Table 1 Study details on all prospective studies (N = 75), subdivided by the following biological sublevels: Neuroimaging, Gastrointestinal factors, Immunology, Neurotransmitters, Neurotrophic factors, Hormones: HPA axis, HPG axis, HPS axis, and HPT axis

Bress et al. [39]^a

Y

b

USA

68

16 O

[15–17]

100

2

DISC; ICD-10

Frontal brain areas (ERP)

Activity: EEG

During reward task. comparing loss-gain

↓ Frontal ERP

7

Davey et al. [49]^a

Y

?

Australia

56

8 O

17

45

2

K-SADS-PL; DSM-IV (nm)

Amygdala-sgACC connectivity

RSFC fMRI

Resting state

↑ Amygdala–sACC connectivity

4

Foland-Ross et al. [51]^b

Y

b

USA

33

18 O

13

100

5

K-SADS-PL and SCID; DSM-IV (nm)

32 brain regions were used for support vector machine classification

Structure: MRI

Volumes

The most important classifiers for MDD were ↓ mOFC, PCG, ACC, and ↑ insula

7

Little et al. [73]^c

Y

?

Australia1

99

26 O

13

29

6

K-SADS-PL. MINI; DSM-IV (nm)

Hippocampus, amygdala, OFC and ACC

Structure: MRI

Volumes

↓ Hippocampus

NS amygdala, OFC and ACC

6

Little et al. [74]

?

?

Australia1

137

36 O

13

52

6

K-SADS-PL; DSM-IV (nm)

Hippocampus

Structure: MRI

Volumes

↓ hippocampus

5

Nusslock et al. [83]^a

?

b

USA

40

13 O

20

43

3

SADS-C; DSM-IV

Left frontal brain areas (Alpha power)

Activity: EEG

Rest. eyes-open and eyes-closed

↓ Left frontal activity

6

Papmeyer et al. [85]^a

Y

b

UK1

204

19 O

21

56

2

Clinical interview; DSM-IV (nm)

Frontal brain regions

Structure: MRI

Gray matter thickness

↓ Right parahippocampal and fusiform gyrus

9

Papmeyer et al. [86]^a

Y

b

UK1

204

19 O

21

56

2

Clinical interview; DSM-IV

Subcortical brain regions

Structure: MRI

Gray matter thickness

NS

9

Whalley et al. [99]^a

Y

b

UK1

156

20 O

21

52

2

SCID ; DSM-IV

Brain regions activated by the task

Activity: fMRI

Cognitive sentence completion task: increase in difficulty

↑ Insula activity

7

Whalley et al. [100]^a

Y

b

UK1

50

11 O

23

Unclear

2

SCID; DSM-IV

Amygdala, insula, hippocampus, ACC, thalamus

Activity: fMRI

View emotional images vs neutral

↑ Thalamus, insula, ACC activity

7

Nickson et al. [81]^c

N

b

UK1

131

30 O

21

69

4.9

SCID; DSM-IV

VBM

Structure: MRI

Whole brain VBM

↑Amygdala gray msatter volume

7

Macoveanu et al. [77]^b

Y

b

Denmark

85

12 O

39

65

7

SCAN 2.1; (ICD-10)

VBM

Structure: MRI

Whole brain VBM

↑ACC gray matter volume

5

Belden et al. [36]^a

Y

?

USA

129

24 O + R

[6–12]

48

10

Clinician and TRD; DSM-5

Anterior insula

Structure: MRI

Volumes

↓ Insula volume

6

Rao et al. [91]^a

Y

a

USA

83

29 O + R

15

58

5

LIFE. PSR; DSM-IV (at baseline)

Hippocampus

Structure: MRI

Volumes

↓ Hippocampus

5

Serra-Blasco et al. [95]^a

N

a

Spain

24

10 O + R

48

75

5

Life-Chart Manual for Recurrent Affective Illness; Unclear

Whole brain (Freesurfer)

Structure: MRI

Volumes

↓rIFG, ACC, rMFG volumes predict recurrence

8

Allen et al. [31]^a

Y

b

USA

9

3 R

49

61

0.5

SCID; DSM-III-R

Frontal asymmetry

Activity: EEG

Rest. pre- post-tryptophan depletion

↑ Right frontal activity after vs before TD predict lower change of recurrence at follow-up

