Prevalence of bipolar disorder in a sample of older adults Prevalência de transtorno bipolar em uma amostra de idosos

AIMS: To estimate the prevalence of bipolar disorder in a sample of older adults and to examine associated socio-demographic and clinical factors. METHODS: Cross-sectional population-based study of individuals aged 60 years or older registered with the Family Health Strategy in Porto Alegre, Rio Grande do Sul, Brazil, selected randomly from 30 different basic health units. Participants were subjected to diagnostic assessment for bipolar disorder and suicide risk using the Mini International Neuropsychiatric Interview Plus 5.0.0. Categorical variables were described as absolute and relative frequencies. Quantitative variables were expressed as means and standard deviations. The Pearson chi-square or Fisher’s exact tests were used as appropriate to evaluate potential associations between the independent variables suicide attempt and risk of suicide. To control for possible confounders and assess variables independently associated with the outcome of interest, the strength association among different risk factors was assessed by means of prevalence ratios, which were estimated with a controlled Poisson model or multivariate Poisson regression. The significance level was set at 5% (p≤0.05). RESULTS: The sample comprised 550 older adults. The lifetime prevalence of bipolar disorder in the sample was 5.8% and the point prevalence was 1.5%. In those with bipolar disorder, 59.4% were type I and 40.6% type II. Significant associations were observed between bipolar disorder regardless of type and female gender (prevalence rate [PR] 2.42, 95% confidence interval [CI] 1.01-5.81), living with a partner (PR 2.52, 95%CI 1.21-5.24), history of suicide attempt (PR 3.16, 95%CI 1.53-6.25), and suicide risk (PR 2.98, 95%CI 1.47-6.06). When analyzed each type of bipolar disorder, statistically significant associations were found between age under 70 years and type I bipolar disorder; having companion and type II bipolar disorder; and risk of suicide was associated with both types of bipolar disorder. CONCLUSIONS: Regardless of the type of bipolar disorder, women and those living with a partner were more affected. History of suicide attempts and suicide risk were more frequent in elderly subjects with bipolar disorder than in those without the disorder.


INTRODUCTION
Aging process currently constitutes an unprecedented phenomenon, with major public health implications.Many countries are facing the challenge of dealing with an ever-growing elderly population affected by chronic diseases, such as mood disorders.As of 2012, an estimated 810 million people worldwide were over the age of 60, accounting for 11.5% of the global population; this figure is expected to reach 1 billion in less than a decade and to double by 2050 to represent 22% of the world population [1].
In Brazil, the elderly population was 23.5 million in 2011 (accounting for 12.1% of the population overall), and during the following four decades, this segment of the population has grown at a rate of 3.2% a year [2].In the Southern Brazilian state of Rio Grande do Sul, older adults account for 13.6% of the population, and the state capital, Porto Alegre, has the highest proportion of individuals aged 60 or older living alone in the country (12.8%).In view of current trends in population aging, it is increasingly important to evaluate the health status of older adults and identify disabling conditions, to ensure quality of life for each individual [5].
A history of mental illness may be present in approximately one in six older adults living in the community.These conditions constitute a major clinical and public health issue, and are associated with poor health-related quality of life [6,7].Mood or affective disorders are among the most common psychiatric disorders in the elderly, and are an important cause of loss of independence in this population, as well as causing deterioration of comorbid pathological conditions [3,4] .Within this context, comorbidities, suicidality and risk of suicide, social and professional impairments, and low treatment adherence increase the burden and cost of disease and are implicated in poor prognosis; consequently, mood disorders are associated with increased health services utilization by older adults [7].
Among the mood disorders, bipolar disorder (BD) is a chronic mental illness that constitutes the most severe form of mood disturbance, characterized by mood swings and recurrent depressive and manic episodes during life, and has a significant impact on quality of life [3,4].BD may affect 1-3% of the overall population, and 0.1-1.0% of adults over the age of 60 years [8][9][10][11].Furthermore, it is associated with psychosocial impairment, reduced quality of life, and increased suicide risk and suicide attempts [7][8][9][12][13][14][15].There is a strong association between psychiatric conditions and suicide; major depression and bipolar depression are particularly associated, and are implicated in 65-90% of suicides [15,16].
In the general population, BD is strongly associated with suicidal ideation, suicide attempts, and completed suicide.The prevalence of suicide is 30fold greater among patients with BD [15][16][18][19][20].However, recognition of BD in older adults in primary care settings is considered low or imprecise.This is related to underdiagnosis or diagnostic delay in major depressive episodes and in patients with chronic diseases [21] and to failure to detect BD in patients presenting with depressive episodes [22].Most first episodes of mania occur early in life, but may present at any age, and have been reported in the elderly [11,23].
Epidemiological studies of geriatric BD are scarce; accordingly, the prevalence of BD in older adults varies widely across studies, probably due to issues attributed to the complexity and heterogeneity of the disease [10,11,[24][25][26][27][28][29][30][31][32].There are no precise estimates of the incidence and prevalence of BD in random samples of older adults in Brazil [24], which further hinders the possibility of evaluating potential associations between BD and other variables in the elderly.The present study sought to estimate the prevalence of BD in older adults and examine potential associations of this disorder with suicide risk and socio-demographic parameters.Variables used in this research deriving from a "General Questionnaire", that collected information about sociodemographic and economic data, and a "Psychiatric evaluation", that applied instrument and diagnostic criteria for psychiatric disorders and suicide risk.

