OBJECTIVE: Endoglin, an anti-angiogenic glycoprotein expressed on endothelial cells, has been proposed recently as a biomarker of pre-eclampsia (PE). Given that PE is characterised by an imbalance of angiogenic factors, we sought to determine the clinical utility of urinary soluble endoglin, relative to the soluble fms-like tyrosine kinase 1 to placental growth factor (PlGF) ratio, in the diagnosis of PE during gestation. DESIGN: Prospective observational cohort. SETTING: Tertiary referral university hospital. POPULATION: Two hundred and thirty-four pregnant women were enrolled prospectively in the following groups: healthy controls, n = 63; gestational age (GA), median (interquartile range), 33 weeks (27-39 weeks); chronic hypertension, n = 27; GA, 33 weeks (30-36 weeks); mild PE, n = 38; GA, 37 weeks (34-40 weeks); severe PE, n = 106; GA, 32 weeks (29-37 weeks). METHODS: Free urinary levels of soluble endoglin, soluble fms-like tyrosine kinase 1 and PlGF were measured by sensitive and specific immunoassay. Levels for all urinary analytes were normalised to creatinine. MAIN OUTCOME MEASURES: Urinary soluble endoglin, and the soluble fms-like tyrosine kinase 1 to PlGF ratio. RESULTS: In healthy controls, urinary soluble endoglin levels were increased significantly at term relative to those earlier in gestation. Severe PE was characterised by an increased urinary level of soluble endoglin, soluble fms-like tyrosine kinase 1, protein to creatinine ratio and soluble fms-like tyrosine kinase 1 to PlGF ratio compared with all other groups. There was a direct correlation between urinary soluble endoglin and proteinuria that remained after GA correction (R = 0.382, P < 0.001). Urinary soluble endoglin could not differentiate mild PE from severe preterm PE. Overall, soluble endoglin had the ability to discriminate PE from chronic hypertension and healthy controls only in women who were evaluated at <37 weeks of GA. The sensitivity, specificity and accuracy of urinary soluble endoglin alone in the diagnosis of PE or in the identification of women with PE requiring a mandated delivery before 37 weeks of gestation were 70%, 86% and 76%, respectively. These values were inferior to those of the soluble fms-like tyrosine kinase 1 to PlGF ratio (P < 0.001). The addition of urinary soluble endoglin did not improve the diagnostic accuracy of the soluble fms-like tyrosine kinase 1 to PlGF ratio alone. CONCLUSIONS: We have provided evidence that soluble endoglin is present and elevated in the urine of women who develop preterm PE. Urinary soluble endoglin has only limited ability to determine the severity of PE and to distinguish between PE and chronic hypertension both preterm and at term. Compared with urinary soluble endoglin, the soluble fms-like tyrosine kinase 1 to PlGF ratio remains a better marker of disease presence, severity and outcome.
OBJECTIVE:Endoglin, an anti-angiogenic glycoprotein expressed on endothelial cells, has been proposed recently as a biomarker of pre-eclampsia (PE). Given that PE is characterised by an imbalance of angiogenic factors, we sought to determine the clinical utility of urinary soluble endoglin, relative to the soluble fms-like tyrosine kinase 1 to placental growth factor (PlGF) ratio, in the diagnosis of PE during gestation. DESIGN: Prospective observational cohort. SETTING: Tertiary referral university hospital. POPULATION: Two hundred and thirty-four pregnant women were enrolled prospectively in the following groups: healthy controls, n = 63; gestational age (GA), median (interquartile range), 33 weeks (27-39 weeks); chronic hypertension, n = 27; GA, 33 weeks (30-36 weeks); mild PE, n = 38; GA, 37 weeks (34-40 weeks); severe PE, n = 106; GA, 32 weeks (29-37 weeks). METHODS: Free urinary levels of soluble endoglin, soluble fms-like tyrosine kinase 1 and PlGF were measured by sensitive and specific immunoassay. Levels for all urinary analytes were normalised to creatinine. MAIN OUTCOME MEASURES: Urinary soluble endoglin, and the soluble fms-like tyrosine kinase 1 to PlGF ratio. RESULTS: In healthy controls, urinary soluble endoglin levels were increased significantly at term relative to those earlier in gestation. Severe PE was characterised by an increased urinary level of soluble endoglin, soluble fms-like tyrosine kinase 1, protein to creatinine ratio and soluble fms-like tyrosine kinase 1 to PlGF ratio compared with all other groups. There was a direct correlation between urinary soluble endoglin and proteinuria that remained after GA correction (R = 0.382, P < 0.001). Urinary soluble endoglin could not differentiate mild PE from severe preterm PE. Overall, soluble endoglin had the ability to discriminate PE from chronic hypertension and healthy controls only in women who were evaluated at <37 weeks of GA. The sensitivity, specificity and accuracy of urinary soluble endoglin alone in the diagnosis of PE or in the identification of women with PE requiring a mandated delivery before 37 weeks of gestation were 70%, 86% and 76%, respectively. These values were inferior to those of the soluble fms-like tyrosine kinase 1 to PlGF ratio (P < 0.001). The addition of urinary soluble endoglin did not improve the diagnostic accuracy of the soluble fms-like tyrosine kinase 1 to PlGF ratio alone. CONCLUSIONS: We have provided evidence that soluble endoglin is present and elevated in the urine of women who develop preterm PE. Urinary soluble endoglin has only limited ability to determine the severity of PE and to distinguish between PE and chronic hypertension both preterm and at term. Compared with urinary soluble endoglin, the soluble fms-like tyrosine kinase 1 to PlGF ratio remains a better marker of disease presence, severity and outcome.