ABSTRACT

Objective:

To explore the relationship between different patterns of multimorbidity and the use of sleeping medications in women.

Methods:

Population-based cross-sectional study with 1,128 women (aged 20–69 years) in Southern Brazil. Data on sleeping medications were obtained from the question "Do you take/use any medication to be able to sleep?" and identified by the Anatomical Therapeutic and Chemical classification. Multimorbidity patterns were derived by the Principal Component Analysis of 26 chronic conditions and two obesity parameters (≥30 kg/m²; ≥40 kg/m²). The association was analyzed by Poisson regression with robust variance using different adjustment models, stratified by age.

Results:

Three multimorbidity patterns were derived: cardiometabolic, endocrine-articular, and psychosomatic. Age stratification showed a change in effect in the relationship investigated. Women under 45 years and high score of cardiometabolic and endocrine-articular patterns were about twice as likely to use sleeping medications [prevalence ratio (PR) 1.85, 95% confidence interval (CI) 1.09–3.12; PR 2.04, 95%CI 1.18–3.51, respectively]. Those with psychosomatic pattern were around five times more likely [PR 4.91, 95%CI 3.00–8.04].

Conclusions:

The study provided the first evidence on the association researched and demonstrated that young women (<45 years) with a high score of the identified patterns are up to five times more likely to use sleeping medications, configuring early use. This unprecedented finding suggests the need for greater health promotion for young adults and actions to raise awareness about risks and the clear indication of the use of sleeping medications.

Keywords:
Multimorbidity; Multiple chronic conditions; Pharmaceutical sleep aids; Women

RESUMO

Objetivo:

Explorar a relação entre diferentes padrões de multimorbidade e o uso de medicamentos para dormir em mulheres.

Métodos:

Estudo transversal de base populacional com 1.128 mulheres (20–69 anos) no Sul do Brasil. Dados sobre medicamentos para dormir foram obtidos por meio da pergunta "Você toma/usa algum medicamento para conseguir dormir?" e identificados pela Classificação Anatômica Terapêutica e Química. Os padrões de multimorbidade foram derivados pela Análise de Componentes Principais de 26 condições crônicas e dois parâmetros de obesidade (≥30 kg/m²; ≥40 kg/m²). A associação foi analisada por regressão de Poisson com variância robusta utilizando diferentes modelos de ajuste, estratificados por idade.

Resultados:

Foram derivados três padrões de multimorbidade: cardiometabólico, endócrino-articular e psicossomático. A estratificação etária mostrou mudança de efeito na relação investigada. Mulheres com idade inferior a 45 anos e alto escore nos padrões cardiometabólico e endócrino-articular tiveram cerca de duas vezes mais probabilidade de usar medicamentos para dormir [razão de prevalência (RP) 1,85, intervalo de confiança (IC) 95% 1,09–3,12; RP 2,04, IC95% 1,18–3,51, respectivamente]. Aquelas com padrão psicossomático tiveram cerca de cinco vezes mais probabilidade (RP 4,91, IC95% 3,00–8,04).

Conclusões:

O estudo forneceu as primeiras evidências sobre a associação investigada e demonstrou que mulheres jovens (<45 anos) com alto escore dos padrões identificados têm até cinco vezes mais probabilidade de usar medicamentos para dormir, configurando uso precoce. Essa constatação inédita sugere a necessidade de maior promoção da saúde dos jovens e de ações de conscientização sobre os riscos e da indicação clara do uso de medicamentos para dormir.

Palavras-chave:
Multimorbidade; Múltiplas afecções crônicas; Medicamentos para dormir; Mulheres

INTRODUCTION

The use of pharmacotherapy for sleep disorders has increased considerably in recent years. Sleep-aiding psychotropic drugs have been widely prescribed, and the duration of treatment often exceeds the recommended period of up to three months¹1 Weymann D, Gladstone EJ, Smolina K, Morgan SG. Long-term sedative use among community-dwelling adults: a population-based analysis. CMAJ Open 2017; 5: E52–E60. https://doi.org/10.9778/cmajo.20160056
https://doi.org/10.9778/cmajo.20160056... . Epidemiological studies seeking to elucidate the factors associated with increased prescription and use of sleep aids are still inconclusive²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... . Sleep disorders are a frequent complaint in public health but predisposing conditions are not well established³3 Fegadolli C, Varela NMD, Carlini ELA. Uso e abuso de benzodiazepínicos na atenção primária à saúde: práticas profissionais no Brasil e em Cuba. Cad Saúde Pública 2019; 35(6): e00097718. https://doi.org/10.1590/0102-311X00097718
https://doi.org/10.1590/0102-311X0009771... .

