Depression in later life: who benefits most from antidepressants plus exercise?


There has been a great deal written about exercise in depression, not least in previous Elf blogs (e.g. Tomlin A, 2012 a & b; Hasselman H, 2014). In recent years some studies, particularly TREAD (Chalder et al 2012), suggested no benefit from adding exercise to routine care, but the conclusion of NICE remains that exercise is effective (NICE, 2013).

Depression in later life is often accompanied by significant cognitive impairment, disability, suicide risk and physical co-morbidity and there is increasing evidence for the effectiveness of physical exercise for depressed older people (Bridle et al, 2012; Chi et al, 2013; Johnson S, 2014). The problem is that translation of research findings into clinical practice can be difficult and the majority of older people with depression are treated in busy primary care settings.

The aim of this study by Zanetidou and colleagues (2016) was to identify which participant characteristics (moderators) and contextual factors influenced the translation into clinical practice of interventions based on physical exercise (PE) as an adjunct to antidepressants (AD) for the treatment of late life major depression.

NICE recommend exercise for depression, but this is based on sparse evidence of effectiveness in clinical practice.

NICE recommend exercise for depression, but this is based on sparse evidence of effectiveness in clinical practice.


This was a secondary analysis of a multicentre, single-blind randomised controlled trial of treatment for late life major depression (LLMD) in primary care (the Safety and Efficacy of Exercise for Depression in Seniors (SEEDS) study) (Belvederi et al, 2015).

In the SEEDS study, participants (aged 65-85 years and meeting 3 criteria- DSM-IV Major Depressive Disorder with a 17 item Hamilton Depression Rating Scale (HAM-D) score of 18 or more, physical condition compatible with exercising and sedentary), received either antidepressants (AD: sertraline) or antidepressants + physical exercise  (AD+PE).

They were recruited from 4 centres, which were already engaged with an established regional primary care consultation-liaison service. Assessments for inclusion were carried out by the study psychiatrist (using HAM-D and Hospital Anxiety and Depression Scale to assess severity), cardiologist, and geriatrician (medical history, comorbidities using the Cumulative Illness Rating Scale (CIRS) and the Montreal Cognitive Assessment (MOCA)). Maximum oxygen intake (VO2 max) was also measured before and after the study.

The main outcome measure was remission defined as a HAM-D score of 10 or less.

The SEEDS intervention: lasted for 24 weeks. Participants were randomised to 3 groups:

  • Antidepressants only (n=42)
  • Antidepressants plus low intensity non-progressive exercise (NPE) (n=37)
  • Antidepressants plus high intensity, progressive aerobic exercise (PAE) (n=42)

Secondary analysis

For moderators of treatment remission:

  • A set of a priori defined factors that might influence response to antidepressants or the ability of older people to exercise was identified, including age, marital status, sex, living status and education.
  • Factors related to participant physical condition were analysed: weight, comorbidity (CIRS comorbidity index), musculoskeletal problems, cardiovascular disease (defined according to CIRS items), level of physical activity (International Physical Activity Questionnaire score), aerobic capacity (VO2max).
  • Polypharmacy was defined as taking two or more medications other than sertraline daily. Other psychotropic drugs such as sedatives and anxiolytics were counted only if participants used them regularly, although the study protocol discouraged it.
  • Next, mental health factors were explored (depression (HAM-D score), age of onset of depression, number of episodes of depression, previous treatment with AD, anxiety (HAD anxiety subscale), psychomotor retardation (HAM-D Item 8 score), mild cognitive impairment (baseline MOCA total score), and executive or visuospatial domain impairment (MOCA item 1 score).

Analyses of moderators were conducted by testing the interaction between moderator and treatment (AD vs. AD + PE), based on logistic regression.

For significant moderators, three measures of treatment effect size were computed:

  • Success rate difference (SRD, difference between proportions of participants remitting with AD + PE and  AD)
  • Number needed to treat (NNT, equivalent to 1/SRD)
  • Moderator effect size (difference of SRDs between participants with and without the characteristics of interest).

Whether the characteristics of the study setting influenced participant flow and attendance at exercise sessions was then explored, and GPs were surveyed regarding their opinions on physical exercise as a treatment for late life major depression.


In the main study those receiving the combined interventions achieved remission from depression more frequently and earlier than those receiving only antidepressants. As differences in the rates of remission from depression in the AD + PAE (low intensity) and AD + NPE (high intensity) were not significant, they were combined in a single group for purposes of moderator analyses.

Participant characteristics

Participant characteristics associated with greater likelihood of achieving remission from depression with AD + PE than with AD alone were:

  • Aged 75 and older (effect size 0.32)
  • Polypharmacy (0.35)
  • Greater aerobic capacity (0.48)
  • Displaying psychomotor slowing (0.49)
  • Less severe anxiety (0.30)

The longer the liaison service had been established at a particular centre, the more individuals were recruited.

