This study has several elements that give it robustness, such as the size of the sample and that this sample can be said to be population-based (it incorporates the entire Swedish population with dementia). This fact is relevant because there are few databases in the world that have this type of information at their disposal.

However, a cause-effect relationship cannot be determined between treatment with antidepressants and the average decrease in the MMSE score (which the authors themselves already acknowledge). First of all, it must be taken into account that the treatment is to modify a situation (depressive mood) that in Alzheimer's disease and other dementias can be camouflaged among the cognitive deterioration itself, or be a consequence of it. At the same time, the brain where these symptoms appear is much more deteriorated than in the general population with depression where the treatment studies have been carried out. It should also be added that the MMSE is a screening test for cognitive deterioration and that it is not a specific instrument to measure amnesic alterations but that the deterioration can affect other domains.

In the analysis by dementia subtypes and by drug subtypes it is relevant and, once again, it is a pleasure to know that there is a registry with this power to be able to carry out this type of work. However, the specific reason why the treatment has been prescribed is not stated, because depending on the symptoms, one type of drug is more appropriate than another, while depending on the type of injury, some types of symptoms are more likely than others. For example, the appearance of apathetic symptoms in vascular dementia is known, while the use of Mirtazapine can occur in patients who present insomnia and/or loss of appetite. There are notable differences between citalopram and escitalopram, with the former having the least impact – although, at some point, there were doubts as to whether the latter would be the only option of the commercial company.

The reason or the specialist who has incorporated these treatments in the patients is also not stated, so few conclusions can be drawn beyond the already known caution regarding these drugs. This caution should be increased in Spain, as it is one of the European countries with the highest prescription of antidepressants and benzodiazepines, mostly to older people.

As the authors themselves acknowledge, there is substantial risk with this study design for confounding by indication, which could explain the results either in part, or entirely. Confounding by indication is where the outcomes we observe are actually due to the underlying reason people take these medications in the first place (e.g. mental health conditions), rather than the medication itself.

More robust study designs, which overcome this very important limitation, are needed before such bold conclusions can be made. When based on limited evidence, these claims can be very damaging for public understanding of antidepressants, which we know help millions of people around the world.

This large population-level study from Sweden uses real-world data and is well-conducted. However, there are some important limitations that should be considered. One major issue is that the severity of depression in dementia patients wasn't fully accounted for, which has the potential to bias the results. Additionally, there may be a "channelling bias," meaning that certain antidepressants like citalopram and sertraline might have been more commonly prescribed to patients with severe dementia, which could also bias the results. 

“Another key limitation is that the study found only a small change in MMSE (Mini-Mental State Examination) scores, which may not be meaningful in everyday clinical practice. Previous research has shown that older adults taking tricyclic antidepressants can experience faster cognitive decline because these drugs interfere with the activity of acetylcholine—a chemical in our brain critical for maintaining cognition. Even among SSRIs (a common type of antidepressant), some, like paroxetine, are known to have stronger anticholinergic effects that could impact cognition negatively. There is also a problem of “residual confounding”,- meaning there could be other risk factors that can affect cognition, and it is unclear if they have accounted for other anticholinergic drugs like oxybutynin, which many dementia patients take to treat their urinary incontinence. The analyses were done on dispensed data (medication sold by pharmacists), and it is unclear if patients actually took them. 

“This study suggests that SSRIs like citalopram and sertraline might also speed up cognitive decline. However, it doesn’t explain how or why this happens at a biological level. Because of these limitations, the study's findings should be interpreted with caution and ideally replicated using other real-world data sources.