Table 2.

Conclusion and strength of the evidence.

	A	B	C	D	E	F	G	H
Prognostic Biomarker	No. of studies with all 3 performance indicators satisfied	No. of pooled meta-analyses showing significant association	No. of pooled meta-analyses showing significant association (excluding high risk of bias)	Non-pooled analyses showed that >=75% of studies had significant association	Persistent association on sensitivity analysis for non-pooled analyses (excluding high risk of bias)	Consistency in definition of prognostic biomarker used in analysis	Conclusion (predictive utility)	Strength of Evidence
NT-proBNP	3	2/2	2/2	Yes	Yes	Yes	High	High
CRP	0	1/2	1/2	No	No	No	Low	Moderate
TnT	4	2/3	1/3	Yes	Yes	No	High	Moderate
CACS	0	0	0	Yes	No	No	Low	Low
CCTA	2	0	0	Yes	No	No	Moderate	Low
SPECT	1	0	0	Yes	No	No	Moderate	Low
PWV	0	1/1	1/1	Yes	Yes	No	Moderate	Moderate
Gal-3	0	0	0	Yes	Yes	No	Low	Low
TnI	1	0	0	No	No	No	Low	Low
Carotid plaque	0	0	0	Yes	Yes	No	Low	Low
GDF-15	0	0	0	No	No	No	Low	Low
TyG	1	1/1	NA	Yes	No	Yes	High	Moderate
GRS-CHD	1	1/1	1/1	Yes	No	No	High	Moderate

Table includes genetic and non-genetic biomarkers showing improvement in prediction performance (i.e. at least one of c-statistic, NRI, IDI) and in more than 1 study (corresponding to all markers in green boxes in Fig. 4 and Fig. 5). Notes: Predictive utility was classified as high (3 point) moderate (2 points) or low (<2 points) based on the criteria defined in columns A, B and D (Column A: >0 = 1 point; Column B: >0 = 1 point, Column D: “Yes” = 1 point). Strength of Evidence was classified as High (4 points), moderate (2 or 3 points) and low (<2 points) based on criteria defined in columns B, C, E and F (Column B: at least one meta-analysis conducted [regardless of outcome] = 1 point; Column C: exclusion of high risk of bias studies did not alter inferences from meta-analyses [same number as in column B] = 1 point; Column E: exclusion of high risk of bias studies did not alter inferences from non-pooled analyses [YES] = 1 point; Column F: “yes” = 1 point).