Antidepressants and dementia

Results Persons who purchased antidepressants once ( N = 687,552) had an increased rate of dementia compared to persons unexposed to antidepressants ( N = 779,831). Nevertheless, the rate of dementia changed over time; thus during the initial prescription periods the rate increased with the number of prescriptions but continued long-term antidepressants treatment was associated with a reduction in the rate of dementia, however, not to the same level as the rate for the general population. This pattern was found for all classes of antidepressants (SSRIs, newer non-SSRI antidepressants and older antidepressants). All findings were replicated in sub-analyses with Alzheimer's disease as outcome. Limitations Methodological reasons for the findings cannot be excluded due to the non-randomized nature of data. Conclusions Continued long-term antidepressant treatment was associated with a reduced rate of dementia, however, not to the same level as the rate for the general population. Keywords Antidepressants Depressive disorder Dementia Alzheimer's disease 1 Introduction Studies suggest that depressive disorder is associated with increased risk of developing cognitive dysfunction ( Kessing, 1998; Castaneda et al., 2008 ) and eventually dementia ( Kessing et al., 1999; Kessing and Nilsson, 2003; Ownby et al., 2006 ) and that this risk increases with every new affective episode ( Kessing, 1998; Kessing and Andersen, 2004 ). On the other hand, antidepressants may have neuroprotective abilities by increasing the proliferation of neural progenitors in the subgranulate zone of the hippocampus as well as the survival of these newborn neurons ( Dranovsky and Hen, 2006; Banasr and Duman, 2007 ) and therefore improve memory processes and cognition ( Schmitt et al., 2006 ). Further, it has been suggested that treatment with an SSRI may improve cognitive function and daily living in patients with Alzheimer's dementia ( Mowla et al., 2007 ). However, the association between antidepressant consumption and the risk of developing dementia has never been investigated. It would be difficult to undertake a prospective longitudinal and controlled study investigating this association as a large number of patients with depression without symptoms on dementia and with and without antidepressant treatment would had to be followed longitudinally for at least 5–10 years. An alternative approach is to use register-based data as done in the present study. We hypothesized that continued treatment with antidepressants decreases the risk of developing dementia. The aim of the present study was to investigate whether continued treatment with antidepressants is associated with reduced risk of dementia by linkage of Danish nation-wide registers of all prescribed antidepressants and all diagnoses of dementia given at hospital in- or outpatient settings in Denmark. It was hypothesized that 1) the use of antidepressants is associated with increased risk of dementia (as depressive disorder is associated with dementia) 2) continued treatment with antidepressants is associated with a decreased risk of developing dementia and further that 3) the risk of dementia decreases with the number of prescriptions of antidepressants. 2 Methods 2.1 The registers Data were obtained by linking Danish population-based registers using the unique personal identification number, which is assigned to all 5.3 million persons living in Denmark, thus ensuring accurate linkage of information between registers, irrespective of changes in name etc. ( Malig, 1996 ). In this way, the Medicinal Product Statistics ( Danish National Board of Health, 2002 ) was linked with the Danish Medical Register on Vital Statistics ( Juel and Helweg-Larsen, 1999 ), the Danish National Hospital Register (DNHR, ( Andersen et al., 1999 )), and the Danish Psychiatric Central Register (DPCR, ( Munk-Jorgensen and Mortensen, 1997 )). The Medicinal Product Statistics contains data on all prescribed medication purchased at pharmacies from January 1, 1995 and onwards ( Danish National Board of Health, 2002 ). In Denmark, all medication, such as antidepressants, which is prescribed by doctors, is purchased at pharmacies, only. Information on indication for the drug is not available. The Danish Medical Register on Vital Statistics ( Juel and Helweg-Larsen, 1999 ) contains data on death. The Danish National Hospital Register ( Andersen et al., 1999 ) contains data on all patients treated at all somatic hospitals as in- or outpatients in Denmark from January 1, 1977 and onwards as a part of the official Danish health survey ( World Health Organisation, 1993 ). Likewise, all psychiatric admissions have been registered in a nationwide register, The Danish Psychiatric Central Register (DPCR, ( Munk-Jorgensen and Mortensen, 1997 )) from April 1, 1970 and onwards. Since 1 January 1994 the ICD-10 has been in use in both registers ( World Health Organisation, 1993 ). The study period was from January 1, 1995 to December 31, 2005 although the period prior to 1995 was used to exclude patients with a prior diagnosis of dementia. 