Additionally, sleeping disorders are more prevalent among women and individuals with chronic diseases⁴4 Helbig AK, Stöckl D, Heier M, Thorand B, Schulz H, Peters A, et al. Relationship between sleep disturbances and multimorbidity among community-dwelling men and women aged 65–93 years: results from the KORA Age Study. Sleep Med 2017; 33: 151-9. https://doi.org/10.1016/j.sleep.2017.01.016
https://doi.org/10.1016/j.sleep.2017.01.... . Women, due to their own reproductive and hormonal characteristics, experience sleep differently through⁵5 Pengo MF, Won CH, Bourjeily G. Sleep in women across the life span. Chest 2018; 154(1): 196-206. https://doi.org/10.1016/j.chest.2018.04.005
https://doi.org/10.1016/j.chest.2018.04.... the use of sleep-aided medications, with a prevalence range observed from 10–28%²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... ,⁶6 Abolhassani N, Haba-Rubio J, Heinzer R, Vollenweider P, Marques-Vidal P. Ten-year trend in sleeping pills use in Switzerland: the CoLaus study. Sleep Med 2019; 64: 56-61. https://doi.org/10.1016/j.sleep.2018.06.022
https://doi.org/10.1016/j.sleep.2018.06.... . The findings reported a linear increase with age, but more significant in the transition from reproductive to non-reproductive life due to the poorer quality of sleep during this period⁵5 Pengo MF, Won CH, Bourjeily G. Sleep in women across the life span. Chest 2018; 154(1): 196-206. https://doi.org/10.1016/j.chest.2018.04.005
https://doi.org/10.1016/j.chest.2018.04.... ,⁷7 Carrasco-Garrido P, Hernández-Barrera V, Jiménez-Trujillo I, Esteban-Hernández J, Álvaro-Meca A, López-De Andrés A, et al. Time trend in psychotropic medication use in Spain: a nationwide population-based study. Int J Environ Res Public Health 2016; 13(12): 1177. https://doi.org/10.3390/ijerph13121177
https://doi.org/10.3390/ijerph13121177... . Moreover, the literature reveals a direct relationship between the number of chronic conditions and sleep problems⁸8 Koyanagi A, Garin N, Olaya B, Ayuso-Mateos JL, Chatterji S, Leonardi M, et al. Chronic conditions and sleep problems among adults aged 50 years or over in nine countries: a multi-country study. PLoS One 2014; 9(12): e114742. https://doi.org/10.1371/journal.pone.0114742
https://doi.org/10.1371/journal.pone.011... . It also indicates that specific conditions such as chronic pain and cardiometabolic disorders, alone or in combination, may be predictors of sleep disorders and potentially contributors to the use of sleep medications⁸8 Koyanagi A, Garin N, Olaya B, Ayuso-Mateos JL, Chatterji S, Leonardi M, et al. Chronic conditions and sleep problems among adults aged 50 years or over in nine countries: a multi-country study. PLoS One 2014; 9(12): e114742. https://doi.org/10.1371/journal.pone.0114742
https://doi.org/10.1371/journal.pone.011... –¹⁰10 Leigh L, Hudson IL, Byles JE. Sleep difficulty and disease in a cohort of very old women. J Aging Health 2016; 28(6): 1090-104. https://doi.org/10.1177/089826431562490
https://doi.org/10.1177/089826431562490... . Recent researches indicate that the presence of two or more chronic conditions in the same individual, defined as multimorbidity, may increase the probability of the occurrence of these disorders⁸8 Koyanagi A, Garin N, Olaya B, Ayuso-Mateos JL, Chatterji S, Leonardi M, et al. Chronic conditions and sleep problems among adults aged 50 years or over in nine countries: a multi-country study. PLoS One 2014; 9(12): e114742. https://doi.org/10.1371/journal.pone.0114742
https://doi.org/10.1371/journal.pone.011... ,¹⁰10 Leigh L, Hudson IL, Byles JE. Sleep difficulty and disease in a cohort of very old women. J Aging Health 2016; 28(6): 1090-104. https://doi.org/10.1177/089826431562490
https://doi.org/10.1177/089826431562490... ,¹¹11 Schmidt MG, Franken DL, Olinto MTA, Costa JSD, Paniz VMV. Impact of multimorbidity on the use of sleep medications among women: a population-based study in Southern Brazil. Sleep Breath 2023; 27(3): 1135-42. https://doi.org/10.1007/s11325-022-02672-5
https://doi.org/10.1007/s11325-022-02672... .

Studies have shown that individuals with multimorbidity are more likely to be prescribed hypnotic/anxiolytic medications⁹9 Linnet K, Gudmundsson LS, Birgisdottir FG, Sigurdsson EL, Johannsson M, Tomasdottir MO, et al. Multimorbidity and use of hypnotic and anxiolytic drugs: cross-sectional and follow-up study in primary healthcare in Iceland. BMC Fam Pract 2016; 17: 69. https://doi.org/10.1186/s12875-016-0469-0
https://doi.org/10.1186/s12875-016-0469-... ,¹²12 Ramos LR, Mari JJ, Fontanella AT, Dal Pizzol TS, Bertoldi AD, Mengue SS, et al. Nationwide use of psychotropic drugs for treatment of self-reported depression in the Brazilian urban adult population. Rev Bras Epidemiol 2020; 23: e200059. https://doi.org/10.1590/1980-549720200059
https://doi.org/10.1590/1980-54972020005... . Common mental disorders (CMD) represent one of the most frequent coexisting diagnoses in different combinations of chronic conditions⁶6 Abolhassani N, Haba-Rubio J, Heinzer R, Vollenweider P, Marques-Vidal P. Ten-year trend in sleeping pills use in Switzerland: the CoLaus study. Sleep Med 2019; 64: 56-61. https://doi.org/10.1016/j.sleep.2018.06.022
https://doi.org/10.1016/j.sleep.2018.06.... ; however, the use of sleeping medications is strongly associated with multimorbidity among women, even in those without medical reports of CMD²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... .