METHODS
Categorical variables were described as absolute and relative frequencies.Quantitative variables were expressed as means and standard deviations.For comparison of variable frequencies, the sample-wide prevalence of BD was estimated with a 95% confidence interval.The Pearson chi-square or Fisher's exact tests were used as appropriate to evaluate potential associations between the independent variables suicide attempt and risk of suicide.To control for possible confounders and assess variables independently associated with the outcome of interest, the strength of association among different risk factors was assessed by means of prevalence ratios, which were estimated with a controlled Poisson model or multivariate Poisson regression.The prevalence ratio (PR) with 95% confidence intervals was the effect measure employed.The significance level was set at 5% (p≤0.05), and all analyses were carried out in the software IBM SPSS version 21.0.

RESULTS
Of the initial selected sample of 1,080 older adults, 550 (51%) met the inclusion and exclusion criteria and agreed to participate in the study.All of them signed an informed consent and attended the second phase of the research.Most were female (62.9%), white (64.9%), "young old" (i.e., 62% were aged 60-69), with a low level of educational attainment (44.6% had not completed primary education), and with a low household income (89.1% had a monthly income of ≤ US$ 900) (Table 1).
The univariate analysis found borderlinesignificant associations between having a partner (p=0.062) or being retired (p=0.066) and diagnosis of BD.Subjects who had a partner or companion had a tendency for a higher prevalence of BD than did those without a companion (8.3% vs. 4.0%).The lifetime prevalence of BD in the sample was 5.8%, and 1.5% of all subjects currently had the disorder.Of those with the disorder, 59.4% had type I and 40.6% had type II BD (Table 1).
Suicide attempts and risk of suicide were significantly associated with BD (p=0.001).Older adults with a history of attempted suicide and those at risk of suicide had a higher lifetime prevalence of BD than those who had never attempted suicide (21.6% vs. 4.8%) and those not at risk of suicide (15.7% vs. 4.3%), respectively.After adjustment by the multivariate model, the variables that remained significantly associated with BD were female gender, having a partner or companion, have attempted suicide, and risk of suicide (Table 2).
After stratification of the sample by BD type, risk of suicide remained associated with both types.Regarding the other factors of interest, age under 70 years remained associate with type I BD.Older adults who had a partner or companion had a greater chance of receiving a type II BD diagnostic compared to subjects who did not have a partner (Table 3).

DISCUSSION
This study found a 5.8% lifetime prevalence of BD in a community sample of older adults receiving primary health care in the city Porto Alegre, RS, Brazil.Of those with the disorder, 59.4% had type I and 40.6% had type II BD.As expected, suicide attempts and risk of suicide were associated with both types.
The prevalence of BD in this community sample of elderly is consistent with the current literature.Population-based studies have suggested BD prevalence values ranging from 0.1% to 9.8% in primary health care samples [25][26][27][28].When the prevalence of BD is assessed in psychiatric inpatient samples, this figure rises to 4-17% [11,23].The variability is attributable to differences in diagnostic modalities and methodology across studies.
Considering any diagnosis of BD (i.e., not stratified by type), we observed a 2.42-fold greater prevalence of BD in women.Carlborg et al. [29] conducted a cohort study and found unequal representation of gender in the studied population of patients with BD, with a greater trend in women than in men.Large, populationbased studies attempting to correlate BD types with gender failed to find significant gender differences in the distribution of types I and II [30][31][32], which is consistent with the results of the present study.

7
This cross-sectional study with secondary data analysis used part of data of the broader research "Multidimensional Study of Elderly of Porto Alegre Family Health Strategy (FHS), Brazil", started in 2010, which was approved by the Research Ethics Committee of Pontifical Catholic University of Rio Grande do Sul (PUCRS) (registry 10/04967) and the Public Health Secretariat of Porto Alegre City (registry 499, process 001.021434.10.7).All participants or their legal representatives signed a consent form.The inclusion criteria were age 60 years or older, FHS registration, and living in the catchment Sci Med.2017;27(4):ID28026 3/Original article Porciúncula LR et al. -Prevalence of bipolar disorder in a sample of older adults area of the selected basic health units.Exclusion criteria were not being able to understand the questions for the research protocol, inability to respond coherently to the tests, withdrawal at any time of the survey, and refusing to participate in the stages of collection.Demographic estimative 2008 of the Instituto Brasileiro de Geografia e Estatística was used to plan the sample size for this analysis.At the time of development of protocol, the FHS in Porto Alegre was made up of 97 basic health units, serving 22,000 seniors.Initially, a random sample of 1,080 old aged was selected, composed by 36 individuals from each 30 different FHS centers, covering all regions of the city (North, South, Center and East).After this randomization, community health workers (CHW) in each basic health unit provided the names of the selected sample.Data collection occurred in two phases: 1) screening and general data collected by the CHW in the homes; and 2) expert evaluation carried out at the São Lucas Hospital of PUCRS.The second stage of the study was characterized by assessments and diagnoses made by professionals in the São Lucas Hospital of PUCRS.At this stage diagnostic instruments were used which include Brazilian version of the Mini International Neuropsychiatric Interview Plus 5.0.0 (MINI according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) rating in current time and lifetime diagnosis of psychiatric disorders and suicide risk.MINI Plus include six questions about suicidal ideas and behavior (current and past).All the professional team had experience in elderly patients with neuropsychiatric disorders and participated in the team of the Brain Aging Clinic of PUCRS (Ambulatório de Envelhecimento Cerebral -AMBEC).
Original articlePorciúncula LR et al. -Prevalence of bipolar disorder in a sample of older adults

Table 1 .
Sociodemographic data of the study sample, constituted by 550 older adults registered in the Family Health Strategy in Porto Alegre, RS, Brazil, between 2010-2015.
* 79 subjects (14.4%) refused to report their income.p values refer to bipolar disorder prevalence variations across variables.

Table 2 .
Multivariate Poisson regression analysis of factors independently associated with bipolar affective disorder regardless of its type, in elderly people.