6

Farb et al. [50]^a

Y

b

Canada

16

10 R

39

69

1.5

SCID; DSM-IV

ROI mPFC

Activity: fMRI

Viewing sad vs natural movie clips

↑ mPFC activity

7

Frodl et al. [53]^c

N

a

Germany1

30

12 R

48

60

1.5

Clinical interview, 2 psychiatrists; DSM-IV

Hippocampus and amygdala

Structure: MRI

Volumes

↓ Hippocampus

NS amygdala

6

Frodl et al. [54]

N

a

Germany1

30

13 R

45

63

3

SCID; DSM-IV

Hippocampus and amygdala

Structure: MRI

Volumes

NS

6

Kronmüller et al. [70]^b

N

a

Germany

57

21 R

44

58

2

SCID / LIFE; DSM-IV

Hippocampus

Structure: MRI

Volumes

↓ Hippocampus

5

Lythe et al. [76]^a

Y

b

UK

95

25 R

34

64

1.2

LIFE; DSM-IV

ROI based on previous studies: ACC, temporal, striatal

Activity: fMRI

Activity during self-blame versus other-blame emotions

↑ sgACC and temporal regions connectivity and putamen and claustrum connectivity in recurrent

9

Nixon et al. [117]^a

Y

b

UK

38

7 R

[24–63]

33

1

SCID; DSM-IV

ROI based on previous studies

Activity: fMRI

Go/Nogo task

↓ Right dmPFC activity following errors and negative feedback compared to correct hits in recurrence vs other groups

6

Workman et al. [101]^a

Y

?

UK

85

17 R

37

64

1.2

SCID; DSM-IV

sgACC

RSFC: fMRI

Rest. Left sgACC to right sgACC connectivity

Recurrent group was intermediate to resilient and control in sgACC connectivity

9

Langenecker et al. [71]^a

Y

b

USA

94

21 R

21

63

4–6

DIGS and longitudinal interval follow evaluation

fMRI

Activity: fMRI & RSFC fMRI

Go/No-Go task: successful vs unsuccessful inhibition

Successful regulation sgACC hyperactivation. Failed regulation; MFG hypoactivation. RSFC: altered MFG and sgACC connectivity.

8

Gastrointestinal factors

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/recurrence of MDD

QA score

Campo et al. [40] °

Y

?

USA

119

14 O

[6–12]

?

7.5

K-SADS-E; DSM-IV (nm)

L-5-hydroxytryptophan injection at baseline

Abdominal discomfort, nausea, or vomiting in response to the L-5HTP infusion; GI distress

L-5HTP infusion; Survival curve (groups high vs. low sensitive)

↑GI distress after serotonin challenge

6

Immunology

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Chocano-Bedoya et al. [43]^c

N

?

USA

4403

81 O

56

100

12

Self-report of diagnosis; Unclear

CRP, IL-6, TNFα-R2

Blood, 1 time point

hs-CRP IA. EIA

NS

4

Haastrup et al. [61]^a

N

?

Denmark

9275

22 O

[18–47]

48

5

prescriptions for antidepressant medication or had a hospital discharge diagnose with the codes F.32 or F.33; ICD-10

suPAR

Blood plasma, 1 time point

ELISA

↑Increased risk (shorter time) to onset of depression with higher SuPAR

6

Rudaz et al. [94]

Y

?

Switzerland1

1524

192 O

51

43

5.5

Diagnostic Interview for Genetic Studies (DIGS); DSM-IV

CRP, IL-1ß, IL-6, TNFα

Blood serum, 1 time point

hs-CRP IA. multiplexed particle-based flow cytometric cytokine assay

↓TNFα, NS for CRP, IL-1ß, IL-6

8

Copeland et al. [46]^c

Y

b

USA

5810

169 O + R

14

49

12

CAPA, YAPA; DSM-IV

CRP

Blood spot, 1 time point

hs-CRP IA

NS

9

Glaus et al. [55]^c

Y/N

?

Switzerland1

2580

608 O + R

[35–66]

54

5.5

DIGS

CRP, IL-6, TNFα

Serum, 1 time pont

IA and latex HS

O + R: ↓TNFα

7

Khandaker et al. [69]^c

N

?