Furthermore, in most investigations, obesity is a risk factor for multimorbidity; however, other studies suggest that it should be considered as a component condition of some combinations of chronic conditions¹³13 Carvalho JN, Cancela MC, Souza DLB. Lifestyle factors and high body mass index are associated with different multimorbidity clusters in the Brazilian population. PLoS One 2018; 13(11): e0207649. https://doi.org/10.1371/journal.pone.0207649
https://doi.org/10.1371/journal.pone.020... ,¹⁴14 Jackson CA, Dobson AJ, Tooth LR, Mishra GD. Lifestyle and socioeconomic determinants of multimorbidity patterns among mid-aged women: a longitudinal study. PLoS One 2016; 11(6): e0156804. https://doi.org/10.1371/journal.pone.0156804
https://doi.org/10.1371/journal.pone.015... . A previous research conducted on women aged 20–69 years that investigated the effect of multimorbidity on sleeping medication use, considering obesity as an independent variable (risk factor or potential confounding factor) and a component of the multimorbidity outcome, found that in multimorbidity women (≥2; ≥3 chronic conditions), the probability of using these drugs was approximately double, even after adjustment for CMD¹¹11 Schmidt MG, Franken DL, Olinto MTA, Costa JSD, Paniz VMV. Impact of multimorbidity on the use of sleep medications among women: a population-based study in Southern Brazil. Sleep Breath 2023; 27(3): 1135-42. https://doi.org/10.1007/s11325-022-02672-5
https://doi.org/10.1007/s11325-022-02672... .

Thus, this study aimed to explore the relationship between different patterns of multimorbidity and the use of sleeping medications in adult women.

METHODS

Study design and target population

A population-based cross-sectional study was conducted in 2015 (from February to October) with a representative sample of women aged 20–69 years residents of the municipality of São Leopoldo, state of Rio Grande do Sul, Brazil. This study is part of a larger study that aimed to evaluate the living and health conditions of adult women. We included women in the age group of interest who lived in the selected sectors and households, except pregnant women.

Sampling and data collection

The sample size was calculated based on the outcomes of interest. We chose the one that required a larger sample size (1,013 women), added 10% for eventual losses and refusals, and 15% to control for confounding factors in the data analysis, totaling 1,281 women. The sampling was probabilistic and performed in multiple stages. Initially, we estimated the average number of women per household in the municipality of São Leopoldo and their proportion in the age group of interest. A total of 371 census tracts in the urban area were classified in descending order according to residents’ monthly income. The sectors were numbered from 1–371, of which 45 were randomly selected and subsequently totaled 36 households in each sector. Then, the blocks and corners of each conglomerate were drawn to start the research. The households were designated according to the following rule: the corner indicated the starting point, always to the left of those facing the initial corner, the first house was selected for the study, skipping two houses and again selected the fourth house, and so on, until 36 households per sector were completed. Trained interviewers collected the data through the application of a standardized questionnaire previously tested in a pilot study. Quality control for data was applied to 10% of the participants.

Variables analyzed

Data on the use of sleeping medications (dependent variable) were obtained through the question "Do you take/use any medication to get to sleep?". The name of the drugs, indication, and duration of use were also requested. Sample calculations were performed using the Epi Info 7.2.5.0 program (https://www.cdc.gov) with an estimated prevalence of 15%, margin of error of 2.5 percentage points, and design effect of 1.37, as detailed in a previous publication¹¹11 Schmidt MG, Franken DL, Olinto MTA, Costa JSD, Paniz VMV. Impact of multimorbidity on the use of sleep medications among women: a population-based study in Southern Brazil. Sleep Breath 2023; 27(3): 1135-42. https://doi.org/10.1007/s11325-022-02672-5
https://doi.org/10.1007/s11325-022-02672... .

To identify chronic conditions, we applied the report on the current consumption of medications of continuous use as a proxy¹⁵15 Huber CA, Szucs TD, Rapold R, Reich O. Identifying patients with chronic conditions using pharmacy data in Switzerland: an updated mapping approach to the classification of medications. BMC Public Health 2013; 13: 1030. https://doi.org/10.1186/1471-2458-13-1030
https://doi.org/10.1186/1471-2458-13-103... ,¹⁶16 Franken DL, Dias-Da-Costa JS, Olinto MTA, Sturmer J, Bordin RB, Paniz VMV. Multimorbidity patterns: obesity as the main modifiable risk factor in adult women in Southern Brazil. Arch Endocrinol Metab 2023; 67(5): e000642. https://doi.org/10.20945/2359-3997000000642
https://doi.org/10.20945/2359-3997000000... . It was determined according to the Anatomical Therapeutic and Chemical (ATC) classification: level 1, anatomical group N; and levels 2 and 3, therapeutic and chemical groups and the prescribed indications mentioned by the participants¹⁷17 WHO Collaborating Centre for Drug Statistics Methodology. Guidelines for ATC classification and DDD assignment 2023. Oslo: Norway; 2022.. Ultimately, 26 chronic conditions were identified, with the most prevalents being: hypertension, CMDs, acid-related digestive disorders, dyslipidemia, thyroid diseases, diabetes mellitus, circulatory disorders, and chronic pain¹⁸18 Franken DL, Olinto MTA, Dias-da-Costa JS, Bairros FS, Paniz VMV. Lifestyle, high Body Mass Index, and markers of socioeconomic conditions associated with multimorbidity in women. Rev Bras Epidemiol 2022; 25: e220007. https://doi.org/10.1590/1980-549720220007
https://doi.org/10.1590/1980-54972022000... .

