Psychiatric Disorders Before and After Dementia Diagnosis

Key Points Question What are the temporal risk patterns of psychiatric disorders, including depression, anxiety, stress-related disorders, substance use disorders, sleep disorders, somatoform/conversion disorders, and psychotic disorders, before and after dementia diagnosis compared with individuals without dementia? Findings In this nationwide cohort study of 796 505 participants conducted in Sweden, compared with participants without dementia, the overall risk of new onset psychiatric disorders was significantly higher among patients with dementia. The risk increased markedly from 3 years before receipt of dementia diagnosis, peaked during the week after diagnosis, and then declined rapidly. Meaning These results suggest that managing psychiatric comorbidities is crucial for individuals with dementia across various disease stages.


Introduction
Dementia was estimated in 2016 to cause 28.8 million disability-adjusted life years across all ages 1 and is the second largest cause of death among individuals older than 70 years. 2Increased longevity has led to an increasing number of individuals living with dementia and a significant burden on social welfare, health care, and financial systems. 3Persons with dementia have elevated rates of premature mortality 4 and increased risk of suicide, 5 especially among patients with dementia having psychiatric comorbidities. 6,7e co-occurrence of dementia and psychiatric disorders, such as depression and anxiety, are common in clinical practice. 8,9While there is substantial research on depression, 10,11 most of the evidence is limited to the period prior to dementia diagnosis. 12There has been comparatively little investigation addressing the period immediately following a dementia diagnosis.Furthermore, other psychiatric disorders, including substance use disorders, 9 stress reaction or adjustment disorders, 13 and sleep disorders, 14 have received less attention.
Accumulating evidence suggests that the period during the receipt of a diagnosis of dementia can be a difficult. 15However, evidence regarding how psychiatric disorders develop after a dementia diagnosis is limited by small sample sizes and short follow-ups. 16,17A longitudinal description characterizing the burden of psychiatric disorders among individuals with dementia is lacking.
Moreover, previous findings relating to rarer dementias, such as frontotemporal dementia (FTD), 18 dementia with Lewy bodies, 14,19,20 and Parkinson disease dementia 21 have been inconclusive due to small sample size, cross-sectional study design, and a short follow-up period.The time-dependent risk patterns of incident psychiatric disorders at different stages of dementia are largely unknown.
Information on such conditions may improve timely assessment and early intervention in this growing group of patients.
In this cohort study, we examined the hypothesis that patients with dementia are at increased risk of psychiatric disorders both before and after dementia diagnosis.

Study Design
Based on nationwide Swedish registers, we conducted a population-based cohort study between January 1, 2000, and December 31, 2017, to compare the time-dependent risk of psychiatric disorders between patients with dementia and individuals without dementia during the periods before and after receipt of dementia diagnosis.The requirement for written informed consent for this study was waived due to the register data being pseudonymized before delivery to our research group.The regional ethics committee in Stockholm approved the study, which complied with the Declaration of Helsinki. 22 The postdiagnostic period was defined as the period from the date of dementia diagnosis onward.To examine the association of dementia disorder progression with the development of new psychiatric disorders, we excluded 14 481 patients who had pre-existing psychiatric comorbidities at the time of enrollment in the cohort (ie, 7 years before the date of diagnosis), leaving 209 245 patients with dementia included in this study.We defined the index date as the date of first diagnostic record of a dementia diagnosis for patients with dementia.

Control Participants
The control participants were selected from the Total Population Register.For each individual with dementia, up to 4 control individuals were selected based on year of birth (±3 years), sex, and region of residence.Control participants were excluded if they were not free of psychiatric comorbidities at entry to the cohort (ie, 7 years before the date of the dementia diagnosis receipt of the matched individuals with dementia).For control participants, the same index date as the matched individual with dementia was assigned.
The study followed up all participants from 7 years prior to the index date (earliest possible entry on January 1, 2000), until a first diagnosis of psychiatric disorder, death, or December 31, 2017, whichever occurred first.

Dementia Diagnosis
Inclusion Dementia with Lewy bodies and Parkinson disease dementia were merged for this study as Lewy body disease (LBD) considering the shared pathological and clinical characteristics. 23,24

Psychiatric Disorders and Medication
We used ICD-10 codes from the National Patient Register to identify any first inpatient or outpatient diagnosis of common psychiatric disorders during follow-up.Psychiatric disorders included depression, anxiety, stress-related disorders, substance use disorders, sleep disorders, somatoform/ conversion disorders, and psychotic disorders (eTable 2 in Supplement 1).To capture conditions that may not be documented in the National Patient Register, we performed a subgroup analysis to examine the use of psychiatric medications associated with psychiatric disorders as assessed by dispensation of prescription medication (eMethods and eTable 3 in Supplement 1).

