Skip to main content

Table 2 Overview of results of meta-analyzed models

From: Depression, cardiometabolic disease, and their co-occurrence after childhood maltreatment: an individual participant data meta-analysis including over 200,000 participants

Model specification

k

Outcome levels

n

Pooled OR [95% CI]

Main models

Binomial regressions

1. Association of childhood maltreatment with depressiona

13

No depression

115,959

ref

Depression

39,910

2.82 [2.40; 3.30]

2. Association of childhood maltreatment with cardiometabolic diseasea

13

No cardiometabolic disease

162,511

ref

Cardiometabolic disease

15,421

1.34 [1.23; 1.46]

2a. Outcome: diabetes mellitus1

12

No diabetes

168,943

ref

Diabetes

7330

1.35 [1.19; 1.53]

2b. Outcome: cardiovascular diseasea

11

No cardiovascular disease

165,502

ref

Cardiovascular disease

9216

1.39 [1.26; 1.54]

Multinomial logistic regressions

3. Association of childhood maltreatment with comorbidity statusa

9

Healthy controls

99,191

ref

Depression only

34,360

2.68 [2.39; 3.00]

Cardiometabolic disease only

9051

1.27 [1.18; 1.37]

Comorbidity

3565

3.04 [2.51; 3.68]

4. Model 3 + adjustment of lifestyle factorsb

9

Healthy controls

90,166

ref

Depression only

30,544

2.57 [2.28; 2.89]

Cardiometabolic disease only

8181

1.25 [1.15; 1.36]

Comorbidity

3175

2.99 [2.46; 3.63]

5a. Model 3 in males onlya

7

Healthy controls

46,327

ref

Depression only

10,402

2.92 [2.51; 3.40]

Cardiometabolic disease only

5314

1.13 [1.01; 1.27]

Comorbidity

1582

2.90 [2.37; 3.53]

5b. Model 3 in females onlya

8

Healthy controls

49,571

ref

Depression only

23,167

2.59 [2.30; 2.91]

Cardiometabolic disease only

3438

1.41 [1.26; 1.57]

Comorbidity

1942

3.49 [2.79; 4.36]

6. Model 3 with predictorsa = 

Physical abuse

6

Healthy controls

88,347c

ref

Depression only

32,198c

1.53 [1.41; 1.66]

Cardiometabolic disease only

8481c

1.45 [1.24;1.71]

Comorbidity

3411c

2.06 [1.78; 2.39]

Emotional abuse

Healthy controls

88,347c

ref

Depression only

32,198c

2.87 [2.56; 3.22]

Cardiometabolic disease only

8481c

1.29 [1.11; 1.49]

Comorbidity

3411c

[2.16; 3.83]

Sexual abuse

Healthy controls

88,347c

ref

Depression only

32,198c

1.66 [1.56; 1.76]

Cardiometabolic disease only

8481c

1.09 [1.01; 1.17]

Comorbidity

3411c

1.60 [1.19; 2.15]

7. Model 3 with cumulation of maltreatment types as predictora = 

1 type (vs. 0 type)

6

Healthy controls

88,347c

ref

Depression only

32,198c

2.32 [2.22; 2.41]

Cardiometabolic disease only

8482c

1.17 [1.07; 1.28]

Comorbidity

3411c

2.37 [1.87; 3.01]

2 or more types (vs. 0 type)

Healthy controls

88,347c

ref

Depression only

32,198c

5.14 [3.93; 6.72]

Cardiometabolic disease only

8482c

1.83 [1.49; 2.26]

Comorbidity

3411c

5.96 [3.59; 9.90]

Sensitivity analyses—multinomial logistic regressions

8. Model 3 with outcome = comorbidity status based on strict definition of cardiovascular diseasea

9

Healthy controls

101,203

ref

Depression only

35,279

2.66 [2.37; 2.98]

Cardiometabolic disease only

7030

1.28 [1.18; 1.40]

Comorbidity

2643

3.09 [2.54; 3.75]

9. Model 3 with outcome = comorbidity status based on broad definition of cardiovascular diseasea

11

Healthy controls

81,817

ref

Depression only

28,612

2.84 [2.41; 3.35]

Cardiometabolic disease only

27,178

1.11 [1.06; 1.17]

Comorbidity

10,364

3.00 [2.69; 3.36]

10. Model 3 with outcome = comorbidity status based on medication intake in addition to reports of diagnosesa

7

Healthy controls

78,989

ref

Depression only

31,266

2.83 [2.27; 3.53]

Cardiometabolic disease only

16,896

1.13 [1.00; 1.26]

Comorbidity

6517

2.76 [2.28; 3.35]

  1. aMinimal adjustment: correction for age, sex, and education level
  2. bMinimal + additional adjustment: correction for age, sex, education level, smoking status, alcohol consumption, and physical activity
  3. cThese sample sizes are specific to each outcome level and across all predictors and predictor levels
  4. Abbreviations: k number of studies included, n sample size used in model, OR Odds ratio, CI Confidence interval, ref outcome reference category in logistic regressions
  5. An overview of the cohorts included in each meta-analyzed model can be found in Additional file 1: Table S6