Explanatory sociodemographic variables and possible confounding factors included: age (20–34, 35–44, 45–54, 55–69 years) analyzed as <45 and ≥45 years; self-reported skin color (white/other–black, yellow, parda, Indigenous, and others); marital status (having a partner or not); education in years of study (≥11, 8–10, 5–7, and ≤4 years); economic class (A+B, C, and D+E) according to the economic classification criteria of the Brazilian Association of Research Companies (ABEP, Associação Brasileira de Empresas de Pesquisa; https://www.abep.org/); household income per capita in quartiles (first ≤R$525.30, second R$525.31–869.00, third R$869.01–1547.00, and fourth ≥R$1547.01) corresponding to about two minimum wages, considering a national minimum wage of R$788.00, approximately U$245 at the time of the study; and occupation (employed/unemployed).

Behavioral variables were adequate consumption of fruits and vegetables (≥5 servings daily)¹⁹19 Stables GJ, Subar AF, Patterson BH, Dodd K, Heimendinger J, Van Duyn MAS, et al. Changes in vegetable and fruit consumption and awareness among US adults: results of the 1991 and 1997 5 A Day for Better Health Program surveys. J Am Diet Assoc 2002; 10296): 809-17. https://doi.org/10.1016/s0002-8223(02)90181-1
https://doi.org/10.1016/s0002-8223(02)90... ; alcohol consumption established based on frequency, type of beverage, and daily amount ingested, considering 15 g of ethanol/day as a cutoff point²⁰20 Barroso WKS, Rodrigues CIS, Bortolotto LA, Mota-Gomes MA, Brandão AA, Feitosa ADM, et al. Diretrizes Brasileiras de Hipertensão Arterial – 2020. Arq Bras Cardiol 2021; 116(3): 516-658. https://doi.org/10.36660/abc.20201238
https://doi.org/10.36660/abc.20201238... ; current smoker; and leisure-time physical activity, according to the International Physical Activity Questionnaire-Short Form (IPAQ-SF) categorized as active (≥150 min/week of moderate/vigorous physical activity) or inactive (<150 min/week)²¹21 Matsudo S, Araújo T, Matsudo V, Andrade D, Andrade E, Oliveira LC, et al. Questionário Internacional De Atividade Física (IPAQ): estudo de validade e reprodutibilidade no Brasil. Rev Bras Ativ Fís Saúde 2012; 6(2): 5-18. https://doi.org/10.12820/rbafs.v.6n2p5-18
https://doi.org/10.12820/rbafs.v.6n2p5-1... . Health variables were body mass index (BMI), classified according to the criteria of the World Health Organization: (low weight <18.5 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... , eutrophy 18.5–24.9 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... , overweight 25.0–29.9 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... , obesity ≥30.0 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... , being Class I obesity 30.0–34.9 kg/m², Class II obesity 35.0–39.9 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... , and Class III obesity ≥40 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... )²²22 Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser 2000; 894: i-xii, 1-53. PMID: 11234459.; and CMD (absent <7 points and present (≥7 points), according to the score obtained through the Self-Reporting Questionnaire (SQR-20)²³23 Gonçalves DM, Stein AT, Kapczinski F. Avaliação de desempenho do Self-Reporting Questionnaire como instrumento de rastreamento psiquiátrico: Um estudo comparativo com o Structured Clinical Interview for DSM-IV-TR. Cad Saúde Pública 2008; 24(2): 380-90. https://doi.org/10.1590/S0102-311X2008000200017
https://doi.org/10.1590/S0102-311X200800... .

Data analysis

Multimorbidity patterns were derived from two approaches using Principal Component Analysis (PCA). In the first, ten chronic conditions were considered (prevalence ≥2%)¹⁸18 Franken DL, Olinto MTA, Dias-da-Costa JS, Bairros FS, Paniz VMV. Lifestyle, high Body Mass Index, and markers of socioeconomic conditions associated with multimorbidity in women. Rev Bras Epidemiol 2022; 25: e220007. https://doi.org/10.1590/1980-549720220007
https://doi.org/10.1590/1980-54972022000... . In the second, obesity was included (BMI≥30 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... ) as well as morbid obesity (BMI≥40 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... )¹¹11 Schmidt MG, Franken DL, Olinto MTA, Costa JSD, Paniz VMV. Impact of multimorbidity on the use of sleep medications among women: a population-based study in Southern Brazil. Sleep Breath 2023; 27(3): 1135-42. https://doi.org/10.1007/s11325-022-02672-5
https://doi.org/10.1007/s11325-022-02672... . The factors obtained were rotated through orthogonal rotation (Varimax), which minimizes the number of variables with high factor loads in each factor, increasing the accuracy of the analysis, and ensuring non-correlation between the factors²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22.. Because each chronic condition was coded as a dichotomous variable, the tetrachoric correlation matrix between all conditions was calculated¹⁴14 Jackson CA, Dobson AJ, Tooth LR, Mishra GD. Lifestyle and socioeconomic determinants of multimorbidity patterns among mid-aged women: a longitudinal study. PLoS One 2016; 11(6): e0156804. https://doi.org/10.1371/journal.pone.0156804
https://doi.org/10.1371/journal.pone.015... . The number of retained factors was based on components with eigenvalues >1 and a scree-plot test. Chronic conditions were considered loaded on one factor if they had an absolute correlation ≥0.3 with the factor²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22.. Before defining the number of chronic conditions included in the PCA, Kaiser-Meyer-Olkin's (KMO) and Bartlett's sphericity tests were used to verify the applicability of the analysis. Factor scores were saved for each participant individually. Multimorbidity patterns were divided into tertiles based on their scores, categorized into low (first and second tertiles) and high (third tertile)¹⁴14 Jackson CA, Dobson AJ, Tooth LR, Mishra GD. Lifestyle and socioeconomic determinants of multimorbidity patterns among mid-aged women: a longitudinal study. PLoS One 2016; 11(6): e0156804. https://doi.org/10.1371/journal.pone.0156804
https://doi.org/10.1371/journal.pone.015... .