Covariates
Socioeconomic characteristics, including educational attainment, marital status, disposable individual income, region of birth, and coresident status were extracted from the longitudinal integrated database for health insurance and labor market studies.The dates of death were extracted from the Cause of Death Register.

Statistical Analysis
We first compared the characteristics of patients with dementia and the control participants at the index date.We then investigated the time-dependent associations between dementia diagnosis and the risk of psychiatric disorders by flexible parametric survival models (Royston-Parmar models). 27spline with 5 df was used for the baseline rate, while 3 df was used for the time-varying effect.We used time since cohort entry as the underlying timescale.Hazard ratios (HRs) and 95% CIs were estimated separately for prediagnostic and postdiagnostic periods.We excluded 26 564 patients with dementia and 34 083 control participants diagnosed as having a psychiatric disorder during the prediagnostic period from the analysis of the postdiagnostic period.To compare the absolute risk of psychiatric disorders between patients with vs without dementia, we plotted cumulative incidence curves for each individual disorder.Considering a high mortality among patients with dementia after diagnosis, competing risk models were applied in the postdiagnostic period.
The proportions for receiving treatment of the studied psychiatric medications from 5 years before dementia diagnosis to 5 years after dementia diagnosis were calculated daily for patients with dementia and control participants.The differences between the 2 groups were examined by χ  1 and Table 2).
The results were largely similar for AD, mixed dementia, vascular dementia and unspecified dementia (Figure 2).The HRs were greater for LBD and FTD from 2 years before diagnosis (LBD HR, We also observed similar results for individual psychiatric disorders, except sleep disorders and somatoform/conversion disorders (eFigure 5 in Supplement 1).Stress-related disorders and psychotic disorders exhibited the most substantial rate increase immediately following dementia diagnosis.
In the subgroup analysis, we also found that the use of antidepressants was persistently higher among patients with dementia compared with controls, and the difference increased from 2 years before dementia diagnosis (15.9% vs 7.9%, P < .001),peaked approximately 6 months after dementia diagnosis (29.1% vs 9.7%, P < .001),and then decreased slowly from 3 years after diagnosis but remained higher than controls 5 years after diagnosis (16.4% vs 6.9%, P < .001)(Figure 3 and eTable 11 in Supplement 1).For anxiolytics and antipsychotics, the increased use was observed 2 years (8.8% vs 6.0%, P < <.001) and 1 year (2.3% vs 0.5%, P < .001)before dementia diagnosis and remained elevated thereafter.The use of hypnotics or sedatives was persistently higher among patients with dementia compared with controls before diagnosis, with the increased difference starting 6 months before diagnosis (18.5% vs 16.1%, P < .001)and decreasing slowly thereafter.A reversed pattern was observed 4 years after diagnosis (9.5% vs 11.2%, P < .001).Similar results were noted for dementia subtypes (eFigures 6, 7, 8, and 9 in Supplement 1).

Discussion
In this nationwide population-based cohort study, we found that compared with participants without dementia, new-onset psychiatric disorders were more common among patients with dementia both before and after receipt of a diagnosis, especially among patients diagnosed as having LBD and FTD.
Patients with dementia had greatly increased risks of almost all psychiatric disorders assessed, including depression, anxiety, stress-related disorders, substance use disorders, and psychotic disorders, as early as 7 years prior to diagnosis.Similar patterns were observed in AD, mixed dementia, vascular dementia, and unspecified dementia.Risk factors associated with psychiatric disorders among dementia patients were being younger, being better educated, receiving a diagnosis in specialist care, and having fewer physical comorbidities or higher MMSE score.
12]28 Several psychiatric disorders have long been associated with cognitive impairment, and there is a growing recognition that cognitive deficits are not solely a consequence of mood disturbances in depression but may persist even after clinical recovery. 29The severity and persistence of enduring deficits have also been emphasized. 30rthermore, depression has been identified as a risk factor for future dementia, and this association cannot be attributed solely to misdiagnosis of early cases of dementia. 11The association persists even for individuals who exhibited depressive symptoms for longer than 2 decades (28 years) prior to the onset of dementia. 11w psychiatric symptoms after the dementia diagnosis, known as neuropsychiatric symptoms, are prevalent among people with AD and related dementias and are associated with the course of disease. 31Previous studies have found that neuropsychiatric symptoms are mostly attributed to dementia symptomatology, which makes looking for new psychiatric diagnoses in a patient with dementia less relevant. 17,32This finding was also confirmed in our study by the high prevalence of psychiatric symptoms during the dementia disease progression and for all dementia disorders.However, it is necessary to differentiate generalized from disorder-specific risk factors over the disease course by evaluating the full range of psychiatric conditions prospectively.
We also found that patients with LBD and FTD had higher risk of psychiatric comorbidities compared with other dementia subtypes, which is in line with previous studies. 33,34These findings suggested that observing neuropsychiatric symptoms could be in favor of diagnoses of LBD and FTD.
The high occurrence of psychiatric disorders and the difference across dementia subtypes highlight the importance of reducing the psychiatric health burden among patients with all dementia subtypes.
To our knowledge, our study constitutes the most comprehensive endeavor to date to quantify associations and to identify temporal patterns of new onset or diagnosis of psychiatric conditions among individuals with dementia before, during, and after diagnosis receipt.Our finding of the timedependent risk pattern is in line with previous studies focusing on health care costs of dementia and injurious falls. 9,35,36A Swedish population-based study reported that patients with dementia exhibited an elevated occurrence of injurious falls, with the highest incidence observed 4 years prior to the diagnosis, peaking in the year of diagnosis, and declining rapidly during the 4 years after diagnosis. 36A case-control study in Germany also found an increase in the use of ambulatory medical care services by patients with dementia of 50% in the year before and of 40% in the year after the incidence, predominantly in primary care and neurology or psychiatry settings. 35Our finding of significant rise in risk observed immediately following a dementia diagnosis and decreased risk several years after diagnosis is consistent with earlier studies that have reported a substantially higher risk of suicide immediately following a dementia diagnosis and a lower risk thereafter. 6,37The markedly elevated risks observed in the year leading up to the diagnosis may be associated with the pre-existing symptoms of dementia and the significant psychological burden of receiving clinical assessment for a suspected disease. 6,37

