In order to successfully conclude their work, the researchers focused on the database created from the ICTUS/DSA study (Impact of Cholinergic Treatment Use/Data Sharing Alzheimer), whose objective was to examine the natural history of Alzheimer’s disease, the effect of symptomatic treatments relying on the use of inhibitors and the socioeconomic impact of the condition. Conducted between 2003 and 2005 in 12 European countries (France, Switzerland, Italy, Spain, Greece, Germany, Belgium, Romania, Great Britain, Holland, Sweden and Denmark), The ICTUS study recruited 1,379 patients with a probable diagnosis of Alzheimer’s, 64.7% of whom were women.
![ICTUS Alz Selection bias]( "ICTUS Alz Selection bias.")
Based on precise exclusion criteria (never having worked, having left the work environment before the age of 50, being still active professionally, having been diagnosed with Alzheimer’s before retirement), the authors of the article published in PLoS One selected 815 individuals from the ICTUS database who were at a mild or moderate state of the disease. Initially, they used a similar measure to that used by Lupton, by distinguishing two elements: the age of diagnosis and the age when symptoms first appeared. The results were as follows: in the first case, each extra year of work delayed the onset of Alzheimer’s by 0.31 years; in the second case the onset of the disease was delayed by 0.30 years. In other words, the trend described by Lupton was not only confirmed but was, in fact, consolidated.
It was now time to check the biases that were attributed to the British study. First, the selection bias. In order to do this, Catherine Grotz and her collaborators restricted the initial sample to 637 retired individuals who were aged 65 or less and who had not developed Alzheimer’s before this age. Under these conditions, the previously recorded effect remained significant but was reduced by half: each extra year of work now only delayed the age of onset of the disease by 0.15 years whatever criterion was applied – the age of diagnosis or the age when the first symptoms appeared.
It was still necessary to deal with the reverse causality bias, that is to say, to ensure in as far as it was possible to do so, that the subjects of the study were not in the prodromal phase of the disease and had not therefore retired due to cognitive difficulties. What was the best way to proceed? The best way to proceed was to only retain for the purposes of the study retired individuals aged 65 or less who developed a form of dementia at least 10 years later. “At this point there were 447 individuals left in the sample”, explains Catherine Grotz. “It emerged that, among this group, an extra year of professional activity delayed the age of onset of Alzheimer’s by 0.06 years on average. This figure was no longer statistically significant; it was in-keeping with the trend”.
This result also requires a prudent approach with regard to the conclusions that can be drawn from the links that exist between these two parameters of the age of retirement and the age of onset of Alzheimer’s disease. “Only prospective longitudinal studies can truly establish the accuracy of the conclusions”, insists the ULg psychologist.