The univariate analysis described the sample, the bivariate analysis used Pearson's chi-square and linear trend tests, and the association between multimorbidity patterns and sleeping medication use was assessed through Poisson regression with robust variance. The crude and adjusted prevalence ratios (PRs) obtained were described with their respective 95% confidence intervals (95%CI). Variables with p≤0.20 in the crude analysis were considered potential confounds. The possible interaction between participants’ age and the main exposure was investigated using the Mantel-Haenszel¹1 Weymann D, Gladstone EJ, Smolina K, Morgan SG. Long-term sedative use among community-dwelling adults: a population-based analysis. CMAJ Open 2017; 5: E52–E60. https://doi.org/10.9778/cmajo.20160056
https://doi.org/10.9778/cmajo.20160056... ,⁷7 Carrasco-Garrido P, Hernández-Barrera V, Jiménez-Trujillo I, Esteban-Hernández J, Álvaro-Meca A, López-De Andrés A, et al. Time trend in psychotropic medication use in Spain: a nationwide population-based study. Int J Environ Res Public Health 2016; 13(12): 1177. https://doi.org/10.3390/ijerph13121177
https://doi.org/10.3390/ijerph13121177... . A p<0.05 in the homogeneity test (M-H) defined the stratification of the analyses by age (<45 and ≥45 years). Multivariate analysis was performed according to different adjustment models: Model I (unadjusted PR), Model II (Model I + sociodemographic characteristics), Model III (Model II + behavioral variables), and Model IV (Model III + health variables), and significance level p<0.05. The diagnosis of each model for each level of analysis (II, III, and IV) was performed using the poisgof command. The statistically significant value (p<0.05) of the test indicates that the model is inappropriate. The detected values ranged from 0.6107 to 0.9997. In sensitivity analyses, we calculated crude and adjusted PRs for sleeping medications use according to the multimorbidity patterns generated by both approaches and additionally without stratification by age group. Statistical analyses were performed using Stata 13.0 (Stata Corp., College Station, Texas, USA).

The study was conducted following the Declaration of Helsinki guidelines and approved by the Research Ethics Committee of the University of Vale do Rio dos Sinos (CAAE 30872914.6.0000.5344, protocol 650.443). All participants provided written informed consent.

RESULTS

The total number of women visited was 1,281, of which 1,128 were interviewed, representing losses or refusals of 11.9%. The participants were 43.4 years on average, standard deviation (SD) ±13.4 years. Most reported working (56.0%), 18.4% were smokers, and 66.8% consumed alcoholic beverages. Overweight/obesity were identified in two-thirds of the sample (66.0%), and CMD in approximately 40.0%. Sleeping medications were used by 14.3% (95%CI 12.2–16.3) of participants, and showed a direct linear association with age, inverse association with schooling, and were more prevalent among women who did not work, who did not consume alcoholic beverages, were former smokers, were overweight/obese, and had CMD (Table 1).

Three patterns of multimorbidity were derived by PCA that explained 45.5% of the total variance: cardiometabolic (dyslipidemia, circulatory disorders, arterial hypertension, and diabetes), endocrine-articular (thyroid diseases, osteoporosis/osteopenia, and rheumatic diseases), and psychosomatic (chronic pain, CMD, and acid-related digestive disorders). The KMO coefficient was 0.720 with p≤0.001 for the Bartlett's test, suggesting the adequacy of the analysis²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22.. Among the ten chronic conditions included, only acid-related digestive disorders were saturated in >1 pattern. Cardiometabolic pattern had the highest percentage (21.9%) of explained variance (Table 2). In PCA, which included obesity, a fourth pattern was derived, explaining 9.4% of the total variance (Supplementary Table 1). The KMO coefficient was 0.711 with p≤0.001 for Bartlett's test.

Table 3 shows that a high score of the identified multimorbidity patterns was verified in older women, with less schooling, unemployed, with adequate consumption of fruits and vegetables, who did not consume alcoholic beverages, were former smokers, and obese. High scores for cardiometabolic and psychosomatic patterns were identified in women with CMD and those on sleeping medications.

In Table 4, after adjusting for confounding factors, there was an increased probability of sleeping medication use in the presence of a high score of multimorbidity patterns identified among younger women (<45 years). A high cardiometabolic and endocrine-articular multimorbidity pattern score doubled the probability of sleeping medication use among women in this age group (PR 1.85, 95%CI 1.09–3.12; PR 2.04, 95%CI 1.18–3.51, respectively). The high score of the psychosomatic multimorbidity pattern increased the probability of sleeping medication consumption by about five times in women aged <45 years and twice among women aged ≥45 years (PR 4.91, 95%CI 3.00–8.04; PR 1.96, 95%CI 1.30–2.94, respectively).