2.2 The sample A random sample consisting of 30% of the Danish population was identified in the Danish Medical Register on Vital Statistics among all inhabitants in Denmark who were alive at January 1, 1995. Additionally, all persons who purchased antidepressants at least once in the study period from January 1, 1995 to December 31, 2005 were identified in the Medicinal Product Statistics. Main and auxiliary diagnoses of dementia given following a hospitalization or outpatient contact were identified in the DNHR (ICD-10 codes: G30.0–G30.9) and the DPCR (ICD-10 codes: F.00–F00.9 + F0.1–F01.9) during the period from January 1, 1995 to December 31, 2005. Patients who were recorded in the DNHR or the DPCR with a diagnosis of dementia before start of an antidepressant were excluded from analyses. The period searching for dementia diagnoses included the period prior to 1995 and included ICD-8 diagnoses also (back to 1977 in the DNHR and back to 1970 in the DPCR). 2.3 Statistical analysis Poisson regression analyses were conducted with a diagnosis of dementia as outcome and with the number of prescriptions of antidepressants as the variable of interest and censoring at death or end of the study period (December 31, 2005). Poisson regression models ( Clayton and Hills, 1993 ) are standard multiple regression models for incidence rates where numbers of diagnosis with dementia and person-years at risk are computed and analyzed in subgroups given by covariates which may be time-dependent. Confounding by indication may occur if periods with antidepressant consumption (1 or more prescriptions) were compared with the period with 0 prescriptions, since one of the main indications for an antidepressant is depressive disorder, which is associated with increased rate of developing dementia ( Kessing et al., 1999; Kessing and Nilsson, 2003 ). Thus, we chose to compare periods with multiple prescriptions with the period with one prescription. We have in previous studies on the associations between lithium and suicide ( Kessing et al., 2005 ) and lithium and dementia ( Kessing et al., 2008 ), respectively, used the same kind of analyses to reduce confounding by indication. Risk time was estimated from the age of 40 years. Antidepressants were categorized into three classes, SSRIs, newer non-SSRI antidepressants and older antidepressants (mostly tricyclic antidepressants (TCA)) according to the Anatomical Therapeutical Chemical (ATC) classification system ( World Health Organization, 1999 ), and the number of antidepressant prescriptions was calculated within each class. Sex, age and calendar period in 1-year periods were included in the model as covariates. Further, combinations of antidepressant purchase, also involving classes different from the class of interest, were included in the model as a time-dependent covariate. Finally, all analyses were repeated with a diagnosis of Alzheimer's disease (ICD-10 codes: G30–30.9, F00–00.9) and a diagnosis of other dementia (ICD-10 codes: F01–0.39), respectively, as outcome. 3 Results During the study period from January 1, 1995 to December 31, 2005 a total of 1,467,383 persons who were older than 40 years were included in the study. Among the included persons, 687,552 persons purchased antidepressants at least once (exposed) and 779,831 persons did not purchase antidepressants (unexposed) according to data from the Medicinal Product Statistics. Characteristics of persons exposed and unexposed to antidepressants can be seen from Table 1 . Table 2 shows the use of various classes of antidepressants at the end of the study period (SSRIs, newer non-SSRIs, older antidepressants, and combinations). As can be seen, 50.21% of the persons exposed to antidepressants used SSRIs only. Further, a large proportion of persons used antidepressants