visualized evidence of selenium nutrition at molecular level for assessing KD prevention, control, and elimination on the etiological perspective.MethodsWe used spatial ecological design for this study. The serum SELENOP levels of the residents were measured by ELISA. ArcGIS version 9.0 was used for spatial description, spatial autocorrelation analysis of SELENOP levels and spatial regression with per capita disposable income.ResultsThe mean serum SELENOP levels of the 6,382 residents in 1,688 counties were 4.62 ± 1.82 μg/mL. The mean serum SELENOP levels of the residents living in the townships and rural areas of KD endemic counties were not statistically significantly lower than those of the KD non-endemic counties. The mean serum SELENOP levels were globally clustered (Moran's I = 0.03, z = 6.37, and P < 0.0001), and 99.3% (553/557) of the cold spots, identified by local autocorrelation analysis (Getis-Ord-Gi* analysis), were located in the KD endemic provinces of Shaanxi, Shanxi, Henan, Hebei, Shandong, Inner Mongolia, Gansu, Hubei, Chongqing, Yunnan, and Sichuan. The serum SELENOP level was positively correlated with per capita disposable income (t = 3.52, P = 0.0004).ConclusionsThe results of this study were the geographically precisely visualized evidence of selenium nutrition at molecular level for assessing KD elimination on the etiological perspective. The cold spot counties found by Getis-Ord-Gi* analysis in the KD endemic provinces should be the high priority of KD precision prevention and control.