UK

2447

422 O + R

9

51

9

CIS-R; ICD-10

CRP, IL-6

Blood serum, 1 time point

hs-CRP essay. ELISA

NS

5

Pasco et al. [87]^a

Y

?

Australia

822

151 O + R

49

100

10

SCID; DSM-IV-TR

CRP

Blood serum, 1 time point

hs-CRP IA

↑ CRP reduced time till depression (HR) 

7

Neurotransmitters

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Johnston et al. [68]^a

N

a

UK

31

20 R

47

71

8

SCID I/P; DSM-III-R

Plasma Norepinephrine, cortisol

Blood plasma, 1 time point

Chromatography, RIA

↓ Norepinephrine, shorter time to first recurrence.

6

Neurotrophic factors and oxidative stress

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/recurrence of MDD

QA score

Pasquali et al. [88]^a

N

b

USA

148

37 O

40

100

3

SCID; DSM-IV

Neurotrophic: BDNF Immunology: HSP70, 3-Nitrotyrosine,, Oxidative stress: Protein carbonyl, Lipid peroxidation, Thiol content

Blood serum, 1 time point

ELISA, quantitative sandwich enzyme immunoassay, colorimetric assay (thiol)

↑ heat-shock protein 70, 3-nitrotyrosine, protein carbonyl, and lipid peroxidation ↓BDNF

6

Vinberg et al. [97]^a

Y

?

Denmark

234

24 O

44

65

7.5

SCAN; ICD 8/ICD-10

BDNF

Blood, 1 time point

Two-site sandwich ELISA

NS

5

Hormones: HPA

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Adam et al. [30]^c

N

?

USA

230

18 O

17

75

1

SCID; DSM-IV

AUC, CAR, diurnal, slope

Saliva, 3 days, 6 times a day

DELFIA

NS

7

Colich et al. [45]^c

Y

b

USA1

89

31 O

12

100

6

K-SADS. SCID; DSM-IV

Cortisol pre- post- trier social stress test

Saliva, 4 time points

LIA

↓ cortisol reactivity in early puberty ↑cortisol reactivity in late puberty

9

Goodyer et al. [56]

?

?

UK2

171

30 O

14

59

1

K-SADS; DSM-IV

Peaks 8:00, cortisol and DHEA

Saliva 4 days, 2 time points

ELISA

↑ cortisol & DHEA

6

Goodyer et al. [57]

?

?

UK2

234

31 O

14

53

1

K-SADS; DSM-IV

Morning and evening cortisol, DHEA

Saliva 4 days, 2 time points

ELISA

↑ DHEA, NS cortisol

7

Goodyer et al. [58]^c

?

b

UK2

367

32 O

14

46

1

K-SADS; DSM-IV

Morning cortisol

Saliva 4 days, 1 time point

ELISA

↑ cortisol

6

Goodyer et al. [59]

?

b

UK2

357

40 O

14

47

1

K-SADS; DSM-IV

Morning cortisol

Saliva 4 days, 1 time point

ELISA

↑ cortisol

6

Grynderup et al. [60]^c

Y

?

Denmark

2920

62 O

56

78

2

SCAN; ICD-10-DSR

Morning, diurnal, evening, morning to evening slope in cortisol

Saliva, 2 hours after awakening, and between 5PM and 4AM

RIA

↓ difference in morning to evening cortisol

9

Harris et al. [63]^c

?

?

UK

116

28 O

39

100

1

SCAN; DSM-IV

DHEA and cortisol

Saliva, 4 days, 2 time points

ELISA

NS

6

Herbert et al. [65]^c

Y

?

UK

279

53 O

37

100

1

SCAN; DSM-IV

Morning cortisol

Saliva 4 days, 1 time point

ELISA

↑ cortisol predicts MDD

8

LeMoult et al. [72]

Y

b

USA1

62

26 O

12

100

6.5

K-SADS. SCID; DSM-IV

Morning cortisol

Saliva, 2 days, 4 time points

ELISA

↑ cortisol predicts MDD

9

Rao et al. [93]^c

Y

b

USA2

89

14 O

15

58

5

K-SADS; DSM-IV

Saliva, NUFC

Saliva, 3 time points before sleep and Urine 1 time point before sleep

RIA

↑ cortisol predicts MDD

6

Carnegie et al. [41]^c

N

?