All sensitivity analyses showed results similar to our main results and provided evidence of the detected interaction (Supplementary Tables 2-5).

DISCUSSION

This study investigated the relationship between patterns of multimorbidity and sleeping medication use in a representative sample of women aged 20–69 years. The consumption of sleeping medications was verified in approximately 15% of women, and three patterns of multimorbidity were identified: cardiometabolic, endocrine-articular, and psychosomatic. The association between high cardiometabolic and endocrine-articular multimorbidity pattern scores and outcome was modified by age. Younger women (<45 years) were twice as likely to use sleeping medications but in the next age group (≥45 years), this association was not verified. For the high psychosomatic pattern score, the probability of using sleeping medications doubled in this age group and was five times greater in younger women, even after adjusting for CMD.

The prevalence of sleeping medication use was higher than those identified in studies carried out in Europe⁶6 Abolhassani N, Haba-Rubio J, Heinzer R, Vollenweider P, Marques-Vidal P. Ten-year trend in sleeping pills use in Switzerland: the CoLaus study. Sleep Med 2019; 64: 56-61. https://doi.org/10.1016/j.sleep.2018.06.022
https://doi.org/10.1016/j.sleep.2018.06.... and lower than that observed in women aged ≥25 years in Spain (18.1%)⁷7 Carrasco-Garrido P, Hernández-Barrera V, Jiménez-Trujillo I, Esteban-Hernández J, Álvaro-Meca A, López-De Andrés A, et al. Time trend in psychotropic medication use in Spain: a nationwide population-based study. Int J Environ Res Public Health 2016; 13(12): 1177. https://doi.org/10.3390/ijerph13121177
https://doi.org/10.3390/ijerph13121177... . However, in a study conducted in Canada with women ≥18 years, a prevalence of 14.6% was found¹1 Weymann D, Gladstone EJ, Smolina K, Morgan SG. Long-term sedative use among community-dwelling adults: a population-based analysis. CMAJ Open 2017; 5: E52–E60. https://doi.org/10.9778/cmajo.20160056
https://doi.org/10.9778/cmajo.20160056... . In Brazil, a national survey observed the prevalence of 10.4% in this age group²⁵25 Kodaira K, Silva MT. Sleeping pill use in Brazil: a population-based, cross-sectional study. BMJ Open 2017; 7(7): e016233. https://doi.org/10.1136/bmjopen-2017-016233
https://doi.org/10.1136/bmjopen-2017-016... . However, the frequency of sleeping medication use has increased in recent decades among women. A cohort study in Australia with a 20–year segment observed that consumption quadrupled over this period²⁶26 Stuart AL, Mohebbi M, Pasco JA, Quirk SE, Brennan-Olsen SL, Berk M, et al. Pattern of psychotropic medication use over two decades in Australian women. Aust N Z J Psychiatry 2017; 51(12): 1212-9. https://doi.org/10.1177/0004867417704056
https://doi.org/10.1177/0004867417704056... . Furthermore, there is a linear increase with age, especially in the age group ≥45 years¹1 Weymann D, Gladstone EJ, Smolina K, Morgan SG. Long-term sedative use among community-dwelling adults: a population-based analysis. CMAJ Open 2017; 5: E52–E60. https://doi.org/10.9778/cmajo.20160056
https://doi.org/10.9778/cmajo.20160056... ,⁷7 Carrasco-Garrido P, Hernández-Barrera V, Jiménez-Trujillo I, Esteban-Hernández J, Álvaro-Meca A, López-De Andrés A, et al. Time trend in psychotropic medication use in Spain: a nationwide population-based study. Int J Environ Res Public Health 2016; 13(12): 1177. https://doi.org/10.3390/ijerph13121177
https://doi.org/10.3390/ijerph13121177... , which is consistent with this study which found that sleeping medication use doubled among women in this age group.

The patterns of cardiometabolic, endocrine-articular, and psychosomatic multimorbidity identified corroborated the literature¹³13 Carvalho JN, Cancela MC, Souza DLB. Lifestyle factors and high body mass index are associated with different multimorbidity clusters in the Brazilian population. PLoS One 2018; 13(11): e0207649. https://doi.org/10.1371/journal.pone.0207649
https://doi.org/10.1371/journal.pone.020... ,¹⁴14 Jackson CA, Dobson AJ, Tooth LR, Mishra GD. Lifestyle and socioeconomic determinants of multimorbidity patterns among mid-aged women: a longitudinal study. PLoS One 2016; 11(6): e0156804. https://doi.org/10.1371/journal.pone.0156804
https://doi.org/10.1371/journal.pone.015... ,²⁷27 Simões D, Araújo FA, Severo M, Monjardino T, Cruz I, Carmona L, et al. Patterns and consequences of multimorbidity in the general population: there is no chronic disease management without rheumatic disease management. Arthritis Care Res (Hoboken) 2017; 69(1): 12-20. https://doi.org/10.1002/acr.22996
https://doi.org/10.1002/acr.22996... . However, methodological differences in pattern derivation, target population, and amplitude of the age groups investigated may explain some specificities. In this sense, the chronic conditions that derived the endocrine-articular and psychosomatic patterns in the present study also composed a unique pattern in a similar study²⁷27 Simões D, Araújo FA, Severo M, Monjardino T, Cruz I, Carmona L, et al. Patterns and consequences of multimorbidity in the general population: there is no chronic disease management without rheumatic disease management. Arthritis Care Res (Hoboken) 2017; 69(1): 12-20. https://doi.org/10.1002/acr.22996
https://doi.org/10.1002/acr.22996... . Another aspect pertinent to the composition of the patterns involves the complex underlying mechanisms of multimorbidity and health determinants, such as psychosocial, behavioral, socioeconomic, and population-level intervention factors that reflect on the structure of health in different countries.

