Effect of trazodone on cognitive decline in people with dementia: Cohort study using UK routinely collected data

Probably not a substantial claim. The paper’s conclusions may not depend on it.

1,733 (78.8%) had a neurodegenerative non-vascular dementia.

“1,733 (78.8%) had a neurodegenerative non-vascular dementia.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Mean baseline MMSE in the total sample was 18.8 (SD 2.6).

“Mean baseline MMSE in the total sample was 18.8 (2.6)”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Mean daily trazodone dose was 101.8 mg (range 50-300 mg).

“mean daily dose was 101.8 mg (range 50-300 mg)”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

In a sample of 55 participants, the primary indication for trazodone was depression, agitation or insomnia.

“The primary indication for use of trazodone in a sample of 55 participants was depression, agitation or insomnia”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Most participants were of white ethnicity (79.8%), 34.2% were married and 32.7% widowed.

“Most were white ethnicity (79.8%), 34.2% were married and 32.7% widowed.”

It may state two separate claims as one.

There was an absence of cognitive benefit from trazodone in all three clinical sites, as well as when combined in meta-analysis.

“We found an absence of cognitive benefit from trazodone in all three clinical sites, as well as when combined in meta-analysis.”

It may state two separate claims as one.

There was a similar pattern of worse cognitive trajectory for trazodone compared with mirtazapine, but the result was not significant due to smaller numbers.

“there was a similar pattern in the comparison between trazodone and mirtazapine although the smaller numbers in these groups meant that the result was not significant.”

It may state two separate claims as one.

The study included 2,199 people with dementia, of whom 406 were prescribed trazodone, 702 mirtazapine and 1,091 citalopram.

“We included 2,199 people with dementia, of whom 406 were prescribed trazodone, 702 mirtazapine and 1,091 citalopram”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Mean follow-up was 2.2 years (SD 1.4).

“The mean follow-up was 2.2 years duration (standard deviation (sd) 1.4 years)”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The sensitivity analysis included 1,236 people with all-cause dementia in South London.

“We included 1,236 people with all-cause dementia in South London”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Mean time between MMSE assessments was between 4 and 7 months in the drug groups.

“the mean time between each MMSE assessment was between 4 and 7 months in the drug groups.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Results in sensitivity analyses were consistent with primary analyses.

“Results in sensitivity analyses were consistent with primary analyses.”

The quote only partly states this claim.

In South London, for people with non-vascular dementia, there was a marginally faster rate of MMSE decline in those taking trazodone compared to mirtazapine (0.31 points more per MMSE assessment, 95% CI 0.00 to 0.62).

“there was a marginally faster rate of MMSE decline in the South London site for people taking trazodone compared to mirtazapine (0.31 (0.00, 0.62) points more per MMSE assessment)”

• AssumptionRests on an assumption: The PERK/eIF2α-P pathway is not pathogenic in vascular dementias.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Participants were on average 79.3 (SD 3.0) years old and 1,435 (65.3%) were female.

“Participants were 79.3 (3.0) years old on average and 1,435 (65.3%) were female.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The cohort comprised 455 participants from North London, 1,263 from South London, and 481 from Oxford.

“455 participants were from North London, 1,263 from South London; and 481 from Oxford.”

It may state two separate claims as one.

The indication for trazodone differed between sites, suggesting that clinical use of the medication may be different, but consistent results were found from the three sites.

“Furthermore, the indication for trazodone differed between sites suggesting that clinical use of the medication may be different but nonetheless we found consistent results from the three sites.”

It may state two separate claims as one.

Among 660 of 1,263 patients in South London with mild all-cause dementia at baseline (61 trazodone, 248 mirtazapine, 351 citalopram), people taking trazodone had faster cognitive decline than those taking mirtazapine (0.45 MMSE difference per assessment (0.02; 0.88), p = 0.04) and citalopram (0.53 (0.12; 0.94), p = 0.01).

“When we included only the 660 of 1,263 patients in South London with mild all-cause dementia at baseline (61 trazodone, 248 mirtazapine, 351 citalopram) we found that people taking trazodone had faster cognitive decline than those taking mirtazapine (0.45 MMSE difference per assessment (0.02; 0.88), p = 0.04) and citalopram (0.53 (0.12; 0.94), p = 0.01) (Appendix 1b).”

It may state two separate claims as one.

In sensitivity analyses, results were consistent when adjusted for neuropsychiatric symptom severity and when only those with mild dementia at baseline were analysed.

“In our sensitivity analyses, we found consistent results when we adjusted for neuropsychiatric symptom severity and when we only analysed those with mild dementia at baseline”

It may state two separate claims as one.

Manual review of the notes of 55 randomly-selected patients found a range of daily doses (50 to 300 mg) and one patient marked as receiving the medication 'PRN'.

“We therefore manually reviewed the notes of 55 randomly-selected patients, finding a range of daily doses (50 to 300 mg) and one patient marked as receiving the medication 'PRN'.”

It may state two separate claims as one.

