Keywords
diseases - multimorbidity - sleep - sleep deprivation
Introduction
Multimorbidity has been defined as the presence of two or more health issues;[1] however, there is no clear consensus regarding its concept and evaluation methods.
A review[2] on the prevalence of multimorbidity that included 68 articles found high heterogeneity,
especially in the number of diseases investigated, which ranged from 4 to 40 among
the studies. Moreover, a recent review[3] found a variety of measures of multimorbidity, including two main groups, one consisting
of a simple count of diseases from a list and another with a weighting for the chronic
conditions included. With the high heterogeneity, it becomes challenging to standardize
a way of measuring multimorbidity. The prevalence of multimorbidity may vary from
29.7% in low- and middle-income countries to 37.9% in high-income countries.[2] Other variations can be observed according to the number of diseases evaluated,
measurement methods (subjective or objective), and cutoff point.[2] In Brazil, ∼ 25% of the population is diagnosed with multimorbidity,[4]
[5] with a higher prevalence in older adults and women.[4]
[5]
[6] Recent evidence[7] suggests an increase in multimorbidity prevalence in Brazilian adults, from 18.7%
in 2013 to 22.3% in 2019. Data from high-income countries also showed that multimorbidity
has increased in the last few years.[8]
[9] Low quality of life,[10] mental health problems,[10] and early death[11] may be listed as consequences of multimorbidity. Moreover, the coronavirus disease
2019 (COVID-19) pandemic impacted the routine of health systems, reducing access to
healthcare among people with chronic diseases.[12]
Population aging may explain in part the occurrence of multimorbidity;[9] however, several other factors might be related to new cases,[9] such as the quality and duration of sleep. People who sleep around seven to eight
hours regularly are more likely to live one to three years more without chronic diseases
when compared with those who sleep less than seven or more than nine hours a day.[13]
Longitudinal studies[14] have evidenced that short sleep duration increases the risk of chronic conditions,
such as stroke and cancer. Moreover, data from 27 cohort samples,[15] including more than 1.3 million participants, evidenced that short sleep duration
increased the risk of all-cause mortality by 12%, whereas long sleep duration increased
it by 30%. A recent meta-analysis[16] showed that people who sleep less than seven hours a day had an increased risk of
all-cause mortality. Furthermore, in comparison to adequate sleep duration (∼ seven
hours a day), an increase or decrease in an hour of sleep increases the risk of cardiovascular
disease.[16]
A study[17] with data from Brazilian Basic Health Care Units patients associated sleep disorders
with chronic diseases such as osteoporosis, arthritis/osteoarthritis, low back pain,
depression, and obesity. Another study[18] evidenced the association of short sleep duration with comorbidities and unhappiness.
Given the growing evidence linking short and prolonged sleep duration to disease accumulation,
carrying out a study with a representative sample of a Brazilian state maybe be helpful
to provide evidence to health professionals and managers regarding case management,
health promotion, and prevention of chronic diseases. In addition, the COVID-19 pandemic
might have changed the time and quality of sleep,[19]
[20] which can impact the future prevalence of chronic diseases. Thus, we aimed to analyze
the association of sleep duration and use of sleeping medication with multimorbidity.
Material and Methods
We conducted a cross-sectional study using data from wave 3 of the Prospective Study
about Mental and Physical Health (PAMPA) cohort, a longitudinal study, with adults
aged 18 or older, performed in the state of Rio Grande do Sul, Brazil. The data 3
was collected between June and September 2021. The Research Ethics Board of the School
of Physical Education of Universidade Federal de Pelotas, Brazil, approved the study
(under protocol number 4.093.170)
Recruitment Phase
We contacted the participants through university professors, social media, local media,
and personal contacts from all macro-regions in the state of Rio Grande do Sul during
the recruitment phase.