Prescribing exercise

Before the study, GPs reasons for not prescribing exercise were: lack of infrastructure (58%), not thought effective (50%), thinking individuals wouldn’t comply (50%), and physical condition (33%).

After participation, the majority of GPs were much more positive about exercise for LLMD, and 80% stated that they had prescribed exercise for their elderly patients to treat their depressive symptoms.

This study identifies a number of patient characteristics associated with achieving greater remission from depression when combining exercise with antidepressants.

This study identifies a number of patient characteristics associated with achieving greater remission from depression when combining exercise with antidepressants.


  • The combination of PE and sertraline could improve the management of LLMD, especially when customised for individuals with specific clinical features.
  • Liaison programs might influence the implementation of similar interventions in primary care, and PCPs viewed them positively.

Strengths and limitations

The authors of the study consider that it provides insights into implementation of similar interventions into routine practice, and, in my view key strengths are that not only was it conducted in a primary care setting, but it sought to tease out those individual and contextual factors related to both outcome and uptake.

Admitted weaknesses are lack of inclusion in the original study of either an exercise only group or a control group providing ‘routine care’, and the self-selection of referring GPs.

The extensive physical assessment carried out before starting exercise is also unrealistic in routine practice.

This study is limited by the lack of an exercise-only or control group.

This study is limited by the lack of an exercise-only or control group.


The authors make some interesting suggestions that further the debate about who might benefit from exercise in older people with depression. The combination of older age, polypharmacy and greater aerobic capacity might identify a group of people with diseases that don’t significantly affect our aerobic capacity (for example diabetes, osteoporosis) or mild cognitive impairment (e.g., retardation, executive dysfunction, apathy- which don’t respond well to antidepressants but do improve with exercise). These people may take more prescriptions but generally remain physically fitter. Or they could simply be those who are usually better at looking after their health anyway and used to exercise before becoming depressed and sedentary. Whatever the case, the physical assessments carried out before exercise would be reassuring to both GP and patient but unrealistic in routine care (none of the GPs had ever prescribed exercise only after a cardiopulmonary stress test (e.g. V02 max).

What is clear is that collaborative multidisciplinary approaches to care are likely to be more effective for people with complex problems, and there are roles for GPs and psychiatrists to work together in both motivating older people with depression to take up exercise and monitor adverse effects of medication and activity. The degree of engagement in primary care liaison by psychiatrists within this region of Italy (Berardi et al, 2014) is consistently impressive (other health care systems take note).

GPs and psychiatrists should work together to motivate older people with depression to take up exercise.

GPs and psychiatrists should work together to motivate older people with depression to take up exercise.

Conflict of interest

I have co-published with one of the researchers in this study.


Primary paper

Zanetidou S, Belvederi Murri M, Menchetti M, Toni G, Asioli F, Bagnoli L, Zocchi D, Siena M, Assirelli B, Luciano C, Masotti M, Spezia C, Magagnoli M, Neri M, Amore M, Bertakis KD and the Safety Efficacy of Exercise for Depression in Seniors Study Group. (2016) Physical Exercise for Late-Life Depression: Customizing an Intervention for Primary Care. Journal of the American Geriatrics Society Version of Record online : 21 NOV 2016, DOI: 10.1111/jgs.14525

Other references

Belvederi Murri M, Amore M, Menchetti M et al.(2015)  Physical exercise for late-life major depression. Br J Psychiatry  207 235–242.

Berardi D, Ferrannini L, Menchetti M. et al (2014) Primary care psychiatry in Italy. J Nerv Ment Dis  202:460–463. [PubMed abstract]

Bridle C, Spanjers K, Patel S et al. (2012) Effect of exercise on depression severity in older people: systematic review and meta-analysis of randomised controlled trials. British Journal of Psychiatry 201 180-185.

Chalder, M, Wiles NJ, Campbell al (2012) Facilitated physical activity as a treatment for depressed adults: randomised controlled trial. BMJ  344 e2758.

Chi I, Jordan-Marsh M, Guo M. et al. (2013) Tai chi and reduction of depressive symptoms for older adults: A meta-analysis of randomized trials. Geriatrics and Gerontology International 13(1) 3-12 [PubMed abstract]

Hasselmann H. (2014) Does depression make us lethargic or does lack of exercise make us depressed? The Mental Elf 11 Nov 2014.

Johnson S. (2014) Program of regular exercise may be beneficial in reducing depression in older adults. The Mental Elf 18 June 2014.

NICE (2013) Surveillance review of CG90: Depression in adults: management and treatment (update)

Tomlin A. (2012)  Here is the evidence for exercising if you are depressed. The Mental Elf 11 June 2012a

Tomlin A. (2012) Exercise doesn’t help with depression. Have the headline writers got it wrong again? The Mental Elf 7 June 2012b

Photo credits


Share on Facebook Tweet this on Twitter Share on LinkedIn Share on Google+