Strengths and Limitations
This study has several strengths, including a large and national sample, a population-based cohort design, long follow-up, and a range of different types of dementia.Furthermore, sociodemographic characteristics, medical disorders, specific dementia subtypes, characteristics related to dementia, and psychiatric medications were also explored.
This study has limitations.The diagnoses of different types of dementia were only available for patients from SveDem.According to incidence estimates, SveDem captured almost one-third of all expected new dementia cases in Sweden.However, assessing the accuracy of these diagnoses is difficult, although they are consistent with standard clinical practice and comply with national guidelines and do not change with follow-up. 38Furthermore, the assessment of the study's outcomes is based on clinical diagnoses obtained through specialized inpatient and outpatient care, which primarily represents more severe manifestations of the studied psychiatric disorders.In Sweden, first-line psychiatric care is administered through primary care, for which we did not have data.However, including psychiatric medication prescriptions (which should have almost complete coverage) was our attempt to compensate for this.The comparable temporal patterns observed in the augmented use of psychiatric medications (although with a more pronounced effect) confirm the high prevalence of psychiatric symptoms during dementia disease progression and suggest that the diagnosis of dementia may be associated with both severe and mild psychiatric conditions.
Therefore, the combined use of both clinical diagnoses and medication use may capture the overall health burden during the period of diagnosis.However, the validity of using data on prescriptions for psychiatric medications as a surrogate diagnostic marker of the underlying diagnoses needs to be further explored.In addition, co-occurring psychiatric disorders are common among patients with dementia.Patients who develop early dementia or psychiatric illness symptoms or are diagnosed formally with these illnesses will contact medical settings more frequently and thus be more likely to receive a diagnosis of concurrent partner illnesses.The extent to which this possibility may be associated with our findings remains uncertain.

JAMA Network Open | Psychiatry
Participants and caretakers were informed verbally and in writing about the Swedish Registry for Cognitive/Dementia Disorders (SveDem) and could decline participation.This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.Patients With Dementia Using SveDem, the National Patient Register, and the Swedish Prescribed Drug Register, we identified 223 726 patients (Ն40 years old) with a first diagnosis of dementia between May 1, 2007, JAMA Network Open | Psychiatry Psychiatric Disorders Before and After Dementia Diagnosis JAMA Network Open.2023;6(10):e2338080. doi:10.1001/jamanetworkopen.2023.38080(Reprinted) October 17, 2023 2/14 Downloaded From: https://jamanetwork.com/ on 10/22/2023 and December 31, 2017.Details of the 3 register database are presented in the eMethods in Supplement 1.Because SveDem contains records since 2007 and the National Patient Register contains records on both inpatient and outpatient care since 2001, the maximum time for obtaining psychiatric comorbidities before dementia diagnosis for patients diagnosed in 2007 was 7 years.For this reason, we defined the prediagnostic period as 7 years before the date of dementia diagnosis.
criteria were as follows: diagnosis of dementia in SveDem; or first record of a dementia defined using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes F00 to F03, G30, and G31 in the National Patient Register; or record of the Anatomical Therapeutic Chemical classification code N06D (dementia medication) in the Prescribed Drug Register of Sweden (eTable 1 in Supplement 1).For patients with dementia identified from SveDem, dementia diagnoses were coded as Alzheimer disease (AD), mixed dementia, vascular dementia, dementia with Lewy bodies, FTD, Parkinson disease dementia, and unspecified dementia.