from different classes during the study period, e.g. most persons who were exposed to newer non-SSRIs also got an SSRI during the study period. Table 3 presents the rate of dementia according to the use of SSRIs, other newer non-SSRIs and older antidepressants. The rates of dementia is estimated in relation to the number of antidepressant prescriptions within each class (0, 1, 2, 3–5, 6–9, 10–14, 15–19, ≥ 20) and adjusted for age, sex, calendar period and the purchase of antidepressants from another class. Purchase of antidepressants once was chosen as the reference. The table presents four major findings: 1) Persons unexposed to antidepressants had a decreased rate of dementia compared to persons who purchased antidepressants once (e.g., RR = 0.30 (95% CI: 0.29–0.31) for persons unexposed to SSRIs). Conversely, persons who purchased antidepressants once had an increased rate of dementia compared to persons unexposed to antidepressants (RR = 3.36 (95% CI: 3.25–3.48) for persons exposed to SSRIs once; RR = 4.74 (95% CI: 4.43–5.07) for persons exposed to newer non-SSRI antidepressants once; RR = 1.77 (95% CI: 1.67–1.86) for persons exposed to older antidepressants once). 2) Compared to the rate of dementia during the period with one prescription of antidepressants, the rate of dementia was increased during subsequent prescription periods but decreased again during periods with multiple prescriptions (≥ 10 prescriptions for SSRIs, ≥ 6 prescriptions for newer non-SSRI antidepressants and older antidepressants). 3) Although the rate of dementia was decreased during periods with multiple prescriptions the rate did never decrease to the level among persons unexposed to antidepressants 4) The rate of dementia was in general decreased for patients using older antidepressants compared to users of other classes of antidepressants. Figs. 1–3 present these findings graphically for every new purchase of antidepressants. Finally, all analyses were repeated with Alzheimer's disease as outcome and with dementia of other kind as outcome, respectively. A total of 7.532 persons got a diagnosis of Alzheimer's disease for the first time ever, during the study period and 27,253 persons a diagnosis of other dementia ( Table 1 ). The findings were generally the same with these two outcome measures as when a diagnosis of dementia was the outcome ( Table 3 ). 4 Discussion This is the first study investigating the association between continued treatment with antidepressants and the risk of subsequently developing dementia. We confirmed our first hypothesis but not, or only partly, our second and third hypotheses, i.e., 1) the use of antidepressants was associated with increased risk of dementia but 2) continued treatment with antidepressants was not systematically associated with a decrease in the risk of developing dementia and 3) the risk of dementia did not steadily decrease with the number of prescriptions of antidepressants. These findings were replicated in analyses with Alzheimer's disease and dementia of other kinds, respectively, as outcome. In relation to hypotheses two and three, the rate of dementia was increased during periods with 2 to 6–10 prescriptions but it should be noted that the rate decreased subsequently during periods with multiple prescriptions, however not to the level for persons not exposed to antidepressants. These findings are partly in contrast to our recent findings regarding lithium and dementia ( Kessing et al., 2008 ). In similar analyses as those undertaken in the present study, we found that continuing taking lithium, i.e., purchasing more than one prescription, decreased the rate of dementia to the same level as the rate for the general population. We used register data of all purchased antidepressants in Denmark including prescriptions by specialist within hospital, private practice settings and general practitioners in a period from 1995 to 2005. Thus, our results pertain to all patients treated with antidepressants nation-wide. The register contains no data on indications for treatment but the vast majority of antidepressants are prescribed for depressive disorder ( Claxton et al., 2000 ). Similarly, the register contains no data on the prescribed dose of the antidepressant. It is likely that older patients may have been prescribed lower doses of antidepressants, particularly regarding the non-SSRIs and TCAs, but it is not clear how this might have affected our findings. During the 