. 2011 Sep 28;2011:765071. doi: 10.4061/2011/765071

Table 3.

Cox Regression Models Predicting Mortality at 20-year followup by BMI, and age, controlling for smoking status (Ever versus Never).

Full sample

Age groups

75–84 year olds

85–94 years old

CD and NH residents

N = 1369

n = 1200

n = 809

n = 752

n = 560

n = 448

Background

Age

1.068^c

1.055–1.082

1.067^c

1.053–1.081

1.077^c

1.047–1.108

1.076^c

1.045–1.108

1.062^b

1.025–1.100

1.065^b

1.025–1.107

Gender (male = 1)

1.267^b

1.104–1.453

1.265^b

1.092–1.465

1.429^c

1.195–1.709

1.449^c

1.202–1.746

1.087

.871–1.357

1.051

0.823–1.342

Origin

East

.849

0.715–1.008

0.867

0.723–1.040

0.960

0.767–1.201

0.993

.788–1.252

0.730^a

0.559–0.955

0.715^a

0.533–0.958

West

1.030

0.887–1.197

1.042

0.887–1.224

1.024

0.846–1.240

1.033

.844–1.264

1.028

0.805–1.313

1.033

0.786–1.358

Years of education

0.994

0.982–1.007

0.996

0.982–1.009

0.998

0.981–1.015

0.999

.981–1.017

0.992

0.974–1.011

0.993

0.972–1.015

Additional income

0.940

0.829–1.070

0.932

0.811–1.071

0.844^d

0.711–1.003

0.858^d

0.717–1.026

1.040

0.845–1.279

0.993

0.791–1.247

Had children

0.705^b

0.570–0.872

0.684^b

0.538–0.869

0.661^b

.504–.867

0.629^b

0.468–0.844

0.791

0.554–1.130

0.780

0.506–1.204

Smoking status
(1= ever)

1.106

0.969–1.262

1.103

0.959–1.269

1.153^d

0.974–1.364

1.151

0.966–1.371

1.015

0.816–1.263

0.993

0.781–1.263

BMI

Obese (>30)

1.122

0.934–1.347

1.096

.902–1.332

1.291^a

1.022–1.631

1.315^a

1.032–1.676

0.946

0.704–1.273

0.853

0.614–1.187

Underweight (<22)

1.411^c

1.199–1.661

1.404^c

1.179–1.671

1.267^a

1.005–1.597

1.304^a

1.025-1.657

1.475^b

1.163–1.871

1.400^a

1.075–1.824

χ ² (df = 2)

16.346^c

13.699^b

7.003^a

7.789^a

10.852^b

8.185^a

^a P < .05, ^b P < .01, ^c P < .001.^d.1 > P > .05.

HR= Hazard Ratio= Exp (b); CI= Confidence Interval.