Sample Size
The sample size was calculated based on the three primary outcomes of the cohort:
low back pain, depressive and anxiety symptoms, and access to the healthcare system.[21] According to the last Brazilian Census (2010), the total population of the state
of Rio Grande do Sul was of 10,693,929 inhabitants in 2010. A required sample of 1,767
participants was defined considering a 95% confidence level, a margin of error of
1.8, and a possible loss-to-follow-up rate of 30%. The state was divided into seven
macro-regions (with the names in Portuguese): Serra, Norte, Nordeste, Centro-Oeste,
Vales, Metropolitana, and Sul. The required sample size was divided proportionally
to the number of people living in each region.
Outcomes
The primary outcome was multimorbidity, defined as the presence of two or more and
three or more health issues. We assessed multimorbidity using the same question previously
used by the Brazilian Telephone-based Surveillance System for Noncommunicable Diseases:[22] “Has any doctor ever told you that you have the following health issues?”. Multimorbidity
was defined based on the following diseases: hypertension or high blood pressure,
diabetes, high levels of cholesterol, cancer, arthritis/arthrosis/fibromyalgia, asthma/bronchitis,
back problems, heart disease, depression, memory problems, HIV/AIDS, and other chronic
diseases.
Exposures
The primary exposure was the duration of sleep, assessed through the following question:
“Thinking about the last two weeks, how many hours per night did you sleep on average?”.
Based on this question, sleep duration was categorized as follows: 1) five hours or
less; 2) six hours; 3) seven to eight hours (adequate sleep); and 4) nine hours or
more. Sleeping medication was used as additional exposure, categorized through the
following yes-no question: “In the last two weeks, did you use any medication to sleep?”.
The following sociodemographic variables were considered as potential confounders:
sex (male or female); age in years (18–39, 40–59, and ≥ 60 years); skin color (white
or non-white); level of schooling (secondary or lower, higher education, and specialized,
masters, or Ph.D.). The behavioral variables included: practice of physical activity
in the previous week, categorized as “No” or “Yes,” as well as smoking, which was
classified as “smoker,” “former smoker,” or “never smoked.”
Data Analysis
Descriptive data were reported as proportions with a 95% confidence interval (95%CI).
Data were presented for the total sample and stratified according to multimorbidity
occurrence (two or more diseases and three or more diseases). For the association
analysis, seven to eight hours of sleep per night was considered the reference, whereas
the other categories were used as exposure. As for sleeping medications, their use
was considered exposure, whereas non-use was considered the reference. Crude and multivariable
analyses were performed using the logistic regression model and reported as odds ratio
(ORs) with 95%CIs. Regarding the multivariable analysis, two models were built: 1)
adjusting sleeping duration to sociodemographic variables (age, sex, skin color, and
level of schooling), and; 2) adjusting sleeping duration to model 1, behavioral variables
(physical activity and smoking), in addition to sleep duration and sleeping medication.
The analyses were performed using the Stata (StataCorp, College Station, TX, US) software,
version 15.1.
Results
[Table 1] shows the sociodemographic and behavioral characteristics of the general sample
and according to the occurrence of multimorbidity. A total of 2,936 subjects were
included (79.1% of them women). Most of the participants were aged between 18 and
39 years (54.2%; 95%CI: 52.5–55.9%) and were white (88.9%; 95%CI: 87.8–90.0%).
Table 1
Sociodemographic and behavioral characteristics of the study participants (N = 2,936).
