Table 3

Logistic regression analysis for incident ischaemic heart disease by H pylori strain

AdjustmentHp+ and CagA− v Hp−Hp+ and CagA+ v Hp−Test3-150 for difference between odds ratios (Ho:OR1 = OR2)
OR195% CIOR295% CI
All available data (n=370) 1.27(0.71 to 2.26)1.33(0.81 to 2.18)0.86
(1) Unadjusted
Data restricted to subjects with complete information on confounders (n=354)
(1) Unadjusted1.26(0.70 to 2.28)1.31(0.78 to 2.17)0.89
(2) Model 1 plus age, BMI, systolic blood pressure, smoking history, and total cholesterol1.07(0.57 to 2.03)1.21(0.71 to 2.07)0.66
(3) Model 2 plus own and father's social class1.19(0.61 to 2.29)1.33(0.77 to 2.30)0.69
(4) Model 3 plus height and lung function1.17(0.61 to 2.28)1.28(0.74 to 2.23)0.76
(5) Model 4 plus fibrinogen and white cell count1.25(0.64 to 2.46)1.38(0.78 to 2.43)0.75
  • 3-150 Based on difference in fit between logistic regression models, one containing a two level factor for H pylori (Hp+, Hp−) and one containing a three level factor forH pylori (Hp−, Hp+/Cag A−, Hp+/Cag A+).

  • BMI, body mass index; CI, confidence interval; OR, odds ratio.