Figure 2 .
Figure 2. Hazard Ratios (HRs) and 95% CIs of Psychiatric Disorders Before and After Dementia Diagnosis in a Matched Cohort Study, by Dementia Types in Sweden, 2000 to 2017

Figure 3 .
Figure 3. Use of Antidepressants, Anxiolytics, Hypnotics and Sedatives, and Antipsychotics Before and After Dementia Diagnosis in a Matched Cohort Study in Sweden, 2006 to 2017

Table 2 .
Hazard Ratios and 95% CIs of Psychiatric Disorders Among Patients With vs Those Without Dementia After Dementia Diagnosis, by Participant Characteristics a a All models were adjusted for age, sex, educational attainment (<9 years, 9-12 years, b Disease status within 3 years before the diagnosis of dementia disease.cAmongpatients from SveDem and matched controls; MMSE score categories, with higher scores indicating milder cognitive impairment.JAMA Network Open | PsychiatryPsychiatric Disorders Before and After Dementia Diagnosis JAMA Network Open.2023;6(10):e2338080. doi:10.1001/jamanetworkopen.2023.38080(Reprinted) October 17, 2023 7/14 Downloaded From: https://jamanetwork.com/ on 10/22/2023 Flowchart of study patients with dementia and control participants eFigure 2. Cumulative incidences and 95% confidence intervals of individual psychiatric disorders between people with and without dementia before and after dementia diagnosis in the study population in Sweden, 2000 to 2017 eFigure 3. Hazard ratios and 95% CIs of individual psychiatric disorders before and after dementia diagnosis in patients from SveDem and matched controls, 2000 to 2017 eFigure 4. Hazard ratios and 95% CIs of individual psychiatric disorders before and after dementia diagnosis in patients from other registers and matched controls, 2000 to 2017 eFigure 5. Hazard ratios and 95% CIs of individual psychiatric disorders before and after dementia diagnosis in a matched cohort study in Sweden, 2000 to 2017 eFigure 6. Use of antidepressants by dementia types before and after a specific dementia diagnosis in patients from SveDem and matched controls, 2006 to 2017 JAMA Network Open | Psychiatry Psychiatric Disorders Before and After Dementia Diagnosis Use of anxiolytics by dementia types before and after a specific dementia diagnosis in patients from SveDem and matched controls, 2006 to 2017 eFigure 8. Use of hypnotics/sedatives by dementia types before and after a specific dementia diagnosis in patients from SveDem and matched controls, 2006 to 2017 eFigure 9. Use of antipsychotics by dementia types before and after a specific dementia diagnosis in patients from SveDem and matched controls, 2006 to 2017 eTable 1. ICD-10 and ATC codes of inclusion criteria for patients with dementia from the National Patient Register and the Swedish Prescribed Drug Register, and exclusion criteria for dementia-free subjects eTable 2. ICD-10 codes for psychiatric disorders eTable 3. ATC codes for psychiatric medications eTable 4. Characteristics of dementia patients and the control participants at the index date eTable 5. Distribution of psychiatric disorders, person-years and incidence rates in a matched cohort study in Sweden, 2000 to 2017 eTable 6. Distribution of psychiatric disorders, person-years and incidence rates in patients from SveDem and matched controls, 2000 to 2017 eTable 7. Distribution of psychiatric disorders, person-years and incidence rates in patients from other registers and matched controls, 2000 to 2017 eTable 8. Distribution of psychiatric disorders, person-years and incidence rates by a specific dementia diagnosis in patients from SveDem and matched controls, 2000 to 2017 eTable 9. Hazard ratios and 95% CIs of psychiatric disorders before a specific dementia diagnosis in patients from SveDem and matched controls, 2000 to 2017 a eTable 10.Hazard ratios and 95% CIs of psychiatric disorders after a specific dementia diagnosis in patients from SveDem and matched controls, 2000 to 2017 a eTable 11.Use of antidepressants, anxiolytics, hypnotics/sedatives and antipsychotics before and after dementia diagnosis in a matched cohort study in Sweden, 2006 to 2017 a JAMA Network Open.2023;6(10):e2338080. doi:10.1001/jamanetworkopen.2023.38080(Reprinted) October 17, 2023 13/14 Downloaded From: https://jamanetwork.com/ on 10/22/2023 eFigure 7.