|
Characteristics
|
Total sample
(N)
|
Total sample
(%): mean (range)
|
Proportion of participants with
MM2: mean (range)
|
Proportion of participants with
MM3: mean (range)
|
|
Sex
|
|
Male
|
678
|
20.9 (19.6–22.4)
|
38.0 (34.3–42.0)
|
20.0 (17.0–23.7)
|
|
Female
|
2,560
|
79.1 (77.6–80.4)
|
50.7 (48.6–52.7)
|
30.6 (28.8–32.5)
|
|
Age (years)
|
|
18–39
|
1,746
|
54.2 (52.5–55.9)
|
33.7 (31.5–36.1)
|
17.1 (15.4–19.1)
|
|
40–59
|
1,137
|
35.3 (33.7–37.0)
|
64.3 (61.3–67.2)
|
41.3 (38.3–44.3)
|
|
≥ 60
|
338
|
10.5 (9.5–11.6)
|
72.2 (66.7–77.1)
|
46.5 (40.7–52.3)
|
|
Skin color
|
|
White
|
2,866
|
88.9 (87.8–90.0)
|
48.0 (46.1–50.0)
|
28.1 (26.5–30.0)
|
|
Non-white
|
357
|
11.1 (10.0–12.2)
|
48.4 (42.9–54.0)
|
30.6 (25.8–36.0)
|
|
Level of Schooling
|
|
Secondary or lower
|
1,163
|
35.9
34.2–37.6
|
48.2
45.1–51.3
|
30.6 (27.8–33.5)
|
|
Higher education
|
780
|
24.1
22.6–25.6
|
47.9
44.3–51.6
|
27.4 (24.3–30.9)
|
|
Specialized, masters, PhD
|
1,298
|
40.0 (38.4–41.7)
|
48.0 (45.2–50.8)
|
27.2 (24.7–29.7)
|
|
Sleep duration (hours)
|
|
< 5
|
1,650
|
11.3 (10.1–12.5)
|
65.1 (59.8–70.1)
|
47.8 (42.4–53.3)
|
|
6
|
324
|
22.9 (21.4–24.5)
|
49.9 (46.1–53.7)
|
28.8 (25.5–32.4)
|
|
7–8
|
659
|
57.3 (55.5–59.1)
|
43.3 (40.9–45.7)
|
23.7 (21.7–25.8)
|
|
≥ 9
|
246
|
8.5 (7.6–9.6)
|
53.3 (47.0–59.4)
|
34.1 (28.5–40.3)
|
|
Sleeping medication
|
|
No
|
2,454
|
84.6 (83.3–85.9)
|
43.2 (41.2–45.1)
|
23.5 (21.9–25.2)
|
|
Yes
|
445
|
15.4 (14.1–16.7)
|
75.1 (70.8–78.9)
|
55.5 (50.8–60.1)
|
|
Physical activity (previous week)
|
|
No
|
1,455
|
49.6 (47.8–51.4)
|
55.1 (52.5–57.6)
|
33.3 (30.9–35.7)
|
|
Yes
|
1,479
|
50.4 (48.6–52.2)
|
41.1 (38.6–43.6)
|
23.4 (21.3–25.6)
|
Abbreviations: MM2, multimorbidity: two or more diseases; MM3, multimorbidity: three or more diseases.
Regarding sleep duration, 57.3% (95%CI: 55.5–59.1%) of the participants slept 7 to
8 hours. A linear positive association between multimorbidity and age was observed,
with older individuals having a higher prevalence of multimorbidity than younger subjects,
both for two or more and three or more diseases. Comparing the practice of physical
activity in the previous week, the prevalence of multimorbidity was lower among those
who reported practicing physical activity at both cutoff points for multimorbidity.
[Fig. 1] shows that the prevalence of multimorbidity was higher among those with short sleep
duration compared with the total sample, as well as for those with long sleep duration.
Regarding patients with adequate sleep time (seven to eight hours a day), the proportion
of multimorbidity was lower. The prevalence of multimorbidity was higher among participants
who reported the use of sleeping medication.
Fig. 1 Frequency of sleep duration and use of sleeping medication for the total sample and
among multimorbidity participants.
[Table 2] shows the association of sleep duration and sleeping medication use with multimorbidity,
considering multimorbidity as ≥ 3 diseases, ≤ 5 hours of sleep a day increased the
odds by 195% (OR: 2.95; 95%CI: 2.31–3.78). Nine or more hours also increased the multimorbidity
odds compared with the reference group (OR: 1.67; 95%CI: 1.25–2.22). The use of sleeping
medication increased the odds of multimorbidity by 306% (OR: 4.06; 95%CI: 3.29–5.00).
Prolonged sleep duration (nine hours or more per day) was associated with multimorbidity,
increasing its odds by 83% (OR: 1.49; 95%CI: 1.14–1.95). Compared with non-users of
sleeping medication, the users had a high probability of having multimorbidity (OR:
3.96; 95%CI: 3.15–4.98).