Considering the explanatory variables examined, the highest prevalence of patterns of multimorbidity and sleeping medication use was verified in older women, those with low schooling, unemployed, not consuming alcoholic beverages, with CMD, and obese, converging with the literature⁶6 Abolhassani N, Haba-Rubio J, Heinzer R, Vollenweider P, Marques-Vidal P. Ten-year trend in sleeping pills use in Switzerland: the CoLaus study. Sleep Med 2019; 64: 56-61. https://doi.org/10.1016/j.sleep.2018.06.022
https://doi.org/10.1016/j.sleep.2018.06.... ,¹³13 Carvalho JN, Cancela MC, Souza DLB. Lifestyle factors and high body mass index are associated with different multimorbidity clusters in the Brazilian population. PLoS One 2018; 13(11): e0207649. https://doi.org/10.1371/journal.pone.0207649
https://doi.org/10.1371/journal.pone.020... ,²⁵25 Kodaira K, Silva MT. Sleeping pill use in Brazil: a population-based, cross-sectional study. BMJ Open 2017; 7(7): e016233. https://doi.org/10.1136/bmjopen-2017-016233
https://doi.org/10.1136/bmjopen-2017-016... . However, inconsistency has been identified in the association with smoking. Studies point to smoking as a predictor of sleeping medication use due to the stimulating effect of nicotine⁷7 Carrasco-Garrido P, Hernández-Barrera V, Jiménez-Trujillo I, Esteban-Hernández J, Álvaro-Meca A, López-De Andrés A, et al. Time trend in psychotropic medication use in Spain: a nationwide population-based study. Int J Environ Res Public Health 2016; 13(12): 1177. https://doi.org/10.3390/ijerph13121177
https://doi.org/10.3390/ijerph13121177... ,²⁵25 Kodaira K, Silva MT. Sleeping pill use in Brazil: a population-based, cross-sectional study. BMJ Open 2017; 7(7): e016233. https://doi.org/10.1136/bmjopen-2017-016233
https://doi.org/10.1136/bmjopen-2017-016... . In our study, women who quit smoking were more likely to consume sleeping medications. Thus, it is considered relevant to investigate the duration of nicotine abstinence and the need for sleeping medications in future studies.

In this study, the relationship assessed was modified by the participants’ age. Although sleeping medication use has shown a direct linear association with age, the probability of use was higher among younger women with high scores for the identified patterns. Some authors suggest that chronic pain present in certain patterns, cardiovascular diseases, or dysregulated hormone levels negatively impact sleep and may affect the sleep-wake cycle⁸8 Koyanagi A, Garin N, Olaya B, Ayuso-Mateos JL, Chatterji S, Leonardi M, et al. Chronic conditions and sleep problems among adults aged 50 years or over in nine countries: a multi-country study. PLoS One 2014; 9(12): e114742. https://doi.org/10.1371/journal.pone.0114742
https://doi.org/10.1371/journal.pone.011... ,¹⁰10 Leigh L, Hudson IL, Byles JE. Sleep difficulty and disease in a cohort of very old women. J Aging Health 2016; 28(6): 1090-104. https://doi.org/10.1177/089826431562490
https://doi.org/10.1177/089826431562490... . Still, in premenopausal women thyroid disorders are common²⁸28 Vanderpump MPJ. The epidemiology of thyroid disease. Br Med Bull 2011; 99: 39-51. https://doi.org/10.1093/bmb/ldr030
https://doi.org/10.1093/bmb/ldr030... and the symptoms include inability to rest, agitation, and anxiety that lead to difficulty sleeping. It is plausible to think that the presence of this condition in the endocrine-articular pattern could lead to the consumption of these medications in younger women.

Our findings evidenced the high score of the psychosomatic pattern as an important predictor of sleeping medication use, especially in younger women. Evidence points toward an emerging increase in the prevalence of CMD among younger women, which could explain, at least in part, our results¹²12 Ramos LR, Mari JJ, Fontanella AT, Dal Pizzol TS, Bertoldi AD, Mengue SS, et al. Nationwide use of psychotropic drugs for treatment of self-reported depression in the Brazilian urban adult population. Rev Bras Epidemiol 2020; 23: e200059. https://doi.org/10.1590/1980-549720200059
https://doi.org/10.1590/1980-54972020005... . Chronic pain and acid-related digestive disorders also compose this pattern. The literature indicates a bidirectional relationship, permeated by biological and psychological factors, between CMD and chronic pain, which could interfere with sleep quality, leading to the consumption of sleeping medications²⁹29 Goesling J, Lin LA, Clauw DJ. Psychiatry and pain management: at the intersection of chronic pain and mental health. Curr Psychiatry Rep 2018; 20(2): 12. https://doi.org/10.1007/s11920-018-0872-4
https://doi.org/10.1007/s11920-018-0872-... . It is important to note that the prescription of sleeping medications without elucidation of the predisposing conditions that permeate sleep disorders in multimorbid and polymedicated women exposes this population to harmful effects that increase mortality and burden health services. Finally, the presence of acid-related digestive disorders in this pattern of multimorbidity is hypothesized to be due to pharmacotherapy associated with chronic pain, including non-steroidal anti-inflammatory drugs related to injury to the stomach epithelium by inhibiting the protective factors of the gastric mucosa³⁰30 Stafford G, Villén N, Roso-Llorach A, Troncoso-Mariño A, Monteagudo M, Violán C. Combined multimorbidity and polypharmacy patterns in the elderly: a cross-sectional study in primary health care. Int J Environ Res Public Health 2021; 18(17): 9216. https://doi.org/10.3390/ijerph18179216
https://doi.org/10.3390/ijerph18179216... .