UK

841

46 O + R

15

49

3

CIS-R; ICD-10

AUC, CAR, DD, bedtime (m), waking (m), TEC

Urine, over 24 h

EIA

NS

5

Vrshek-Schallhorn et al. [98]^c

Y

?

USA

270

42 O + R

17

72

4

SCID; DSM-IV

AUC, CAR, diurnal slope

Saliva, 3 days, multiple time points

Time-resolved fluorescent-detection IA

↑ cortisol recurrence

9

Appelhof et al. [32]^c

N

a

Spain

45

22 R

50

44

22

At baseline: SCID. Relapse: HRSD. MADRS. and BD; DSM-IV

Post dex cortisol, Max ACTH, Delta ACTH, Max cortisol, delta cortisol, discharge, difference cortisol

Blood, 2 days, multiple time points before and after DEX/CRH combined with TRH

luminescen-ce enzyme IA

↑ maximal cortisol after DEX/CRH predicts relapse

6

Aubry et al. [34]^b

Y

a

Switzerland

34

12 R

44

56

1

MINI; DSM-III-R/IV. ICD-10

Cortisol after DEX/CRH test

Blood, 2 days, multiple time points before and after DEX/CRH

Immulite 2000 analyser

↑ AUC and delta in relapse

8

Banki et al. [35]^b

N

a

Hungary

24

9 R

51

100

0.5

Hospital diagnosis; DSM-III-R

CRH, SRIF

CSF, 1 time point

Sensitive and specific IA

↑ CRF in relapse

3

Bockting et al. [37]^c

Y

b

Netherlands1

55

43 R

44

73

5.5

SCID; DSM-IV

Morning and evening cortisol

Saliva, 2 days, 1–2 time points a day

RIA

↓ cortisol related to relapse

7

Bouhuys et al. [38]^c

N

b

Netherlands

77

21 R

44

66

2

CIDI; DSM-IV

24 h urine

RIA

NS

6

Charles et al. [42]^c

N

?

Belgium

13

7 R

[33–67]

77

1.5

SADS-L; DSM-III and RDC

Morning after DST

Blood plasma, 2 time points after taking DEX

RIA

↑ cortisol (non suppression at recovery) higher rates of MDD readmission

4

Chopra et al. [44]^b

Y

b

Canada

55

28 R

39

64

1.5

SCID; DSM-IV

Morning/evening (before mood induction)

Saliva, 4 time points

EIA

↑ cortisol

4

Cosgriff et al. [48]^c

N

b

New Zealand

13

4 R

51

54

0.25

Clinical readmission; Unclear

Mean cortisol, Delta TSH

Blood, 10–12 time points

EIA/RIA

↑ cortisol

5

Hardeveld et al. [62]^c

Y

a

Netherlands

549

193 R

45

71

4

CIDI; DSM-IV

Salivary CAR, evening, DST

Saliva, 2 days, 6 times day 1, 1 time day 2 after DEX

EIA

AUC increase differed, and related to time to recurrence. Other measures, mean evening, DST and AUC were not predictive

7

Hatzinger et al. [64]^b

N

a

Switzerland

20

12 R

52

70

1

SCID; ICD-10

Cortisol after DEX/CRH

Blood plasma, 1 time point after DEX/CRH

RIA

DEX/CRH test

7

Lok et al. [75]

Y

b

Netherlands1

187

102 R

44

68

2

SCID; DSM-IV

Morning and evening cortisol

Saliva, 2 days, 3 time points

RIA

NS

8

Mander et al. [78]^c

N

?

UK

70

32 R

Unclear

Unclear

3

SCID; DSM-III

DEX suppression

Saliva, 1 day, 3 time points day after DEX

RIA

NS

5

Mocking et al. [118]^c

Y

b

Netherlands1

187

154 R

44

68

10

SCID; DSM-IV

Cortisol/ DHEAS ratio

Saliva, 2 days, 2 time points a day.