Table 2
Association of sleep duration and sleeping medication use with multimorbidity (N = 2,869).
|
Crude
OR (95%CI)
|
Model 1
OR (95%CI)
|
Model 2
OR (95%CI)
|
|
Two or more diseases
|
|
Daily sleep duration (hours)
|
|
7–8
|
Ref.
|
Ref.
|
Ref.
|
|
< 5
|
2.45 (1.90–3.14)
|
2.41 (1.88–3.10)
|
1.80 (1.36–2.38)
|
|
6
|
1.31 (1.09–1.57)
|
1.30 (1.08–1.56)
|
1.26 (1.03–1.54)
|
|
≥ 9
|
1.49 (1.14–1.95)
|
1.60 (1.20–2.13)
|
1.44 (1.07–1.94)
|
|
Sleeping medication
|
|
No
|
Ref
|
Ref
|
Ref
|
|
Yes
|
3.96 (3.15–4.98)
|
3.85 (3.06–4.85)
|
3.41 (2.67–4.36)
|
|
Three or more diseases
|
|
Daily sleep duration (hours)
|
|
7–8
|
Ref.
|
Ref.
|
Ref.
|
|
< 5
|
2.95 (2.31–3.78)
|
2.81 (2.19–3.60)
|
2.12 (1.61–2.79)
|
|
6
|
1.30 (1.06–1.60)
|
1.29 (1.05–1.58)
|
1.24 (0.99–1.55)
|
|
≥ 9
|
1.67 (1.25–2.22)
|
1.72 (1.27–2.33)
|
1.57 (1.14–2.15)
|
|
Sleeping medication
|
|
No
|
Ref.
|
Ref.
|
Ref.
|
|
Yes
|
4.06 (3.29–5.00)
|
3.93 (2.19–4.86)
|
3.46 (2.77–4.32)
|
Abbreviations: 95%CI, 95% confidence interval; N, number; OR, odds ratio.
Notes: Model 1: adjusted for age, sex, skin color, and level of schooling. Model 2:
adjusted for model 1 plus physical activity, smoking. When the sleep duration was
exposure, the model was adjusted for sleeping medication and vice versa.
Discussion
The present cross-sectional study evaluated the association between sleep duration
and sleeping medication with multimorbidity during the COVID-19 pandemic. Our results
showed that the prevalence of multimorbidity was lower among individuals with regular
sleep (seven to eight hours a day) compared with the total sample. Otherwise, the
prevalence was higher among those sleeping less than five hours a day and nine hours
or more. Short and prolonged sleep duration were associated with multimorbidity (for
two or more or there or more diseases), even after the adjustment for sociodemographic
and behavioral characteristics. The use of sleeping medication increased the probability
of having multimorbidity in the crude and adjusted analyses. This association may
occur due to drug interaction; therefore, more studies evaluating this association
are required. Despite this, a study[23] presented null results for chronic diseases (OR: 1.89; 95%CI: 0.56–6.35).
Cross-sectional analyses from the Canadian Longitudinal Study on Aging (CLSA) evidenced
a high probability of multimorbidity among participants with short or long sleep duration.[24] A longitudinal study[25] using data from the Swedish National Study of Aging and Care in Kungsholmen, showed
that moderate and severe sleep disturbances were associated with the accumulation
of chronic diseases, in addition to neuropsychiatric and musculoskeletal conditions.
However, for cardiovascular diseases, the results were not significant.[25] In older adults from the Cooperative Health Research in Augsburg (KORA), Germany,
neither short nor long daily sleep duration was associated with multimorbidity among
men. In contrast, among women, both short (≤ 5 hours a day) and prolonged (≥ 10 hours
a day) sleep duration were significantly associated with multimorbidity.[26] In Luxembourg, between 2013 and 2015, Ruiz-Castell et al.[23] presented an association between short sleep duration and the number of chronic
conditions, regardless of socioeconomic and behavioral characteristics. Short sleep
duration may be a risk factor for several chronic conditions, including cardiovascular
and metabolic diseases,[27] which represents the current findings of the present study. So far, the results
for some diseases, such as cardiovascular diseases, are inconclusive for men, and
more studies are needed. However, for the accumulation of conditions (multimorbidity),
the association is more solid.
