Table 2 The association between mBMIΔ/quintiles of mBMIΔ and CVD outcomes (prevalent and incident) in the AusDiab cohort

From: Metabolic phenotyping of BMI to characterize cardiometabolic risk: evidence from large population-based cohorts

mBMIΔ

Prevalent CVD (n = 577 cases versus 9690 controls)

10-year incident CVE (n = 414 events versus 7936 non-events)

10-year incident IHD (n = 304 events versus 8046 non-events)

Odds ratio (95% CI)a

p value

Hazard ratio (95% CI)b

p value

Hazard ratio (95% CI)c

p value

mBMIΔ (continuous scale)

1.3 (1.1, 1.4)

3.40 × 10–5

1.11 (1.01, 1.22)

4.30 × 10–2

1.13 (1.01, 1.2)

3.6 × 10–2

Q1 (Ref)

–

–

–

–

–

–

Q2

1.4 (1.01, 2.1)

7.30 × 10–2

1.1 (0.8, 1.5)

7.00 × 10–1

1.0 (0.7, 1.5)

9.00 × 10–1

Q3

1.3 (0.9, 2.0)

1.70 × 10–1

1.0 (0.7, 1.3)

8.00 × 10–1

0.9 (0.6, 1.4)

7.00 × 10–1

Q4

1.7 (1.1, 2.4)

1.10 × 10–2

1.4 (1.02, 1.9)

4.00 × 10–2

1.4 (1.02, 2.1)

4.10 × 10–2

Q5

2.1 (1.5, 3.1)

6.40 × 10–5

1.3 (0.9, 1.7)

1.30 × 10–1

1.3 (0.9, 1.8)

2.00 × 10–1

  1. Significant two-sided p values (<0.05) are shown in bold.
  2. CVD cardiovascular disease, CVE cardiovascular event, IHD ischemic heart disease.
  3. aLogistic regression between the mBMIΔ /quintiles of mBMIΔ and prevalent CVD adjusting for age, sex, BMI, smoking status and diabetes history. Odds ratios and 95% CIs for subjects with prevalent CVD compared to control groups were computed.
  4. bProportional hazard Cox-regression between the mBMIΔ /quintiles of mBMIΔ and major incident CVE adjusting for age, sex, BMI, smoking status and diabetes history. Hazard ratios and 95% CIs for subjects with incident CVE compared to non-events are presented.
  5. cProportional hazard Cox-regression between the mBMIΔ /quintiles of mBMIΔ and incident IHD adjusting for age, sex, BMI, smoking status and familial diabetes history. Hazard ratios and 95% CIs for subjects with incident IHD compared to non-events are presented.
Back to article page