RIA

↓DHEAS diurnal course, ↑ cortisol/DHEAS ratio diurnal course

6

Morris et al. [80]^c

Y

a/b

USA

32

9 R

23

63

0.75

SCID; DSM-IV

Cortisol pre- post- TSST

Saliva, 6 time points

ELISA

NS

8

Pintor et al. [89]

N

a

Spain1

43

18 R

51

53

2

SCID; DSM-IV

Cortisol and ACTH after CRF injection

Blood plasma, 6 time points around CRF injection

EIA/RIA

NS Cortisol and ACTH AUC, ACTH after CRF

6

Pintor et al. [90]^c

N

a

Spain1

43

18 R

51

46

2

SCID; DSM-IV

Cortisol and ACTH after CRF injection

Blood plasma, 6 time points around CRF injection

EIA/RIA

↑ cortisol (NAUCC) after CRF, ↓ACTH after CRF

5

Rao et al. [92]

N

b

USA2

47

20 R

15

64

3.5

K-SADS-PL; DSM-IV

NUFC

Urine before and after sleep

RIA

↑ cortisol

5

Tsuru et al. [96]^c

N

a

Japan

25

9 R

41

64

10

SCID; DSM-IV

ACTH and cortisol, TSH

Blood, 2 days, day 1 TRH test 5 time points, day 2 DEX /CRH-test, 10 time points

IRMA

NS cortisol and ACTH,↑TSH after TRH test in recurrence

5

Zimmerman et al. [102]^c

N

?

USA

165

47 R

40

73

0.5

Unclear; DSM-III

DEX suppression

Blood, 2 time points after DEX

RIA

NS

6

Zobel et al. [103]

N

a

Germany2

40

10 R

50

65

0.5

Unclear; DSM-IV

DEX/CRH test

Blood, 9 time points DEX/CRH-test

Unclear

↑ cortisol after DEX/CRH at discharge predicts MDD relapse

3

Zobel et al. [104]^c

N

a

Germany2

74

13 R

50

59

0.5

Unclear; DSM-IV

DEX/CRH test, cortisol and ACTH

Blood, 5 time points DEX/CRH-test

RIA

↑ cortisol predicts MDD, ACTH NS

5

Hormones: HPG

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measure

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Asselmann et al. [33]^a

Germany

1760

165 O

45

50

9

DIA-X/M-CIDI

Testosterone, Androstenedione, sex hormone-binding globuline

Blood serum 8AM-7PM

Liquid-chromatography-tandem mass spectrometry and RIA

NS

6

Hormones: HPS axis

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measurement

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Coplan et al. [47]^c

Y

?

USA

34

13 O

15

52

9.6

K-SADS & K-SADS-E & SADS-LA; RDC

Growth hormone (sleep)

Blood over 2 nights, 72 time points

RIA

↑GH secretion

8

Franz et al. [52]^c

N

a

USA3

29

22 R

40

100

3

SADS; RDC

Growth hormone (sleep)

Blood before onset of and during sleep, every 20 min

RIA

↑GH secretion

5

Jarrett et al. [66]^b

N

a

USA3

29

22 R

40

100

3

SADS; RDC

Growth hormone (sleep)

Blood before onset of and during sleep, every 20 min

RIA

NS

6

Owashi et al. [84]^c

N

b

Japan

26

6 R

57

76

0.5

Unclear; DSM-IV

Growth hormone, ACTH, cortisol

Blood after DEX/CRH and GHRH test, 5 time points

RIA

↓GH after GHRH at time of discharge. Cortisol and ACTH around DEX/CRH: NS

4

Hormones: HPT axis

Baseline MDD excl. Y = yes, N = No

Baseline symptoms a = above cut-off, b = below cut-off, ? = unclear

Country and cohort information (nr indicates similar cohorts)

Total N

Onset (O) or relapse/ recurrence (R) of MDD (N)

Baseline age (mean or range)

% female

Follow-up time (years; mean or max)

MDD diagnostic interview; Diagnostic criteria

Biomarker of interest

Measurement

Technical details

Direction result of biomarker predicting onset, relapse/ recurrence of MDD

QA score

Joffe et al. [67]^a

N

a

Canada

75

71 R

38

66

10

SADS-L; RDC

T4, T3, TSH

Unclear

Unclear

↓T3 was significantly related to time to recurrence.

T4, TSH NS

4