Serum metabolites predict response to angiotensin II receptor blockers in patients with diabetes mellitus

Table 3 Risk prediction of change in UAE in response to ARB therapy

	R²	p value*	Discrimination of >30 % decrease in UAE
	R²	p value*	ROC	95 % CI		p value*	IDI	95 % CI		p value*
Discovery cohort
Clinical parameters^a	0.10	Ref.	0.72	0.57	0.87	Ref.		Ref.
+Serum metabolites classifier	0.70	<0.001	0.95	0.89	1.00	0.001	0.50	0.36	0.63	<0.001
+Subset of 7 metabolites^b	0.50	<0.001	0.90	0.81	0.99	0.012	0.33	0.19	0.47	<0.001
Validation Cohort
Clinical parameters^a	0.20	Ref.	0.74	0.59	0.89	Ref.		Ref.
+Serum metabolites classifier	0.53	<0.001	0.89	0.79	0.99	0.063	0.30	0.15	0.46	<0.001
+Subset of 7 metabolites^b	0.35	0.002	0.78	0.64	0.92	0.460	0.19	0.09	0.47	0.055

* Comparing clinical parameters + metabolites to only clinical parameters
^aBaseline Age, Sex, SBP, HbA1c, GFR, UAE
^bMetabolites assigned to both drug interference:direct disease phenotype and disease progression processes (seven metabolites: ADMA, citrulline, lysoPC a C16:0, lysoPC A C16:1, PC aa C36:0, PC aa C42:2, tryptophan)

ISSN: 1479-5876