Keywords
COVID-19 - obstructive sleep apnea - sleep - polygraphy
Introduction
The coronavirus disease 2019 (COVID-19) pandemic caused by the severe acute respiratory
syndrome coronavirus 2 (SARS-CoV-2) is still under investigation. The main risk factors
associated with the most severe forms of SARS-CoV-2 manifestation are cardiovascular
and metabolic diseases[1]
[2]
Overweight and obesity are also risk factors for severe pneumonia in individuals with
COVID-19,[3] and great concern exists regarding the incidence of SARS-CoV-2 in patients who are
overweight, obese, or both. Another factor is the relationship between obstructive
sleep apnea (OSA) and major comorbidities associated with severe COVID-19.[4]
[5]
Most clinical case series of patients with severe COVID-19 reported associations between
OSA and male sex, obesity, age, and cardiometabolic disorders.[6] The low incidence of OSA diagnosis in high-risk populations is congruent with the
underdiagnosis of this disease.[7]
Another important factor is that OSA is associated with reduced lung function and
increased lung inflammation even when obesity is controlled. This partially explains
why patients with OSA present a high risk of developing pneumonia, which plays a significant
role in the progression of COVID-19 infection.[8]
Furthermore, unexplored mechanisms that may link the imbalance between angiotensin-II
receptor, angiotensin-II converting enzyme (the entry receptor for SARS-CoV-2)[9] and severe COVID-19 infections may also be applied to OSA. Studies with patients
with untreated OSA demonstrated increased angiotensin-converting enzyme[9] expression and dysregulation of the renin-angiotensin system mainly due to chronic
intermittent hypoxia.[10]
Understanding the possible link between OSA and severe COVID-19 outcomes may generate
important information and encourage investments in treating sleep disorders. Therefore,
this study investigated OSA as a factor associated with COVID-19 severity.
Materials and Methods
Study Design
This descriptive observational study was conducted in a tertiary hospital in Ceará
(Brazil) from April to May 2021. The data corresponded to the second wave of SARS-CoV-2
infection in Brazil.
Participants
Participants were recruited by convenience from COVID-19 isolation wards at hospital
discharge. We included individuals aged ≥ 18 years with clinical diagnosis of COVID-19
confirmed by reverse transcriptase-polymerase chain reaction (RT-PCR) and able to
perform sleep polygraphy. Individuals on oxygen therapy, invasive or noninvasive mechanical
ventilation, or tracheostomized were not included. The exclusion criteria were previous
pulmonary disease causing significant obstructive or restrictive disorder, previous
neuromuscular disease, or pulmonary hypoventilation by any cause.
The study followed the Declaration of Helsinki[11] and resolution 466/2012 of the National Health Council. Assessments were initiated
after written informed consent and understanding about the protocol. The study was
approved by the research ethics committee of the Hospital Universitário Walter Cantídio
(CAAE: 40734920.8.3002.5045).
Initial Assessment and Hospitalization Data
Data regarding laboratory and imaging examinations, length of hospital stay, need
for invasive and noninvasive mechanical ventilation, need for oxygen therapy, and
clinical complications were collected from medical records. Data from medical records
refer to the worst clinical situation presented by the participants throughout the
hospital stay.
Coronavirus disease 2019 severity was classified as mild (mild clinical symptoms and
no signs of pneumonia on imaging examination), moderate (fever and respiratory symptoms
with radiological evidence of pneumonia), severe (respiratory distress [> 30 breaths/min],
oxygen saturation < 93% at rest, arterial partial pressure of oxygen/fraction of inspired
oxygen of < 300 mm Hg, or chest imaging showing lesion progression of > 50% within
24–48 hours), or critical (respiratory failure and mechanical ventilation, shock,
or other organ failure requiring intensive care unit [ICU]).[12]
Quantitative Sleep Assessment
All participants underwent respiratory polygraphy at bedside using a type-III portable
multichannel device (PolyWatch, BMC Medical, Beijing, China), which includes a nasal
flow cannula, chest strap, and oximeter for recording peripheral oxygen saturation
and heart rate. Participants also used the ActTrust2 Actigraph (Condor Instruments,
Vila Madalena, SP, Brazil), a noninvasive method for monitoring rest and activity
cycles. The actigraph was used to assess whether the participant was in a sleep state
during polygraphy to attenuate the limitation of the typ-III sleep study for diagnosing
respiratory disorders. Thus, participants with short sleep time (according to the
actigraph) were excluded from the study.
