IPA Bulletin
Recent Advances

By Dr. Robert Barber and Professor Robert Baldwin

Special Retirement Announcement
After many, many years of distinguished service, Robert Barber and Robert Baldwin are retiring as editors of the Research and Practice column and this will be their last submission. Their outstanding work made this column an IPA member favourite.

It has been a privilege and a pleasure working with Drs Barber and Baldwin throughout the years and they have set a very high standard that the new editors look forward to achieving.

On behalf of the IPA Board of Directors, the IPA Bulletin Editors, IPA members, and the IPA Secretariat who have thoroughly enjoyed this column over the years -- Thank you for your immeasurable contributions to IPA.


Very little is known about the epidemiology of late life depression in non-high income countries. Guerra and colleagues from the 10/66 population based study (British Journal of Psychiatry, 2009; 195: 510-515) studied 5 areas in Latin America, urban and rural Peru, urban and rural Mexico and Venezuela. They used the automated computer based GMSS (Geriatric Mental State) system with the Level 1 output (which basically categorises symptoms on a level of severity) to estimate prevalence. They then organise this data into prevalence by DSM IV, ICD-10 and the Euro-D Scale. They also measured, medical problems, disability and social circumstances. For DSM IV, the overall prevalence varied between 1.3 and 2.8% and for ICD-10 between 4.5 and 5.1%. However using the GMS AGECAT scores (a computerised diagnostic schedule) for ‘clinically significant depression’, the rates were between 30.0 and 35.9% and for the Euro-D 26.1-31.2%. There was a strong association between sub-syndromal depression and disability, supporting the notion that the narrower operational criteria such as ICD-10 and DSM do miss a lot of significant depression. The most striking correlates of being depressed were a previous history of depression and limiting physical impairment. This has important implications for recognising depression in order to prevent future recurrences and for improved Primary Care treatment of common medical disorders. As with many other surveys, the level of treatment of established cases was poor.

Prevention is Better Than Cure
Lyness and colleagues (American Journal of Psychiatry 2009; 166: 1375-1383) enrolled 617 subjects from Primary Care Practices and followed them up for a maximum of 4 years (405 one year, 338 at two years, 259 at three years, 54 at four years). They assessed the incidence of major depression (5.3%) and then looked for indicators that would have predicted the depression. The combination of minor (subsyndromal) depression, lowered functional status and a history of depression was the best set of variables (from quite a large variety) with which to predict depression. Note that it was functional impairment rather than medical illness burden that was the main predictor suggesting that this may be more important than individual diseases in predicting depression incidence. The number needed to treat (NNT) using these variables in combination is 5 although, as the authors point out, this assumes a fully effective treatment. This is less likely in the face of medical morbidity and impairment. Nevertheless, this study is important because it shows how in Primary Care the prevention of depression is a practical proposition because the risk factors lend themselves readily to chronic disease management models and the disease case registers that often accompany such models.

Antidepressants and the heart
There is a longstanding controversy regarding antidepressants and heart disease. Some studies showed SSRIs benefit heart disease patients but others (for example the Cardiac Arrhythmia Suppression Trial) issued a caution about tricyclics because of an increase in mortality. Therefore, the study by Smoller et al (Archives of Internal Medicine, 2009; 169: 2128-2139) from the US Women’s Health Initiative Studies is of relevance. Antidepressants at baseline were divided into SSRIs, TSAs, other or none and depression was assessed with a short 8 item scale. End points were occurrence of coronary heart disease including fatal infarction, fatal stroke or non-fatal stroke. The cohort of antidepressant users (N = 5496) were followed up for an average of 6 years. All participants were post-menopausal women with the largest group being 60 to 69 years. SSRI use showed a 45% increased relative risk of death. The risk of death was likewise increased for TCAs. There was a significant increased risk of haemorrhagic stroke for SSRI users. These findings were independent of vascular risk factors. The fact that all cause mortality was increased in antidepressant users raises the question of whether residual depression rather than antidepressants was the causative ingredient. As the accompanying editorial points out, this question remains unanswered and there is always a need to balance the risks of depression itself against potential risks of antidepressants. Nevertheless, this is an important study.

