Table 3.

Associations of CYP17 common haplotypes with prostate cancer in the BPC3

Haplotype copiesCases (n = 8,138)*Controls (n = 9,033)*OR (95% CI)P heterogeneity
Block 1
CTAGG03,2873,5861.00 (ref.)0.33
13,6964,1230.99 (0.93-1.06)
21,1561,3240.98 (0.89-1.08)
P trend = 0.27
TTGAG04,4574,8741.00 (ref.)0.99
13,1283,4970.98 (0.92-1.05)
25536620.91 (0.80-1.03)
P trend = 0.16
CCAGT05,2796,0111.00 (ref.)0.25
12,5042,6571.06 (0.99-1.13)
23553641.06 (0.91-1.23)
P trend = 0.02
TTGGG06,1716,8761.00 (ref.)0.95
11,8041,9791.02 (0.94-1.10)
21631781.03 (0.82-1.29)
P trend = 0.65
Global test block 1: χ2 = 3.62, 4 df, P = 0.460
Block 2
TAAA04,6175,0481.00 (ref.)0.84
13,0063,3750.98 (0.92-1.05)
25156100.91 (0.80-1.04)
P trend = 0.17
CACA04,9685,5181.00 (ref.)0.69
12,7753,0651.01 (0.94-1.08)
23954500.96 (0.83-1.11)
P trend = 0.92
CACG05,4086,1821.00 (ref.)0.29
12,3892,5281.07 (0.99-1.14)
23403231.14 (0.97-1.34)
P trend = 0.002
CAAA05,3185,9071.00 (ref.)0.92
12,3962,6011.04 (0.97-1.12)
24245250.93 (0.80-1.07)
P trend = 0.53
CGCA06,0316,6551.00 (ref.)0.33
11,9282,1720.98 (0.91-1.05)
21792061.00 (0.81-1.23)
P trend = 0.50
Global test block 2: χ2 = 7.59, 5 df, P = 0.180
  • * Numbers do not sum to total due to missing genotype data.

  • The ORs and 95% CIs were estimated using conditional (on matching factors: age, ethnicity, and country within EPIC) logistic regression and adjusted for cohort.

  • P value for heterogeneity in ORs across cohorts.