Before the exam, participants were instructed on the placement and maintenance of
the devices, and recordings started only after answering the doubts of participants.
Instructions were also given to stop recordings upon awakening in the morning by pressing
a button on the polygraph monitor.
Data Analysis
The PolyLogic Sleep Analysis software (BMC Medical) was used to analyze the respiratory
variables of the polygraphy, while the ActStudio software (Condor Instruments) assessed
the actigraphy records. For sleep staging, the polygraphy record was divided into
time intervals of 30 seconds, and respiratory events were divided into 120 seconds.
Two certified professional experts in sleep studies manually analyzed the data, and
sleep and associated events were scored according to the AASM Manual for the Scoring
of Sleep and Associated Events.[13] Obstructive sleep apnea severity was classified according to the respiratory event
index as mild (between 5 and 14.9/hour), moderate (between 15 and 29.9/hour), or severe
(≥ 30/hour). The respiratory event index and minimum oxygen saturation were analyzed
as study outcomes. After these analyses, participants were divided into two groups
based on the presence (OSA + ) or absence (OSA-) of OSA.[13]
Statistical Analysis
Data distribution was analyzed using the Shapiro-Wilk test, and results were presented
as mean ± standard deviation (SD). Continuous data were compared using the unpaired
t-test while the χ2 test analyzed categorical data. Qualitative variables were expressed
in absolute and relative frequencies. Data were analyzed using the Statistical Package
for Social Science, version 20.0 (IBM Corp., Armonk, NY, USA); significance was set
at p < 0.05.
Results
Eligibility
Nineteen participants with COVID-19 meeting the eligibility criteria were contacted.
Of these, five refused participation when informed about using the devices for nighttime
polygraphy, while two presented technical failures in the polygraph records. Thus,
this study included 12 participants: 7 with OSA- and 5 with OSA+ ([Fig. 1]).
Fig. 1 Flowchart of participant selection.
Sample Characteristics
[Table 1] presents the demographic data, symptoms, and comorbidities of participants. The
most reported symptoms were dyspnea and cough; comorbidities (e.g., hypertension,
diabetes, anemia, and heart failure) were similarly distributed in both groups.
Table 1
Demographic data, symptoms, comorbidities, and functioning according to groups.
|
Individuals with COVID-19
|
|
|
OSA+ (n = 5)
|
OSA- (n = 7)
|
p-value
|
|
Age, mean (±SD)
|
55 (13.5)
|
53.42 (10.43)
|
0.847[a]
|
|
Sex, N (%)
|
|
|
|
|
Male
|
3 (60)
|
3 (42.85)
|
0.558[a]
|
|
Symptoms, N (%)
|
|
|
|
|
Headache
|
1 (20)
|
2 (28.57)
|
0.734[b]
|
|
Dyspnea
|
5 (100)
|
6 (85.71)
|
0.377[b]
|
|
Fever
|
2 (20)
|
3 (42.85)
|
0.921[b]
|
|
Myalgia
|
1 (20)
|
3 (42.85)
|
0.408[b]
|
|
Cough
|
5 (100)
|
5 (71.42)
|
0.190[b]
|
|
Comorbidities, N (%)
|
|
|
|
|
SAH
|
1 (20)
|
1 (14.28)
|
0.793[b]
|
|
DM
|
1 (20)
|
2 (28.57)
|
0.735[b]
|
|
HF
|
1 (20)
|
1 (14.28)
|
0.793[b]
|
|
Obesity
|
1 (20)
|
1 (14.28)
|
0.793[b]
|
|
Anemia
|
1 (20)
|
1 (14.28)
|
0.793[b]
|
|
No comorbidities
|
2 (40)
|
3 (42.85)
|
0.921[b]
|
Abbreviations: DM: diabetes mellitus; HF, heart failure; OSA, obstructive sleep apnea;
SAH, systemic arterial hypertension.