Dysexecutive problems, dementia and antidepressants
Two studies from the American Journal of Geriatric Psychiatry, February Edition, outline the use of SSRIs in the setting of impaired cognition. Sneed and colleagues from New York (American Journal of Geriatric Psychiatry 2010; 28: 128-135) examined whether patients with late life depression respond less well to antidepressants if they have dysexecutive impairment, using the Stroop Test. The sample (N=174) was sub-divided by treatment with Citalopram or placebo sub-divided by whether dysexecutive symptoms were present or absent, making 4 groups in all (about a quarter of those actively treated and in the placebo group had such impairment). Although we know that patients who have dysexecutive problems respond less well to antidepressants, the surprising finding of this study was that those with this impairment, as measured with the Stroop, responded much better to placebo than to Citalopram. This was of an order that would be detected clinically. Conversely, those treated with Citalopram did better than placebo provided there was no evidence of dysexecutive problems. The authors postulate that dysexecutive problems are often associated with fronto-striatal disruption that limits the effects of antidepressants. The placebo response affects different areas of the brain. It is important to remember that the Stroop Test assesses only one aspect of executive difficulties.

In a study by Rosenberg and colleagues from Baltimore (American Journal of Geriatric Psychiatry, 2010; 18: 136-145), patients with depression in Alzheimer’s disease were allocated either to placebo (N=64) or Sertraline (N=67) to a dose of 100 mg. Over 12 weeks there was no difference between the groups either on clinical global outcome measure or on the Cornell Scale. Both groups had support with a psychosocial intervention at every study visit and with phone calls, so this may have modified treatment effects, but even so, the Sertraline group was associated with a higher rate of adverse reactions which included tremor and respiratory problems. A disappointing result, but soon we should have data from a larger UK study (HTA-SADD) in which two antidepressants, mirtazepine and sertraline, were compared with placebo. We will hope for better news.

Nursing home depression
Depression in nursing homes is known to run a chronic course. In a study from Norway by Barca and colleagues (Journal of Affective Disorders, 2010; 120: 141-148), the hypothesis was that the incidence and persistence of depression in nursing homes is high and predicted by length of stay, baseline depression symptoms and physical health problems. Using the Cornell Scale (for depression), a sample of 902 was followed up at 12 months. The mortality rate was 26.7% and many of the survivors were hard to evaluate either because of aphasia or severe dementia. Nevertheless, using a score of 8 and above on the Cornell Scale, there was a 15% incidence of depression and a high persistence rate (44.8%). This is consistent with, or a little lower than, other studies. Predictors of depression at 12 months were a high Cornell Score at baseline, a shorter stay in a nursing home (presumably because of the difficult adjustment of giving up one’s home) and the use of antidepressants. The latter is intriguing as the rates of antidepressant for those with higher depression scores was only 43%. One conclusion is that depression in this setting is simply very hard to treat. A strength of the study is that it did not exclude residents with dementia. Also of interest is the difficulty that the authors had in the use of the Cornell in patients who were severely demented because of the overlap between dementia and depression for symptoms such as irritability, retardation and agitation.

Depression and age
Does depression prevalence increase with age? This question was asked by Mossaheb and colleagues from Vienna (Journal of Clinical Psychiatry, 2009; 70: 500-508) in the Vienna Transdanube Aging Study (VITA) of a cohort of 75 year olds who were followed up at 30 months. Of 331 participants who were not depressed at baseline, 31.4% developed minor or major depression at 30 months. 14.2% had mild cognitive impairment, 42.5% of whom were diagnosed with a new onset depression. Among the many risk measures, including genetic factors and medical illness, only social factors seemed to be triggers, especially problematic relationships with relatives. Since only 10.8% were depressed at baseline and over 30% were at follow up, the authors suggest that the prevalence of depression does indeed increase with age but not in a way that can be clearly linked to biological factors, save perhaps cognitive impairment. It is surprising that vascular risk factors did not predict these new onset cases, all of whom were late onset depressions.

Leading on from the previous report, perhaps age of onset rather than actual age is critical in determining the extent of pathology and prognosis. Coryell et al (Psychological Medicine, 2009; 39: 1689-1695) examined 20 year follow up data to explore the outcome of 3 groups of patients with depression or schizo-affective disorder. Age of onset of the disorder was recorded. 93 subjects were aged under 30 when they entered the study, 85 were aged 30-34 and 42 were aged over 45 years. There was no evidence of increasing or decreasing morbidity across the age periods, thus depressive symptoms did not change as individuals moved from their fourth to sixth decade. However age of onset did predict symptom persistence, with early age of onset being associated with the worst prognosis. Gender differences were confined to the youngest age group with women tending to worsen whilst men improved. Importantly, the pattern of relapse seemed to be set early in the illness and did not vary much.