In models adjusted for agitation and depressed mood, people with mild dementia taking trazodone performed worse on successive MMSE assessments than those taking mirtazapine (0.43 (0.00; 0.86), p = 0.05) and citalopram (0.52 (0.10; 0.93), p = 0.01).

“In models adjusted for agitation and depressed mood, people with mild dementia taking trazodone performed worse on successive MMSE assessments than those taking mirtazapine (0.43 (0.00; 0.86), p = 0.05) and citalopram (0.52 (0.10; 0.93), p = 0.01).”

It may state two separate claims as one.

Patients in the trazodone groups had lower MMSE scores at baseline and more neuropsychiatric symptoms than the comparator groups.

“Patients in the trazodone groups had lower MMSE scores at baseline and more neuropsychiatric symptoms than the comparator groups.”

It may state two separate claims as one.

Results were similar to primary analyses with no difference in cognitive decline between those taking trazodone and mirtazapine (0.21 (-0.08, 0.49)) or citalopram (0.25 (-0.02, 0.52)).

“Results were similar to primary analyses (Appendix 1b) with no difference in cognitive decline between those taking trazodone and mirtazapine (0.21 (-0.08, 0.49)) or citalopram (0.25 (-0.02, 0.52)).”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

HoNOS is a clinician-rated scale with acceptable psychometric properties.

“HoNOS is a clinician-rated scale with acceptable psychometric properties which is usually completed at regular clinical assessments”

• cites [27]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. It may state two separate claims as one.

In mouse models of neurodegenerative disease, trazodone restored protein synthesis, was neuroprotective, restored memory deficits, prevented neurodegeneration, and prolonged survival.

“A recent study found that trazodone has effect on eIF2α-P in mouse models of neurodegenerative disease 2 ; it restored protein synthesis, was neuroprotective, restored memory deficits, prevented neurodegeneration, and prolonged survival in these models of prion disease and frontotemporal dementia.”

• cites [2]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Trazodone may be used as a treatment for depression or other neuropsychiatric symptoms, and these symptoms may be risk factors for or symptoms of worse dementia trajectory.

“As trazodone may be used as a treatment for depression or other neuropsychiatric symptoms, and these symptoms may be risk factors for or symptoms of worse dementia trajectory, 12 we also aimed to take into account the potential confounding effect of neuropsychiatric symptoms.”

• cites [12]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Trazodone is prescribed for non-cognitive symptoms of dementia, which may reflect more rapidly progressive dementia.

“As trazodone is prescribed for non-cognitive symptoms of dementia which may reflect more rapidly progressive dementia, 11 we compared trazodone with citalopram and mirtazapine”

• cites [11]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Funders had no role in the study design and the collection, analysis, and interpretation of data and the writing of the article and the decision to submit it for publication.

“Funders had no role in the study design and the collection, analysis, and interpretation of data and the writing of the article and the decision to submit it for publication.”

Probably not a substantial claim.

MMSE has good psychometric properties for determining disease severity.

“MMSE is a 30point scale assessing global cognitive function regularly recorded by clinicians, has good psychometric properties for determining disease severity, 23 and has been used as a cognitive outcome measure in trials.24”

• cites [23]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

MMSE has been used as a cognitive outcome measure in trials.

“has been used as a cognitive outcome measure in trials.24”

• cites [24]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Citalopram and mirtazapine do not have effect on eIF2α-P.

“antidepressant drugs which do not have effect on eIF2α-P 2 but are prescribed for similar indications.”

• cites [2]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. It may state two separate claims as one.

Any relevant record of the drugs of interest during the study window would be detected by the GATE algorithm, so drug exposure was measured longitudinally.

“Any relevant record of the drugs of interest during the study window would be detected by the GATE algorithm so drug exposure was measured longitudinally.”

• cites [18]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [19]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

In post-hoc sensitivity analyses, the study examined whether confounding by indication affected the results.

“In post-hoc sensitivity analyses, we examined whether confounding by indication affected our results”

Probably not a substantial claim. It may state two separate claims as one.

The main analysis was pre-specified and adjusted for age, sex and the length of time between first and last MMSE scores.

“Our pre-specified main analysis was adjusted for age, sex and the length of time (as a continuous variable) between first and last MMSE scores.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The initial improvement may reflect the combined effect of simultaneous initiation of pharmacological or psycho-social cognition-enhancing treatments, improvement of cognition as a result of the antidepressants alleviating neuropsychiatric symptoms, or practice effect in MMSE performance.

“This improvement may reflect the combined effect of simultaneous initiation of pharmacological or psycho-social cognitionenhancing treatments, improvement of cognition as a result of the antidepressants alleviating neuropsychiatric symptoms, or practice effect in MMSE performance.44”

• cites [44]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The drugs may have been used differentially for depression, meaning that trazodone may have been used for people with treatment resistance, which could have ameliorated the drug's potential cognitive benefit.

“In addition, the drugs may have been used differentially for depression meaning that trazodone may have been used for people with treatment resistance which could have ameliorated the drug's potential cognitive benefit.”

It may state two separate claims as one.

