Table 1.

DCCT/EDIC study naïve and genome‐wide bootstrap bias‐reduced logHR estimates (taken in absolute value)

				Univariable models
Gene		SNP	MAF (%)	P‐value	Naïve logHR estimate	Genome‐wide bootstrap estimate	Percentage reduction of naïve by genome‐wide bootstrap	Power by naïve (%)	Power by genome‐wide bootstrap (%)
1	LIPC	rs1968685	48.3	2.22 × 10−4	0.51	0.29	43.1	90.1	35.3
2	LIPC a	rs7178362	13.7	2.36 × 10−4	0.56	0.22	60.1	63.4	7.7
3	FLT1	rs7999615	6.2	6.13 × 10−4	0.75	0.27	64.0	57.3	5.7
4	SOD1 a	rs17880135	5.6	8.18 × 10−4	0.69	0.17	75.4	43.2	2.4
5	VDR	rs2254210	36.3	2.61 × 10−3	0.39	0.16	59.0	60.4	7.9
6	PRKCB1	rs411103	39.2	3.82 × 10−3	0.41	0.16	61.0	67.7	8.2
7	FLT4	rs307806	14.9	3.85 × 10−3	0.46	0.12	73.9	46.4	3.8
8	MMP2	rs17859935	16.9	4.36 × 10−3	0.46	0.14	69.6	53.0	3.8
9	ABCA1 a	rs2472448	10.6	5.00 × 10−3	0.86	0.46	46.5	91.5	32.2
10	VEGFA	rs3025035	7.2	5.34 × 10−3	0.58	0.13	77.6	35.4	1.9
11	PON1 a	rs854555	34.4	6.22 × 10−3	0.41	0.17	58.5	64.7	8.8
12	UCP1	rs2043125	26.5	6.52 × 10−3	0.46	0.17	63.0	69.3	7.5
13	PARP1 a	rs3219065	15.9	7.75 × 10−3	0.41	0.09	78.0	38.0	1.9
14	PARP1 a	rs2027440	15.9	8.18 × 10−3	0.41	0.08	80.5	38.0	1.6
15	MMP2	rs17859970	14.8	9.99 × 10−3	0.43	0.10	76.7	40.8	2.1

^aMinor allele is the risk allele, otherwise major allele is associated with risk.

Analysis of time to severe nephropathy (1,361 individuals with 115 events) based on a significance threshold selection criterion of P < 0.01. SNPs with MAF ≤ 5% are excluded. Bias reduction ranges from 43.1% to 80.5%. The final two columns display post hoc power calculations for a similar sample assuming logHRs are set to the value of the naïve or the genome‐wide bias‐reduced estimates.