Table 2.

Summary of primary prevention studies in the general population (39 studies)

Item	Category	Number of studies (%)
Study design	RCT	4/39 (10.3 %)
	Prospective cohort study	25/39 (64.1 %)
	Nested case-control study	10/39 (25.6 %)
Follow-up	1 yr to <2 yrs	1/39 (2.6 %)
	2 yrs to <5 yrs	4/39 (10.3 %)
	5 yrs to <10 yrs	20/39 (52.3 %)
	10 yrs+	11/39 (28.2 %)
	Not reported or unclear	3/39 (7.7 %)
Gender	Males only	13/39 (33.3 %)
	Females only	2/39 (5.1 %)
	Mixed males and females	24/39 (61.5 %)
Age	<65 yrs	36/39 (92.3 %)
	Elderly ≥ 65 yrs	3/39 (7.7 %)
Ethnicity	Mixed	3/39 (7.7 %)
	Korean	2/39 (5.1 %)
	Taiwanese	1/39 (2.6 %)
	Japanese	1/39 (2.6 %)
	Native American Indian	1/39 (2.6 %)
	White (majority populations)	15/39 (38.5 %)
	Not reported/unclear	16/39 (41.0 %)
Model methoda	Cox proportional hazards	20/39 (51.3 %)
	Logistic regression	11/39 (28.2 %)
	Conditional and unconditional logistic regression	1/39 (2.6 %)
	Conditional logistic regression	7/39 (17.9 %)
	Discriminant analysis	0/39 (0 %)
Model variables	Includes LDL-C as model variable	19/39 (48.7 %)
	Does not include LDL-C or unclear/not reported	19/39 (48.7 %)
Lp(a) assay	Isoform dependent	1/39 (2.6 %)
	Isoform independent	12/39 (30.8 %)
	Isoform independent and dependent	1/39 (2.6 %)
	Isoform independence NR or unclear	25/39 (61.4 %)
Sample type	Fresh plasma samples	5/39 (12.8 %)
	Frozen plasma samples	25/39 (64.1 %)
	Mixture of frozen and fresh samples	2/39 (5.1 %)
	Not reported or unclear	7/39 (17.9 %)
Risk of bias	Low	2/39 (5.1 %)
	Moderate	22/39 (56.4 %)
	High	7/39 (17.9 %)
	Not enough information	8/39 (20.5 %)

LDL-C low density lipoprotein; Lp(a) lipoprotein (a); NR not reported; RCT randomised controlled trial; yrs years

^aNote some studies report multiple types of models