Use of data from three different services covering around 5% of the UK's population, and inclusion of almost all eligible people, ensures that the findings are generalizable and avoids the selection bias common in randomised controlled trials.

“Our use of data from three different services covering around 5% of the UK's population, as well as the inclusion of almost all eligible people in this naturalistic study, ensures that our findings are generalizable and avoids the selection bias common in randomised controlled trials.45,46”

• cites [45]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [46]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The initial improvement from the first to second assessment is consistent with other naturalistic studies examining cognitive function in routine practice.

“and this is consistent with other naturalistic studies examining cognitive function in routine practice.16”

• cites [16]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

Use of mixed linear models is a strong analytic approach as it uses every cognitive assessment to reduce measurement error.

“Our use of mixed linear models is a strong analytic approach as it uses every cognitive assessment to reduce measurement error.”

It may state two separate claims as one.

We could not derive reliable data on concordance with prescribed medication, nor dose, which may modulate any cognitive effect of trazodone.

“Unlike studies which use pharmacy data and scrutinise drug use, we could not derive reliable data on concordance with prescribed medication, nor dose, which may modulate any cognitive effect of trazodone.”

It may state two separate claims as one.

Three other studies support our finding as they did not find cognitive benefit or reduced dementia incidence in people without dementia.

“Three other studies support our finding as they did not find cognitive benefit or reduced dementia incidence in people without dementia.”

• cites [41]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [42]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [43]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim.

Patients with continuing clinical concerns, and likely worse disease course, may be more likely to receive follow-up and cognitive assessment.

“Patients with continuing clinical concerns, and likely worse disease course, may be more likely to receive follow-up and cognitive assessment.”

It may state two separate claims as one.

This is the largest study to examine the effect of trazodone on cognition with any comparator and with the longest follow-up.

“This is the largest study to examine the effect of trazodone on cognition with any comparator and with the longest follow-up”

It may state two separate claims as one.

MMSE assessments were unblinded, meaning that there could be patient or observer bias if clinicians were expecting different cognitive benefit from the drugs of interest.

“Finally, MMSE assessments were unblinded meaning that there could be patient or observer bias if clinicians were expecting different cognitive benefit from the drugs of interest”

Probably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.

Such bias is unlikely as, at the time of these observations, it was not widely thought that these drugs were of cognitive benefit.

“but such effect is unlikely as, at the time of these observations, it was not widely thought that these drugs were of cognitive benefit.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

There is mixed literature on the cognitive effect of citalopram and mirtazapine in dementia, including early preclinical evidence of neuroprotective effects.

“There is mixed literature on the cognitive effect of citalopram and mirtazapine in dementia, including early preclinical evidence of neuroprotective effects.35,36”

• cites [35]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [36]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

The sample consisted of older people for whom a smaller dose is required to achieve the same plasma level, and clinicians were likely to have prescribed sufficient doses targeting clinical effect.

“However, our sample consisted of older people for whom a smaller dose is required to achieve the same plasma level, and clinicians were likely to have prescribed sufficient doses targeting clinical effect.”

Probably not a substantial claim. It may be too vague to check against evidence.

The consistent results between these settings lends support to our conclusions.

“However, the custom in these clinical settings is to follow patients receiving active intervention and the consistent results between these settings lends support to our conclusions.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

There are several potential limitations of routinely collected data resources, including non-representative patient populations, misclassification bias or confounding by indication, meaning that conclusions using such data should be cautious.

“There are several potential limitations of these data resources, including non-representative patient populations, misclassification bias or confounding by indication, 45 meaning that conclusions using such data should be cautious.”

• cites [45]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper may give no reason why it holds.

These factors are unlikely to have confounded the main analyses as they would not have affected the decision about prescribing of trazodone.

“although these factors are unlikely to have confounded our main analyses as they would not have affected the decision about prescribing of trazodone.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Although duration of drug exposure was variable, this was accounted for in the analysis.

“Although duration of drug exposure was variable, this was accounted for in our analysis.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The dose used in the mouse-model study was equivalent to 194 mg daily in humans.

“The dose used in the mouse-model study was equivalent to 194 mg daily in humans”

It may state two separate claims as one.

A cross-sectional observational study in UK Biobank participants without dementia found no cognitive benefit associated with trazodone and instead a slowing of reaction time.

“Secondly, a cross-sectional observational study examining the cognitive effects of a range of medications in participants without dementia in the UK Biobank found no cognitive benefit associated with trazodone and instead a slowing of reaction time, 42”

• cites [42]Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. It may be too vague to check against evidence.

This research is evidence of the potential utility of large scale routinely collected research data for the rapid examination of research hypotheses to guide future experimental studies.

“Our research is evidence of the potential utility of large scale routinely collected research data for the rapid examination of research hypotheses to guide future experimental studies.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

We lacked other information of interest, including data on comorbid medications, hospitalisations, physical health or activity of daily living performance.

“We lacked other information of interest including as we were not able to derive data on comorbid medications, hospitalisations, physical health or activity of daily living performance”