Data on age are presented as mean and standard deviation and other variables as absolute
and relative values.
a Unpaired t-test.
b : χ2 test.
Polygraphic Characteristics
Polygraphic recordings showed a significantly lower oxygen saturation in the OSA+
than OSA- group (77.6% ± 7.89% vs 84.4% ± 2.57%; p = 0.041) ([Fig. 1]). However, no difference (p = 0.180) was found in apnea duration between the OSA+ and OSA- groups (25.4 ± 19.7 second
vs 12.1 ± 12.3 seconds) ([Fig. 2] and [3]).
Fig. 2 Minimum oxygen saturation of individuals with COVID-19 in the OSA+ and OSA- groups.
Abbreviation: OSA: obstructive sleep apnea. Unpaired t-test. *p = 0.041.
Fig. 3 Apnea duration (in seconds) of individuals with COVID-19 in the OSA+ and OSA- groups.
Abbreviation: OSA: obstructive sleep apnea. Unpaired t-test. *p = 0.180.
[Table 2] presents the associations between COVID-19 severity and OSA diagnosis. Coronavirus
disease 2019 was more severe in individuals from the OSA+ (100%) than OSA- group (28.57%)
(p = 0.013). The OSA+ group also received oxygen therapy for a longer time than the
OSA- group (p = 0.038). Regarding length of hospital stay, no significant difference was observed
(p = 0.268).
Table 2
Associations between COVID-19 severity, OSA diagnosis, and clinical variables.
|
Individuals with COVID-19
|
|
|
OSA + (n = 5)
|
OSA- (n = 7)
|
p-value
|
|
COVID-19 severity, N (%)
|
|
|
|
|
Moderate
|
0 (0.0)
|
5 (71.42)
|
*0.013[a]
|
|
Severe
|
5 (100)
|
2 (28.57)
|
*0.013[a]
|
|
Clinical variables, mean (±SD)
|
|
|
|
|
Length of hospitalization (days)
|
15 (3)
|
12 (4)
|
0.268b
|
|
Oxygen therapy time (days)
|
12 (4)
|
5 (2)
|
*0.038b
|
Abbreviation: OSA, obstructive sleep apnea.
a : χ2 test; b: Student's t-test. *p-value < 0.05.
[Table 3] presents the comparisons between groups considering blood cell and biochemical variables.
The OSA+ group presented a thrombocytosis profile (p = 0.008) compared with the OSA- group. Higher rates of D-dimer were also found in
the OSA+ (1443 ± 897 ng/ mL) compared with the OSA- group (648 ± 263 ng/ mL) (p = 0.019).
Table 3
Comparison of laboratory variables and presence of obstructive sleep apnea.
|
Individuals with COVID-19
|
|
|
OSA+ (n = 5)
|
OSA- (n = 7)
|
p-value
|
|
Laboratory variables, mean (±SD)
|
|
|
|
|
RBCs (millions/ µL)
|
4.28 (0.81)
|
4.77 (0.50)
|
0.242
|
|
Hemoglobin (g/ dL)
|
10.5 (3.53)
|
13.53 (1.18)
|
0.374
|
|
Hematocrit (%)
|
36.37 (7.25)
|
40.27 (3.80)
|
0.250
|
|
Neutrophil (/µL)
|
67.08 (9.13)
|
64.23 (14.64)
|
0.710
|
|
Leukocyte (/mm3)
|
11,846.80 (2,771.30)
|
8895.43 (2384.39)
|
0.076
|
|
Lymphocyte (/mm3)
|
23.60 (0.39)
|
27.18 (13.36)
|
0.602
|
|
Platelets (/mm3)
|
449,240 (75,503)
|
229,827 (132,865)
|
0.008*
|
|
PAT (s)
|
11.95 (1.02)
|
12.1 (2.54)
|
0.601
|
|
APTT (s)
|
29.35 (2.86)
|
24.25 (0.35)
|
0.500
|
|
PCr (mg/dL)
|
0.84 (0.73)
|
1.01 (0.91)
|
0.542
|
|
Urea (mg/dL)
|
40 (9)
|
51 (32)
|
0.394
|
|
Creatinine (mg/dL)
|
0.76 (0.09)
|
0.81 (0.20)
|
0.578
|
|
D-dimer (ng/mL)
|
1,443 (897)
|
648 (263)
|
0.019*
|
|
CPK (U/L)
|
228 (318)
|
61 (54)
|
0.458
|
Abbreviations: APTT, Activated partial thromboplastin time; CPK, creatine phosphokinase;
CRP, C-reactive protein; PAT, prothrombin activity time; RBCs, red blood cells.