10 years study period, the use of antidepressants such as SSRIs and newer non-SSRI antidepressants has increased mainly due to the improved side effect profile. All analyses in the study were adjusted for the effect of calendar period in 1-year periods — in this way also taking account for any changes in the health care system during the study period. ICD-10 diagnoses ( World Health Organisation, 1992 ) were used in Denmark during the entire study period. The validity of the diagnosis of dementia in the DNHR and the DPCR has been found to be high with a correct register diagnosis in 85.8% of the cases ( Phung et al., 2007 ). The ICD-10 subtype diagnosis of dementia was rather low, but Alzheimer's disease, although underregistered, also had a good validity once the diagnosis was registered. It is estimated that at least two-thirds of patients with dementia in Denmark will get the diagnosis at a contact to secondary hospital health care ( Phung et al., 2008 ). Nevertheless, it should be noted that general practitioners or private practicing specialists in psychiatry or neurology might diagnose persons with dementia without this being recorded in the registers. In this way, it is most likely that only patients with the more severe forms of dementia are recorded in the registers. The Danish population is ethnically and socially homogeneous and with a very low migration rate. Psychiatric and medical care is well developed so persons can easily come in contact with general practitioners or specialists in psychiatry or neurology. Psychiatric and other medical treatment is available free of charge in Denmark, and as antidepressants have a 75% refunding from the state, the study is not likely to be biased by socio-economic differences. 4.1 Interpretation of the results The first finding in our study is in accordance with our previous finding of an increased risk of developing dementia among patients who got a diagnosis of depressive disorder during psychiatric hospitalization ( Kessing et al., 1999; Kessing and Nilsson, 2003 ), as antidepressants mainly are prescribed for depressive disorders, and further, for the first time ever, generalize our prior findings to the much larger population of all patients treated with antidepressants in Denmark. Although continued use of antidepressants did not decrease the rate of dementia to the rate for the general population continued long-term use was in fact associated with a somewhat decreasing rate. During periods with prescriptions of 15th or more of newer non-SSRIs or older antidepressants the rate decreased to the level when persons purchased their first prescription (RR is around 1). As data from our prior studies suggest that the rate of dementia increases with the number of depressive episodes in patients with depressive disorder ( Kessing and Andersen, 2004 ) it is possible that the use of antidepressants at least partly counterbalance this potential deteriorating effect of the illness. For persons who purchased older antidepressants, the rate of dementia was generally only slightly increased compared to the rate among users of other antidepressants. It is likely that older antidepressants, especially tricyclic antidepressants (TCA), are not prescribed to patients whom the clinician believe will develop dementia later on, e.g. patients with mild cognitive impairment as clinicians may believe that TCAs may induce confusion in such patients. Nevertheless, also among users of older antidepressants the rate of dementia was increased compared to the rate among the general population (RR = 1.77 (95%CI: 1.67–1.86). Further, the way the rate of dementia changed over time — with an initial increase and a subsequent decrease — was the same, although less pronounced, as for the other two classes of antidepressants. To sum up, there are two possible explanations to our findings: 1) Continued treatment with antidepressants decreases the rate of developing dementia somewhat, however not to the rate among the general population. This effect may be mediated directly due to neuroprotective abilities or other mechanisms, or indirectly by decreasing the risk of developing depressive episodes associated with cognitive decline. 