Table 2 Characteristics of included reviews and summary of findings - general older population

Prevalence and co-occurrence

Seitz [23] See 1A

Mar 2009

BPS

Dep

Anx

Care home

35

Prevalence dep symptoms in long term care: 29% (14 to 82%)

-

- Developing countries

- Multinational studies

- Collaboration across centres

- Adoption of standard survey methods

- Effective and safe interventions

Original studies

- Small sample size

- Conducted in developed countries

- Included relatively few long term care facilities

- Many studies conducted several years ago

3

Luppa [28]

May 2010

Dep

Older (60+)

24

Prevalence of dep disorders ranged from 4.5 to 37.4%. Pooled prevalence: 17.1% (95% CI 9.7 to 26.1)

Pooled prev major dep: 7.2% (95%CI 4.4 to 10.6) Dep disorders: 17.1 (9.7 to 26.1)

- Large scale

- Population-based

- Prospective studies

- Also covering oldest age segments

- Comorbidity, cognition and function

- Suitable depression diagnostics

Original studies

- Methodological differences in study design, sampling structure and study quality

5

Chen [20]

Jun 1997

Dep

Older (60+)

10

Prevalence dep mood: 14.8 (14.2 to 15.6%), higher in rural communities

Prev dep mood: 14.8% (14.2 to 15.6)

- Similar methodology

- Culture-specific validated instruments

- Risk factors and understanding dep

Original studies

- Much variation

- Cultural acceptability of instruments

4

Beekman [29]

1996

Dep

Older, community dwelling (55+)

34

The reported prevalence rates vary enormously (0.4 to 35%). Minor dep: 9.8% (8.3 to 14.3) Clinical dep symptoms: 13.5% (2.8 to 35%)

-

- Focus on those most at risk and in adverse socio-economic conditions

- Improving comparability of the data

Original studies

- Methodological differences

- Bias translating instruments

Review

- Formal meta-analysis was not considered justified

3

Meeks [30]

Jan 2010

Dep

Older (55+)

153

Dep was generally at least two to three times more prevalent than major dep. Prevalence lower in community settings (9.8%, 4.0 to 22.9) than primary care (15.1 to 35.9%) and LTC (4.0 to 30.5%).

-

- Incidence

- Prevalence

- Various clinical settings,

- Diverse geographical areas

- Cultural/socioeconomic groups

- Neurobiology

- Treatment

- Terminology of depression

- Associations with psychopathology

Review

- Could not conduct a meta-analysis due to data heterogeneity

- Review did not include data on early or mild adulthood subthreshold depression, limiting extrapolation of findings to other age groups

2

Djernes [31]

Sep 2004

Dep

Older (65+)

122

Prevalence clinical relevant depressive symptoms: 7.2 to 49%

-

- Target risk factors, improvement of prevention and treatment of chronic somatic and mental illnesses, adequate social support, prevention social isolation

- Education and information dep in elderly

- Comparability of methodology

- Focus on nursing home residence

Original studies

- Methodological differences

- Rates of participation; depressed elderly may be particularly prone to refuse research invitations

- Subjective variations in the assessment of the presence or absence of a diagnostic criterion

- Differences between instruments

2

Alwahhabi [32]

2001

Anx

Older (55+)

119

-

See disease outcome

See disease outcome

1

Course and progression

Huang [35]

Aug 2007

Dep

Older (55+)

17

Non-dementia cognitive impairment vs without: incidence dep: OR = 1.5, 95% CI 0.9 to 2.5 prevalence dep: RR = 1.1, 95% CI 0.6 to 2.0. Dem vs. no dem: incidence OR = 1.8, 85% CI 1.2 to 2.9, prevalence RR = 3.9, 95% CI 1.9 to 8.0

See summary of results

- Risk for cognitive impairment for depression

Review

- No conclusion if dep was risk factor for dem

- No hand-search of journals and no attempt to identify unpublished studies. English language only

- Heterogeneity among included studies

- Confounding comorbidity other psychiatric disorders

- Data only gathered until august 2007

- Only four longitudinal studies included

5

Meeks [30]

Jan 2010

Dep

Older (55+)

153

8 to 10% of subthreshold dep developed major dep per year. Median remission rate to non-dep status 27% after > 1 year.

