Population risk factor estimates for abdominal aortic aneurysm from electronic medical records: a case control study

Table 2 Risk factor estimates for AAA

Variable	Meta^*estimate	Meta^*SE	P ^†	OR	95% CI
Peripheral artery disease	1.49	0.0040	3.71E-16	4.42	3.08–6.35
Smoking, ever/never	1.36	0.0036	3.92E-15	3.91	2.77–5.53
Coronary stenosis	1.07	0.0032	2.01E-12	2.91	2.15–3.93
Type 2 diabetes	−0.84	0.0043	1.69E-05	0.43	0.29–0.64
Systolic blood pressure	0.03	0.0001	2.20E-07	1.03	1.02–1.05
Diastolic blood pressure	−0.05	0.0003	1.84E-05	0.95	0.93–0.98
Age	0.05	0.0001	1.63E-15	1.05	1.04–1.06
Weight	−0.01	3.94E-05	9.13E-07	0.99	0.98–0.99
Height	0.12	0.0001	2.78E-05	1.13	1.06–1.20
Sex	0.65	0.0044	0.00172	1.92	1.24–2.97
Benign neoplasm	−0.39	0.0029	0.00263	0.67	0.51–0.89
Pulmonary disease	0.54	0.0035	0.00061	1.72	1.24–2.38
Hypertension	0.62	0.0038	0.00056	1.86	1.28–2.72
Myelogenous neoplasm	−0.32	0.0032	0.02131	0.73	0.54–0.99

Controls chosen randomly from all non-AAA subjects (2,500 iterations, 500 cases, 2,000 controls), but weighted on sex and age (5 y classes) to match census population structure for the GHS catchment area. For outline of the study design, see Figure 1.
Age classes below 35 were consolidated into a single age class to avoid oversampling the sparse classes of [18, 20, 25] and [25, 30]. These classes are typically underrepresented in physician visits.
*Meta.Est is the meta-analysis estimate (random effects, variance weighted, aggregate) of the betas from bootstrap iterations, Meta SE is the standard error of this estimate.
^†Two-sided P-value.

ISSN: 1471-2261