Unpaired t-test.
Discussion
The findings revealed that individuals hospitalized due to COVID-19 and with OSA presented
more severe symptoms than those without OSA as well as higher platelet and D-dimer
counts, longer oxygen therapy time, and worse peripheral oxygen saturation. Our results
corroborate a previous study exploring the relationships between greater COVID-19
severity and OSA.[14]
Regarding clinical outcomes during hospitalization, our results can be compared with
those of Mashaqi et al.,[14] who conducted a cohort analysis with 1,738 individuals with COVID-19 and OSA and
1,599 without OSA and observed a statistical significance in ICU admission. However,
this association was attenuated when the model was adjusted for age, sex, body mass
index (BMI), and comorbidities. A retrospective study by Cade et al.[15] with 443 patients also found that the increased risk of intubation, ICU admission,
or hospitalization associated with OSA was attenuated after adjustment for demographic
data, BMI, and comorbidities. Both studies reported OSA diagnosis by reviewing medical
and health records.
Our study also revealed that individuals with OSA need oxygen therapy for longer,
which may be associated with low nocturnal oxygen saturation due to apnea and hypopnea
and high COVID-19 severity. In this context, oxygen therapy must be well evaluated,
considering that continuous positive airway pressure is the gold standard for reversing
apnea and hypopnea events and improving saturation.[16] In addition, prolonged and unnecessary administration of oxygen therapy may expose
the patient to the harmful effects of oxygen.[17]
In contrast to Mashaqi et al.,[14] we found elevated platelet levels and higher D-dimer rates in individuals with OSA.
These findings can be explained by the hypercoagulability of OSA[18] and characteristics of COVID-19, which may increase the risks of developing thrombotic
conditions.
Previous studies investigating the influence of OSA on other lung injuries warned
about the possible relationships with COVID-19 severity and increased risk of developing
community-acquired pneumonia[19] and perioperative acute respiratory distress syndrome.[20] Pathophysiological mechanisms may explain the associations between greater disease
severity and worse clinical outcomes. Untreated OSA progresses with repeated airway
obstruction and generates negative intrathoracic pressure; thus, associated shear
forces may favor inflammatory processes with worsening lung injury. Also, increased
sympathetic outflow during OSA episodes promotes catecholamine release, which may
increase the risk of cardiovascular complications (e.g., arrhythmias, cardiac ischemia,
and hypercoagulability).[18]
Our study has clinical implications. Given the repercussions on the clinical outcome
of individuals hospitalized due to COVID-19, the presence of OSA should be viewed
as a potential comorbidity and risk factor for adverse COVID-19 outcomes. Patients
with suspected OSA should be carefully evaluated and promptly treated. Considering
the modulating and regulatory effects of sleep on the immune system,[21] adequate treatment may reduce the COVID-19 evolution or other lung lesions.
Some limitations also need to be highlighted. A small sample size increases the chances
of β error and limits the external validity, while clinical data collected using medical
records increases the risk of information bias. In addition, the cross-sectional design
of the study hinders the capacity to infer associations, different from longitudinal
studies. Another limiting factor was that the sample was composed of hospitalized
individuals, which may impair polygraphic recordings because data were collected during
sleep in the ward. We encourage longitudinal studies with a larger sample size to
ensure the external validity of the results.
The greatest strength of our study is the use of type-III polygraphy, which enhances
the precision and reliability of OSA diagnosis and enables the extraction of quantitative
sleep data. In addition, we used actigraphy to ensure the adequacy and comparability
of the polygraph recording time with the sleep time recorded by the actigraph.
Conclusion
High COVID-19 severity was associated with OSA diagnosis. Also, individuals with COVID-19
and OSA presented longer oxygen therapy time, higher platelet count and D-dimer, and
worse peripheral oxygen saturation than those without OSA.