-

- Longitudinal course

See prevalence

2

Jorm [33, 36, 37]

End 2000

Dep

Dem/Older

11, 15, 2

1991: history of dep (late onset cases) associated with AD (late onset). 2000: Dep increased risk of dem in case control, 95% CI 1.2 to 3.5 and prospective studies, 95% CI 1.1 to 3.2.; 2001: Update 2000: case control studies: RR = 2.0, 95% CI 1.2 to 3.5, prospective studies 1.9, 95% CI 1.1 to 3.2

Too many results

1991:

- Prospective studies

- History of psychiatric disorders other than dep and psychiatric treatments

2000/2001:

- Large sample size

- Mechanisms association dep and dem

Review

1991

- The pooled analyses cover only a small number of exposures from the domain of psychiatric history

0

Ohayon [38]

2003

Sle

Adult ("healthy or normal")

65

Total sleep time, sleep efficiency, percentage of slow-wave sleep, percentage of REM sleep and REM latency all significantly decreased with age. Sleep latency, waking after sleep, waking after sleep duration and the percentage of stage 1 and 2 sleep increase with age, but only sleep efficiency continued to significantly decrease after 60 yr.

Age - sleep: TST: r = -0.76 P < 0.0001 Sleep efficiency: r = -0.82, %SWS: r = -0.56% REM: r = 0.16 Sleep latency: r = 0.16% stage 1 sleep: r = 0.16% stage 2 sleep: r = 0.34 WASO: r = 0.75 All P < 0.0001

- Strict screening methods

- Effect of race

- Take into account subjects' habitual sleep schedules as well as whether PSG recording occurs on weekday or weekend night

Original studies

- No information given in relation with the presence or absence of sex differences, no information about race composition

- Several studies did not include middle-aged subjects

Review

- Limited to peer-reviewed studies

3

Floyd [39]

2002

Sle

Adult ("healthy or normal")

244

Age and REM%: essentially linear, decreasing 0.6% per decade but ceased during mid-70s followed by small increase 75 to 85

Age - REM%: r = -0.17

- REM sleep in women

- More data in old-old population

Review

- Studies did not screen for psychoactive substance use, dep and sleep apnea, few studies of women

- Univariate approach

- Publication bias

2

Floyd [40]

1996

Sle

Adult

41

Night-time sleep amount and the ability to initiate sleep decreased with age. Larger age-related changes when sleep variables were measured by polysomnography rather than self-report.

Age - sleep, effect size: Sleep latency: 0.19 (0.14 to 0.24) WASO frequency: 0.38 (0.34 to 0.42) WASO duration: 0.74 (0.71 to -0.77) Night time sleep amount: -0.33 (-0.37 to -0.28)

- Controlling for health moderators (carefully assessed for levels of depression, sleep apnea and use of psychoactive substances)

- Study women

Original studies

- Inclusion or exclusion of certain covariates may have influenced which predictors emerged as significant

- Very few of the studies examined the effects of collinearity, moderation or mediation among critical predictor variables

- Range of quality scores

Review

- Heterogeneity made the estimation of pooled effects impractical

1

Biological

Huang [43]

Aug 2007

Dep

Older (55+)

28

Significant OR and RR for increased dep in old age: stroke, loss of hearing, loss of vision, cardiac disease or chronic lung disease had a. Significant OR but un-significant RR: arthritis, hypertension and diabetes. Both OR and RR not significant: gastro-intestinal disease

Too many results

Review

- Not hand-search journals, not identify unpublished studies, three databases, only English language

- Risk factors dep might be differently related to the onset, chronicity and recurrence but not differentiated

- Recent life event not taken into account

- Heterogeneity in results

5

Huang [44]

Aug 2007

Dep

Older (55+)

31

Chronic disease - dep: RR = 1.5, 95% CI 1.2 to 2.0. poor SRH - dep: RR = 2.4, 95% CI 1.9 to 3.0.

Chronic disease - dep: RR = 1.5 (1.2 to 2.0) SRH - dep: RR = 2.4 (1.9 to 3.0)

Review

- Not hand-search journals, no attempt to identify unpublished studies, three databases, only English

- Heterogeneity in results

5

Almeida [45]

Dep

Older (70+)

17

High tHcy increased risk of dep: OR = 1.7, 95% CI 1.4 to 2..1 TT vs. CC carriers: OR = 1.2, 95% CI 1.0 to 1.5

High tHcy - dep: OR = 1.7 (1.4 to 2.1)

MTHFR C677T - dep: TT vs CC: OR = 1.2 (1.0 to 1.5) CT vs CC: OR = 1.1 (0.9 to 1.2)

- Sufficiently powered randomised trials

Original studies

- Small sample size (trials)

- Reverse causality (observation studies)

