Endotrophin is a risk marker of complications in CANagliflozin cardioVascular Assessment Study (CANVAS): a randomized controlled trial

Table 2 Association of P-ETP with Outcomes in the CANVAS Study

	Variables retained	HR_adj (95% CI) for P-ETP^a	P	C index (clinical covariates)	C index (P-ETP)	C index combined (∆C)
HF	P-ETP, age, sex, BMI, HbA1c, dBP, sBP, ACR, NTproBNP, prior history of HF, prior history of CV disease	1.57 [1.15–2.14]	0.005	0.823	0.679	0.828 (0.005)
CVD	P-ETP, age, LDL, ACR, NTproBNP, prior history of HF, prior history of CV disease	1.52 [1.23–1.88]	0.0001	0.755	0.646	0.757 (0.002)
HFCVD	P-ETP, age, BMI, HbA1c, ACR, treatment, NTproBNP, smoking, prior history of HF, prior history of CV disease	1.49 [1.23–1.79]	< 0.0001	0.771	0.651	0.776 (0.005)
Mortality	P-ETP, age, BMI, dBP, LDL, ACR, NTproBNP, smoking, prior history of CV disease	1.42 [1.19–1.70]	0.0001	0.720	0.624	0.724 (0.004)
KCOM1	P-ETP, ACR, treatment, NTproBNP	1.54 [1.15–2.06]	0.004	0.774	0.631	0.773 (− 0.001)
KCOM2	P-ETP, age, BMI, LDL, ACR, treatment, NTproBNP, prior history of CV disease	1.45 [1.22–1.74]	< 0.0001	0.757	0.639	0.760 (0.003)
KCOM3	ACR, treatment	1.26 [1.06–1.48]^a	0.008	0.778	0.583	0.777 (− 0.001)

HR_adj, adjusted HR per doubling (log2) of P-ETP. Covariates in the initial model included age, sex, body mass index (BMI), diabetes duration, HbA1c, LDL, diastolic and systolic blood pressure (BP), eGFR, albumin:creatinine ratio (ACR), treatment, NTproBNP, smoking, prior history of HF, and prior history of CV disease
KCOM1 comprised a 40% decrease in eGFR (sustained for two or more consecutive measures), the need for RRT (dialysis or transplantation), or renal death, KCOM2 comprised KCOM1 and CVD, KCOM3 comprised KCOM1 and progression to macroalbuminuria, HF heart failure, BMI body mass index, dBP diastolic blood pressure, sBP systolic blood pressure, ACR albumin:creatinine ratio, CVD cardiovascular death, HFCVD composite of HF and CVD, KCOM kidney composite endpoint
^aUsing the Akaike information criterion in backwards selection, covariates for adjustment of P-ETP were selected. The variables retained in the final model were listed in the column “Variables retained”. C index values are given for clinical covariates only (“C index (clinical covariates only)”, only P-ETP (“C index (P-ETP)”), combined (“C index combined”), and the improvement in the C index when adding P-ETP to the base model only including clinical covariates (“∆C”). ^aDespite the significance, P-ETP was not retained in the final model after the bootstrap analyses

ISSN: 1475-2840