Regarding the quality of sleep, data from the China Health and Retirement Longitudinal
Study[28] showed that participants with poor-quality sleep had a higher probability of having
multimorbidity compared with those with good-quality sleep. However, this parameter
was not evaluated in the present study, making the comparison with other studies unfeasible.
Several mechanisms may explain the association between short sleep duration and health
issues. Among the consequences of sleep disturbances, changes in circadian rhythms
stand out.[29] An unregulated circadian rhythm can exert deleterious effects on the human body.[30] The molecular circadian clock has functions in several cells in the human body and
exerts temporal control over the physiological activity of different tissues and organs.[31] Results from the Whitehall II Study[32] showed that short sleep duration was associated with lower scores on most cognitive
function tests. Sleep deprivation was also associated with reduced attention and working
memory, in addition to long-term memory reduction and decision-making.[33] Moreover, sleep deprivation may result in a complex set of changes in brain activity
and connectivity.[34]
Short sleep duration may result in several consequences for health, such as stress
responsivity, somatic issues, reduced quality of life, emotional distress, in addition
to mental and memory problems.[29] On the other hand, prolonged sleep duration (here categorized as nine or more hours
a day) may lead to chronic diseases, including hypertension, dyslipidemia, cardiovascular
diseases, obesity, metabolic syndrome, and diabetes.[29]
[35]
[36] Such implications may explain the results of the present study, since multimorbidity
comprises most of these health issues.
As observed in our results, in addition to the impact of sleep disorders on several
functions of the human body, these changes can cause the accumulation of different
chronic diseases, worsening quality of life. Strategies such as exercises, meditation,
and healthy eating can improve sleep quality.[37]
[38]
[39] In a meta-analysis with 18 trials and more than 1600 participants, Rusch et al.[38] found evidence that mindfulness meditation may enhance sleep quality. Another meta-analysis,[39] including 557 participants, showed that exercise improved sleep quality without
notable adverse effects.
Our data were collected during the COVID-19 pandemic, and they may provide important
information on both the burden of multimorbidity and sleep patterns that may have
been altered throughout the pandemic. Recently, evaluating data from waves 1 and 2
of the PAMPA study, we found a 27.1% incidence of multimorbidity, with higher rates
among women and the elderly.[40] A cross-sectional study[41] conducted in Brazil during the pandemic pointed out that sleep quality significantly
worsened during social distancing. Furthermore, in an online survey[42] of just over 45 thousand respondents from Brazil, 43.5% of the sample reported sleep
issues during the pandemic.
To the best of our knowledge, the present is the first study with a Brazilian population
investigating the associations of sleep duration and medication use with multimorbidity
during the COVID-19 pandemic. The present study includes a high number of participants
from one of the largest states in the Brazil. However, the study also has some limitations.
First, the fact that the multimorbidity was self-reported may have caused us to include
fewer patients with it than there are in reality, as some participants may not know
they have a specific chronic condition. The second limitation refers to the fact that
our study did not include questions about sleep quality. Finally, the cross-sectional
design may imply reverse causality in our association of interest (multimorbidity
and sleep duration). Despite this, longitudinal studies with other populations[25]
[28] have reported results similar to ours, demonstrating that inadequate sleep duration
and sleep disturbances are associated with multimorbidity. With the present findings,
we recommend that longitudinal studies and clinical trials be performed to confirm
the association found in our analyses.
As a conclusion, short (five hours or less) and prolonged (nine hours or more) sleep
duration increased the odds of multimorbidity, regardless of sociodemographic and
behavioral characteristics. The use of sleeping medication was also associated with
multimorbidity. The results of the present study are important but require caution
due to reverse causality, and longitudinal studies are needed to confirm the findings.