2) The finding is a methodological artifact: firstly, patients who continue taking antidepressants for long periods may presumably be good compliers with improved cognitive function, less alcohol use and a healthier life style associated with decreased risk of developing dementia. However, this explanation cannot systematically explain our findings as the rate of dementia changed over time with an increase during periods with 2 to 6–10 prescriptions and a subsequent decrease during periods with multiple prescriptions. Secondly, patients with cognitive impairment may experience more side effects, such as nausea or headache, associated with antidepressants and consequently discontinue antidepressant treatment more often. This does, however, not seem to be the case, as antidepressants, and perhaps especially SSRIs, are well tolerated among patients with symptoms of dementia ( Sink et al., 2005 ). Thirdly, our finding could be a result of a reverse causation as patients with dementia often are prescribed antidepressants ( Kessing et al., 2007 ). Although this possibility cannot be excluded, it should be emphasized that we excluded all patients who got a diagnosis of dementia before start of an antidepressant. Thus, although we do not find it likely that our results can be explained fully by methodological drawbacks we cannot exclude this possibility due to the non-randomized nature of our data. Stress induced reductions in adult neurogenesis in animals has been reversed by a number of different antidepressants, including fluoxetine ( Malberg and Duman, 2003; Chen et al., 2006; Banasr et al., 2007 ), escitalopram ( Jayatissa et al., 2006 ) and desimipramine ( Chen et al., 2006 ). It was not possible to estimate the rate of dementia for single antidepressant drugs in the present study due to reduced statistical power. It is not possible from the present study to decide whether antidepressants has a partly protective effect against dementia due to increased neurogenesis or other mechanisms or due to mood stabilizing abilities preventing recurrence of affective episodes or other treatment related factors. In naturalistic data, as used in the present study, it is not possible validly to investigate the association between treatment, such as antidepressants, and the course of illness, due to confounding by indication as patients with many episodes tend to take antidepressants for longer time. Or as expressed by Keller et al. (1983) , in naturalistic studies, “treatment itself becomes an outcome, because patient's state [is] likely to help determine the choice of treatment.” 4.2 Conclusion In a nationwide study including all patients treated with antidepressants, the rate of developing dementia was increased compared to the rate among persons unexposed to antidepressants. Nevertheless, the rate of dementia changed over time; thus during the initial prescription periods the rate increased with the number of prescriptions but continued long-term antidepressants treatment was associated with a reduction in the rate of dementia, however, not to the same level as the rate for the general population. Methodological reasons for this finding cannot be excluded due to the non-randomized nature of data. Role of funding source Nothing declared. Conflict of interest No conflict declared. Acknowledgement Lars Vedel Kessing had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. References Andersen et al., 1999 Andersen T.F. Madsen M. Jorgensen J. Mellemkjaer L. Olsen J.H. The Danish National Hospital Register. A valuable source of data for modern health sciences Dan. Med. Bull. 46 1999 263 268 Banasr and Duman, 2007 Banasr M. Duman R.S. Regulation of neurogenesis and gliogenesis by stress and antidepressant treatment CNS Neurol. Disord. Drug Targets 6 2007 311 320 Banasr et al., 2007 Banasr M. Valentine G.W. Li X.Y. Gourley S.L. Taylor J.R. Duman R.S. Chronic unpredictable stress decreases cell proliferation in the cerebral cortex of the adult rat Biol. Psychiatry 62 2007 496 504 Castaneda et al., 2008 Castaneda A.E. Tuulio-Henriksson A. Marttunen M. Suvisaari J. Lonnqvist J. A review on cognitive impairments in depressive and anxiety disorders with a focus on young adults J. Affect. Disord. 106 2008 1 27 Chen et al., 2006 Chen H. Pandey G.N. Dwivedi Y. Hippocampal cell proliferation regulation by repeated stress and antidepressants Neuroreport 17 2006 863 867 Claxton et al., 2000 Claxton A.J. Li Z. McKendrick J. Selective serotonin reuptake inhibitor treatment in the UK: risk of relapse or recurrence of depression Br. J. Psychiatry 177 2000 163 168 Clayton and Hills, 1993 Clayton D. Hills M. Statistical Models in Epidemiology 1993 Oxford University Press Oxford Danish National Board of Health, 2002 Danish National Board of Health, L. Medicinal Product Statistics. www.laegemiddelstyrelsen.dk . 