Table 1.

Associations for BMD-related and non-BMD-related genome-wide significant SNPs for hip fractures

	Non-BMD-related locus			BMD-related loci												n
	APOE (19q13.32)			ETS2 (21q22.2)			SALL1 (16q12.1)			REST (4q12)			HOXC8 (12q13.13)
	rs429358, (EA C, EAF 0.17)			rs11088458 (EA G, EAF 0.70)			rs62028332 (EA G, EAF 0.87)			rs35339719 (EA G, EAF 0.74)			rs4142680 (EA T, EAF 0.41)a
Fractures	OR	95% CI	p	OR	95% CI	p	OR	95% CI	p	OR	95% CI	p	OR	95% CI	p	Cases	Controls
Hip fracture	1.14	(1.10–1.19)	3.8 × 10−11	1.11	(1.07–1.14)	3.7 × 10−10	1.15	(1.10–1.20)	1.4 × 10−9	1.11	(1.07–1.14)	1.8 × 10−9	1.12	(1.08–1.16)	2.2 × 10−9	11,516	723,838
Old	1.33	(1.25–1.43)	3.1 × 10−17	1.11	(1.05–1.18)	1.3 × 10−4	1.18	(1.10–1.26)	6.5 × 10−6	1.08	(1.02–1.14)	7.2 × 10−3	1.09	(1.03–1.14)	1.8 × 10−3	4,700	609,945
Young	1.03	(0.97–1.10)	3.0 × 10−1	1.10	(1.05–1.16)	1.9 × 10−4	1.15	(1.07–1.23)	5.6 × 10−5	1.15	(1.09–1.21)	9.8 × 10−8	1.14	(1.09–1.20)	1.6 × 10−7	4,334	609,945
Men	1.14	(1.06–1.23)	3.9 × 10−4	1.09	(1.03–1.16)	5.8 × 10−3	1.08	(1.00–1.17)	5.8 × 10−2	1.11	(1.04–1.17)	1.3 × 10−3	1.12	(1.05–1.18)	2.5 × 10−4	3,070	266,328
Women	1.17	(1.11–1.24)	3.8 × 10−8	1.12	(1.07–1.17)	3.3 × 10−6	1.21	(1.14–1.28)	1.6 × 10−9	1.11	(1.06–1.17)	4.0 × 10−6	1.12	(1.07–1.17)	7.6 × 10−7	5,980	343,617
Fracture at any bone site	1.01	(0.99–1.02)	6.0 × 10−1	1.05	(1.03–1.06)	4.5 × 10−11	1.04	(1.03–1.06)	6.9 × 10−6	1.02	(1.00–1.03)	8.9 × 10−3	1.01	(0.99–1.02)	2.4 × 10−1	53,184	373,611
Other binary outcomes
Falls	1.02	(1.01–1.03)	6.3 × 10−4	1.01	(1.00–1.02)	8.5 × 10−2	1.00	(0.98–1.01)	5.2 × 10−1	1.00	(0.99–1.01)	8.3 × 10−1	1.01	(1.00–1.02)	1.0 × 10−1	89,076	362,103
Alzheimer′s disease	3.33	(3.20–3.45)	1.2 × 10−881	0.98	(0.95–1.01)	2.3 × 10−1	0.99	(0.94–1.03)	5.4 × 10−1	1.00	(0.97–1.03)	8.3 × 10−1	1.02	(0.98–1.05)	3.3 × 10−1	35,274	59,163
BMD-related traits	Beta	SE	p	Beta	SE	p	Beta	SE	p	Beta	SE	p	Beta	SE	p	n
FN-BMD	−0.010	0.011	4.0 × 10−1	−0.030	0.008	4.1 × 10−4	−0.060	0.012	3.0 × 10−7	0.007	0.008	4.1 × 10−1	−0.04	0.012	9.8 × 10−4	49,998
LS-BMD	0.006	0.013	6.3 × 10−1	−0.025	0.010	1.1 × 10−2	−0.023	0.013	9.0 × 10−2	0.016	0.010	1.1 × 10−1	−0.057	0.012	6.0 × 10−6	44,731
eBMD	0.001	0.003	7.9 × 10−1	−0.044	0.002	6.8 × 10−82	−0.041	0.004	5.2 × 10−20	−0.015	0.002	1.8 × 10−11	−0.002	0.003	6.3 × 10−1b	426,824

Besides the data for the five identified genetic signals in the present hip fracture GWAS meta-analyses, look ups were performed in the following published GWASs: fracture at any bone site,¹¹ falls,¹⁵ Alzheimer′s disease,¹⁴ eBMD,¹¹ and FN-BMD and LS-BMD.¹² Odds ratios (OR) for binary outcomes are given per effect allele. Betas for continuous BMD-related parameters are expressed as SD per effect allele. EA, effect allele; EAF, effect allele frequency; BMD, bone mineral density; FN, femoral neck; LS, lumbar spine; eBMD, estimated bone mineral density in the heel using ultrasound. To achieve effect estimates not confounded by a possible minor dilution by diaphyseal and distal femur fractures and lack of adjustment for height and weight in the publicly available analyses of the FinnGen cohort, we replicated the five genome wide signals for hip fracture in a meta-analyses excluding the FinnGen cohort, yielding similar effect estimates (meta-analyses results excluding FinnGen; APOE, rs429358 OR 1.16 95% CI 1.11–1.21, p = 2.5 × 10⁻¹⁰; ETS2, rs11088458 OR 1.11 95% CI 1.07–1.15, p = 7.4 × 10⁻⁸; in the ETS2 locus, the correlated SNP rs8130983 [r² = 0.73] was genome-wide significant in analyses without FinnGen OR 1.12 95% CI 1.08–1.16, p = 1.1 × 10⁻⁸; SALL1, rs62028332 OR 1.16 95% CI 1.10–1.22, p = 3.1 × 10⁻⁹; REST, rs35339719 OR 1.11 95% CI 1.07–1.15, p = 3.4 × 10⁻⁸; HOXC8, rs4142680 OR 1.12 95% CI 1.08–1.16, p = 2.2 × 10⁻⁹).

^aInformation for rs4142680 and hip fractures was only available in HUNT, UK Biobank, and UFO (8,401 cases and 501,168 controls).

^bDerived from Neale et al. (http://www.nealelab.is/uk-biobank/), n = 206,496 as this SNP was not present in Zheng et al.¹²