Exploration of the role of obesity, either as a risk factor or chronic condition, in the construction of multimorbidity patterns is still controversial in the literature. In this sense, a previous study in this population observed twice the likelihood of using sleeping medications in case of multimorbidity (≥2; ≥3) including obesity as a chronic condition and risk factor; however, the patterns of multimorbidity were not investigated¹¹11 Schmidt MG, Franken DL, Olinto MTA, Costa JSD, Paniz VMV. Impact of multimorbidity on the use of sleep medications among women: a population-based study in Southern Brazil. Sleep Breath 2023; 27(3): 1135-42. https://doi.org/10.1007/s11325-022-02672-5
https://doi.org/10.1007/s11325-022-02672... . In the present study, when investigating obesity for the construction of patterns, the formation of a single pattern was observed regardless of the cutoff point used (≥30 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... ; ≥40 kg/m²2 Fernandes CSE, Azevedo RCS, Goldbaum M, Barros MBA. Psychotropic use patterns: are there differences between men and women? PLoS One 2018; 13(11): e0207921. https://doi.org/10.1371/journal.pone.0207921
https://doi.org/10.1371/journal.pone.020... ). Thus, considering that obesity alters metabolism and generates an inflammatory process that contributes to the development of chronic health conditions³¹31 Carvalho FG, Cunha AMD, Tonon AC, Pereira FS, Matte U, Callegari-Jacques SM, et al. Poor sleep quality associates with self-reported psychiatric and cardiometabolic symptoms independently of sleep timing patterns in a large sample of rural and urban workers. J Sleep Res 2020; 29(5): e12969. https://doi.org/10.1111/jsr.12969
https://doi.org/10.1111/jsr.12969... , it is plausible to think about the possible precursor role of obesity in multimorbidity patterns generation, i.e., obesity would indirectly influence the use of these medications.

Study limitations include the possibility of temporality bias inherent in cross-sectional studies and analyses based on self-reported data. However, the study considered the diagnoses of chronic conditions already under pharmacological treatment prescribed by a physician, validating the presence of these conditions. Regarding PCA, subjective decisions such as the definition of diseases¹⁸18 Franken DL, Olinto MTA, Dias-da-Costa JS, Bairros FS, Paniz VMV. Lifestyle, high Body Mass Index, and markers of socioeconomic conditions associated with multimorbidity in women. Rev Bras Epidemiol 2022; 25: e220007. https://doi.org/10.1590/1980-549720220007
https://doi.org/10.1590/1980-54972022000... , cutoff point for factor loading²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22., and number of factors to be retained²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22. were consistent with the literature¹⁴14 Jackson CA, Dobson AJ, Tooth LR, Mishra GD. Lifestyle and socioeconomic determinants of multimorbidity patterns among mid-aged women: a longitudinal study. PLoS One 2016; 11(6): e0156804. https://doi.org/10.1371/journal.pone.0156804
https://doi.org/10.1371/journal.pone.015... ,²⁴24 Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, eds. Epidemiologia nutricional. Rio de Janeiro: Atheneu; 2007. p. 213-22..

To the best of our knowledge, this is the first study to investigate the relationship between multimorbidity patterns and sleeping medication use in a representative sample of women. Besides, it promoted a new contribution to the area. The strategy employed to identify chronic health conditions allowed the inclusion of an unlimited number of morbidities already established by drug use, rather than a predefined list, as observed in most available studies. Furthermore, the statistical method applied for the derivation of multimorbidity patterns does not establish a priori restrictions on the derivation of factors or the number of factors to be retained; it allows each variable to load more than one factor. The analysis of morbid obesity allowed the exploration of its effect on sleeping medication use, an aspect not previously investigated.

Although both sleeping medication use and multimorbidity patterns were more frequent in older women, our study demonstrated a higher probability of consumption among younger women (<45 years) with high scores for the identified multimorbidity patterns, suggesting that specific aspects of some combinations of chronic conditions can interfere with the quality of sleep. Given the increasing prevalence of multimorbidity and the consumption of sleeping medications at early ages, understanding the relationship between different patterns of multimorbidity and the use of these medications becomes important to guide the clinical management and rational indication of sleep aids, and to design strategies for the prevention and care of women with multimorbidity, especially younger women.

DATA AVAILABILITY:

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

ACKNOWLEDGEMENTS:

MTAO received research productivity grants from the Brazilian Council for Scientific and Technological Development (CNPq, Conselho Nacional de Desenvolvimento Científico e Tecnológico) processes n. 307257/2013-4 and 307175/2017-0. JSDC received research productivity grants from CNPq process n. 310595/2018–0.

REFERENCES

Publication Dates

History