2002. Ref Type: Electronic Citation. Dranovsky and Hen, 2006 Dranovsky A. Hen R. Hippocampal neurogenesis: regulation by stress and antidepressants Biol. Psychiatry 59 2006 1136 1143 Jayatissa et al., 2006 Jayatissa M.N. Bisgaard C. Tingstrom A. Papp M. Wiborg O. Hippocampal cytogenesis correlates to escitalopram-mediated recovery in a chronic mild stress rat model of depression Neuropsychopharmacology 31 2006 2395 2404 Juel and Helweg-Larsen, 1999 Juel K. Helweg-Larsen K. The Danish registers of causes of death Dan. Med. Bull. 46 1999 354 357 Keller et al., 1983 Keller M.B. Lavori P.W. Endicott J. Coryell W. Klerman G.L. “Double depression”: two-year follow-up Am. J. Psychiatry 140 1983 689 694 Kessing, 1998 Kessing L.V. Cognitive impairment in the euthymic phase of affective disorder Psychol. Med. 28 1998 1027 1038 Kessing and Andersen, 2004 Kessing L.V. Andersen P.K. Does the risk of developing dementia increase with the number of episodes in patients with depressive disorder and in patients with bipolar disorder? J. Neurol. Neurosurg. Psychiatry 75 2004 1662 1666 Kessing and Nilsson, 2003 Kessing L.V. Nilsson F.M. Increased risk of developing dementia in patients with major affective disorders compared to patients with other medical illnesses J. Affect. Disord. 73 2003 261 269 Kessing et al., 1999 Kessing L.V. Olsen E.W. Mortensen P.B. Andersen P.K. Dementia in affective disorder: a case-register study Acta Psychiatr. Scand. 100 1999 176 185 Kessing et al., 2005 Kessing L.V. Sondergard L. Kvist K. Andersen P.K. Suicide risk in patients treated with lithium Arch. Gen. Psychiatry 62 2005 860 866 Kessing et al., 2007 Kessing L.V. Harhoff M. Andersen P.K. Treatment with antidepressants in patients with dementia—a nationwide register-based study Int. Psychogeriatr. 19 2007 902 913 Kessing et al., 2008 Kessing L.V. Søndergaard L. Forman J.L. Andersen P.K. Lithium treatment and risk of dementia Arch Gen Psychiatry 65 2008 1331 1335 Malberg and Duman, 2003 Malberg J.E. Duman R.S. Cell proliferation in adult hippocampus is decreased by inescapable stress: reversal by fluoxetine treatment Neuropsychopharmacology 28 2003 1562 1571 Malig, 1996 Malig C. The civil registration system in Denmark. Technical paper No 66 1996 Mowla et al., 2007 Mowla A. Mosavinasab M. Haghshenas H. Haghighi A.B. Does serotonin augmentation have any effect on cognition and activities of daily living in Alzheimer's dementia? A double-blind, placebo-controlled clinical trial J. Clin. Psychopharmacol. 27 2007 484 487 Munk-Jorgensen and Mortensen, 1997 Munk-Jorgensen P. Mortensen P.B. The Danish psychiatric central register Dan. Med. Bull. 44 1997 82 84 Ownby et al., 2006 Ownby R.L. Crocco E. Acevedo A. John V. Loewenstein D. Depression and risk for Alzheimer disease: systematic review, meta-analysis, and metaregression analysis Arch. Gen. Psychiatry 63 2006 530 538 Phung et al., 2007 Phung T.K. Andersen B.B. Hogh P. Kessing L.V. Mortensen P.B. Waldemar G. Validity of dementia diagnoses in the Danish hospital registers Dement. Geriatr. Cogn. Disord. 24 2007 220 228 Phung et al., 2008 Phung, T.K., Waltoft, B.D., Kessing, L.V., Mortensen, P.B., and Waldemar, G., Is dementia underdiagnosed? A nationwide population-based study of incidence and prevalence of dementia using national registers. 2008. Ref Type: Unpublished Work. Schmitt et al., 2006 Schmitt J.A. Wingen M. Ramaekers J.G. Evers E.A. Riedel W.J. Serotonin and human cognitive performance Curr. Pharm. Des. 12 2006 2473 2486 Sink et al., 2005 Sink K.M. Holden K.F. Yaffe K. Pharmacological treatment of neuropsychiatric symptoms of dementia: a review of the evidence JAMA 293 2005 596 608 World Health Organisation, 1992 World Health Organisation The ICD-10 Classification of Mental and Behavioural Disorders: Clinical Descriptions and Diagnostic Guidelines 1992 World Health Organisation Geneva World Health Organisation, 1993 World Health Organisation Klassifikation af sygdomme Systematisk del. International Statistical Classification of Diseases and Health Related Problems Tenth revision (Danish Edition) 1993 Munksgaard Copenhagen World Health Organization, 1999 World Health Organization World Health Organization Guidelines for ATC Classification for DDD Assignment 1999 WHO Collaboration Centre for Drug Statistics Methodology, Nordic Council on Medicines Oslo