Cognition in Schizophrenia
In the first Meta-analysis of its kind, Rajji and colleagues (British Journal of Psychiatry 2009; 195: 286-293) examined the prevalence of cognitive dysfunction in schizophrenia across the lifespan. They divided subjects into youth-onset schizophrenia (age under 19), first onset adult schizophrenia and late onset schizophrenia (aged over 40 although the average age of this sample was 68.4 with a mean age of onset of 60.7 years). They were able to aggregate cognitive scores on a range of up to 20 measures. Those with youth-onset schizophrenia had the greatest cognitive impairment affecting arithmetic, executive function and overall IQ as well as processing speed. This is consistent with youth-onset being a marker of more severe form of schizophrenia. In contrast, those with late onset schizophrenia had relatively preserved cognitive functions such as arithmetic, digit symbol coding and vocabulary. However they were quite severely impaired on measures of auditory and visual attention, fluency, global measures of cognition, IQ and visuo-spatial construction, more so than those with youth onset schizophrenia or first episode schizophrenia on most of these measures. The different patterns might be explained by the longer duration of the late onset group and the pathoplastic effects of ageing along with cerebrovascular disease.

Do antidepressants cause arthritis?
Type 2A 5-HT receptors are obviously found in the brain but also vascular smooth muscle and probably the immune system. Kling and colleagues from Sweden (Arthritis and Rheumatism 2009; 61: 1322-1327) explored the WHO Adverse Reaction Database for Sweden, which has high reliability. They explored relationships between the three most common atypical antidepressants, mianserin, mirtazapine and nefazodone at the time for arthralgia, arthritis, exacerbated arthritis, arthropathy and synovitis between 1990 and 2006. Bisphosphonates were chosen as a reference point. The study involved all ages. These atypical antidepressants were more likely to be associated with adverse reactions in the joints, more so than SSRIs. Type 2A 5-HT receptors are present on activated T lymphocytes so that inflammatory response may be one explanatory mechanism. Type 2A receptors are also present in the ganglia which innervate synovium in the joints. What this means for psychiatric practice remains to be seen but for psychiatric drugs which block Type 2A 5-HT receptors, there may need to be increased awareness of this potential adverse reaction.

Mental Health and Pulmonary Disease
Laurin and colleagues from Canada (Psychosomatic Medicine, 2009; 71: 667-674) prospectively studied 116 patients with chronic obstructive pulmonary disease. They were followed up for an average of 2 years. All were interviewed with an anxiety and depression inventory linked to DSM-IV. The average age was 67 years. Those identified with depression or anxiety had 40% more exacerbations than those without any psychiatric disorder. They were almost twice as likely to have an out-patient exacerbation compared to those without psychiatric disorder such that a quarter of the risk associated with out-patient exacerbation was accounted for by depression or anxiety. Inpatient exacerbations did not differ. The strength of the study is that it is perspective and confined to one condition. Besides showing the importance of psychiatric morbidity on medical outcomes, the authors suggest a variety of explanations from increased distress and increased sensitivity to COPD symptoms, physiological responses that might provoke autonomic dysregulation and weakened immune function. It also suggests that prompt treatment of mental disorder in medically ill patients may be immediately beneficial to the associated medical condition.

Folate Metabolism and Depression
There is evidence that low folate and high homocysteine levels are associated with depression and, in the latter case, with white matter hyperintensities. Using the Neurocognitive Outcomes of Depression in the Elderly (NCODE) study, Hong et al (American Journal of Geriatric Psychiatry, 2009; 17: 847-855) compared 178 older depressed subjects (mean age 70 years) with 85 control subjects. They assessed geno-type for the 5, 10 – Methylene-Tetra-Hydrofolate-Redutase (MTHFR) gene, there being evidence that the T allele is associated with higher white matter lesion volumes in the brain and possibly depression. They also assessed cognition with a test battery. They found an interaction between the T allele version of MTHFR and white matter lesions but not grey matter lesions. There were no associations with neurocognitive tasks. The effect of the gene on white matter lesions exhibited a steeper slope with advancing age. This genetic polymorphism was not specifically associated with depression. This adds to the uncertainty regarding MTHFR and late-life depression but suggests that seeking gene-environment-biological interactions may be more profitable than examining risk factors in isolation. There was no assessment of homocysteine levels, so we do not know if this extra risk of white matter lesions was associated with raised homocysteine levels, as others have previously found.





 

Reprinted from IPA Bulletin, Volume 27, Number 1