- Inconsistent definition phenotype, misclassification bias (genetic studies)

- Lack of reliable information on ethnicity

Review

- Meta-analysis lacked power

4

Stetler [46]

May 2009

Dep

Adult

414

Dep vs no dep: Cortisol d = 0.6 (95% CI 0.5 to 0.7) Adrenocorticotropic-releasing hormone d = 0.28 (95% CI 0.2 to 0.4) Corticotropin-releasing hormone d = 0.02 (95% CI -0.5 to 0.5)

Too many results

- Bioinformatic technologies

- Larger sample size

- Longitudinal

Original studies

- High degree of heterogeneity

- Publication bias possible

- Based on cross-sectional studies

- Arbitrary criteria for minimal methodological quality

- Most of the included studies were underpowered

3

Kuo [47]

Sep 2004

Dep

Adult

19

High concentrations C-reactive protein predictive of cognitive decline and dem. Relations to dep cross and not consistent.

-

- Prospective study c-reactive protein-dep

- Intervention studies to lower c-reactive protein and improved outcomes

NR

3

Kuo [41]

Mar 2004

Dep

Older (55+)

NR

Growing evidence of association hyper-homocysteinemia and cognitive impairment, dem and dep. Proposed mechanisms include angiotoxicity, neurotoxicity, and inhibition of collagen cross-linking

-

- Role of homocysteine in prevention

- Prospective studies association with dep

- Adequate adjustment for possible confounders

NR

3

Camus [48]

Jun 2003

Dep

Older

NR

Potential ways association dep - vascular disease: 1 direct influence vascular disease, 2 direct influence dep, 3 common causes

-

- Pathophysiological and genetic background of vascular depression

NR

1

Vink [49]

Dec 2005

Anx

Dep

Older (50+)

80

Risk factors anx and dep showed many similarities but some differences were found. Biological factors may be more important in predicting dep, and a differential effect of social factors on dep and anx was found.

-

- Intervention (whether manipulation of risk factors reduces the onset of anx/dep)

- Clearer understanding of etiological factors differentiating anx and dep

Review

- Heterogeneity between studies, no meta-analysis

- Only main effects of risk factors on anx and dep

- Heterogeneity limits comparison across studies

- Risk factors that have not yet been studied

- No distinction made between different anx disorders

1

Risk factors

Chen [20]

Jun 1997

Dep

Older (60+)

10

The patterns of risk factors were similar to those in western countries

See prevalence

See prevalence

See prevalence

4

Meeks [30]

Jan 2010

Dep

Older (55+)

153

Risk factors: female, medical burden, disability and low social support; neurological illnesses (Parkinson's disease, stroke, AD)

-

- While some risk factors are well established, others remain to be identified.

See prevalence

2

Djernes [31]

Sep 2004

Dep

Older (65+)

122

Risk factors: female, somatic illness, cognitive and functional impairment, lack of social contacts, history of dep

-

See prevalence

See prevalence

2

Cole [50]

2001

Dep

Older (50+)

20

Risk factors, Qualitative: disability, new medical illness, poor health status, prior depression, poor self-perceived health, and bereavement. Quantitative: bereavement, sleep disturbance, disability, prior depression, female gender

13 risk factors investigated. OR ranged from 1.0 to 3.3, significant risk factors: bereavement, sleep disturbance, disability, prior dep, female gender

- Intervention

Original studies

- Follow-up incomplete in most studies

- Differences in the length of follow-up

- Differences in definitions risk factors and adjustment

- Many potential risk factors not studied adequately

- Cumulative effect of multiple risk factors not studied

- Heterogeneity in the results

Review

- Search by one author only

- Only English or French literature

- Did not assess publication bias

- Abstracted by one author

2

Vink [49]

Dec 2005

Anx, Dep

Older (50+)

80

Risk factors both anx and dep: personality, coping strategies, previous psychopathology, social network, stressful life events, female. Dep: smaller network size, being unmarried.

-

See biological

See biological

1

Disease outcome

Meeks [30]

Jan 2010

Dep

Older (55+)

153

Consequences: disability, greater healthcare utilisation, increase suicide ideation

-

- More sophisticated health economic studies

See prevalence

2

Alwahhabi [32]

2001

Anx

Older (55+)

119

Limitations: understanding expression anx, variable definitions elderly, diagnostic instruments. Anx in elderly potential for negative consequences independent of comorbidity major dep.

-

- Definition of elderly

- Symptom definition and diagnostic instruments

- Clinical trials

Original studies

- No common definition of the lower limit of geriatric age

1