Prevalence of suspected obesity hypoventilation syndrome in Hungarian Intensive Care Units during the COVID‐19 pandemic

Abstract Introduction The symptoms of obesity hypoventilation syndrome (OHS) may be present for years with concomitant progressive comorbidities, and the condition is frequently diagnosed late as a result of acute‐on‐chronic hypercapnic respiratory failure. Although some data exist on intensive care unit (ICU) prevalence, mortality and morbidity of OHS, little is known about the ICU mortality of these chronic respiratory failure patients during the COVID‐19 pandemic. Methods We performed a cross‐sectional observational study in five Hungarian Intensive Care Units for 4 months during the COVID‐19 pandemic. All ICU patients were screened for OHS risk factors by treating physicians. Risk factors were defined as obesity (body mass index [BMI] ≥ 30 kg/m2) and at least one of the following: Epworth Sleepiness Score ≥ 6; symptoms of right heart failure; daytime or night‐time hypoxemia; presence of loud snoring; witnessed apnoea. We calculated prevalence, mortality and factors associated with unfavourable outcome. Results A total of 904 ICU patients were screened for OHS risk factors. Overall 79 (8.74 ± 5.53%) patients were reported to have met the criteria for suspected OHS with a mortality rate of 40.5%; 69% (54 patients) of the cohort displayed at least 3 symptoms related to OHS before their acute illness. COVID‐19 infection was associated with higher mortality in OHS‐suspected patients, independently of actual BMI. Conclusion Despite the increased risk of obese patients, suspected OHS did not show higher prevalence than expected during the COVID‐19 pandemic in critically ill patients. COVID‐19 infection however was a risk for mortality in these patients, independent of actual BMI.


| INTRODUCTION
The growing rates of obesity worldwide pose several clinical challenges, including the emergence of obesity hypoventilation syndrome (OHS) as a leading cause of chronic respiratory impairment. 1,2 OHS is defined as the combination of sleep-related hypoventilation, resting daytime hypercapnia (arterial carbon dioxide [P a CO 2 ] ≥ 45 Hgmm) and a body mass index (BMI) of ≥30 kg˙m À2 in the absence of alternative cause for alveolar hypoventilation. 3 OHS is also associated with significant morbidity, leading to increased rates of chronic heart failure, pulmonary hypertension and metabolic syndrome. 4 Although symptoms indicative of the condition and its comorbidities may be present for years, OHS is frequently diagnosed late in its course as acuteon-chronic hypercapnic respiratory failure. [5][6][7][8] As such, intensive care units (ICUs) are an important healthcare setting for patients with OHS.
Data are limited regarding the prevalence of OHS in the ICU and its effect on mortality in different settings. While the overall prevalence of obesity is 20-30% in ICUs, a previous retrospective monocentric study found that 8% of ICU admissions met the criteria of OHS and extreme obesity (BMI > 40 kg˙m À2 ). 9 Both the obese and OHS patients may benefit from the obesity paradox, which results in similar or even decreased mortality rates compared with non-obese patients. 5,[10][11][12] This protective effect was seemingly diminished by COVID-19. During the pandemic, the obese made up the majority of critically ill patients (up to 48%) and were more likely to require invasive mechanical ventilation and die (with a mortality rate of 28-35%). [13][14][15][16] Despite extensive data on outcomes for the obese in COVID-19, little is known about the ICU prevalence of OHS and the mortality of this cohort during the COVID-19 pandemic.
The current study aimed to assess suspected OHS prevalence and associated outcomes among critically ill patients during the COVID-19 pandemic. We conducted a multicenter cross-sectional investigation in general ICUs. Primary outcomes were the prevalence and mortality of suspected OHS in critically ill patients. Secondary outcomes were relative risk of suspected OHS prevalence and mortality due to COVID-positive status, pneumonia or need for invasive ventilation (IV).

| Design of the study
Eleven high case number (minimal bed number of 12), mixed case patient population ICUs participating in COVID-19 management were invited to take part in the study. Out of the 11, 5 ICU sites took part in the study. Investigating physicians were asked to screen all critical care patients for risk factors of OHS during the study period between the 1st of October and the 30th of November, 2020 and again between the 1st of October and 30th of November, 2021. Overall patient number, COVID-19 prevalence and mortality were recorded for all participating units during the study period. Patients qualifying as suspected OHS were included in the study. Patients were excluded if they developed OHS-related symptoms due to acute illness. We also excluded patients aged under 18 years. Written informed consent was provided by all participating patients or their next of kin. Data were collected anonymously. The study protocol was approved by the regional ethical committee of Semmelweis University (SE RKEB 52/2020.).

| Screening
Suspected OHS was defined as a BMI of ≥30 kg/m 2 and the presence of at least one OHS-related risk factor before the current acute illness (Table 1). 2 Epworth Sleepiness Score was calculated to evaluate daytime sleepiness; a score of ≥6 was considered abnormal. 17 Symptoms of right heart failure were considered according to European Society of Cardiology guidelines for the diagnosis and treatment of acute and chronic heart failure. 18 Daytime and nighttime hypoxemia was based on previous outpatient or in-hospital medical records. The presence of loud snoring and witnessed apnoea was based on history collected from next of kin. In case of missing data, the risk factor was assumed to be absent.
Anthropometry data were collected by critical care physicians based on the medical records of the patients and information given by the patients or relatives. BMI was calculated on admission.
Information about past medical history, comorbidities such as diabetes, congestive heart disease, pulmonary hypertension and ischaemic heart disease, was based on information provided by healthcare professionals, information in previous medical reports and information based on findings during critical care treatment. Chronic kidney disease nomenclature was used according to Kidney Disease Improving Global Outcomes 2012 Clinical Practice Guidelines for the Evaluation and Management of Chronic Kidney Disease. 19 Hospital admission and ICU admission data in the last 12 months were based on the electronic medical record of the patient. COVID-19 status was based on a polymerase chain reaction (PCR) test (sputum, tracheal aspirate or nasopharyngeal swab) performed at any time during the course of the acute illness resulting in the study hospital admission. The total number of days spent in the ICU defined the ICU LOS. The total number of days spent with invasive respiratory support defined LOV. Hypercapnia (p a CO 2 ≥ 45 mmHg) and increased bicarbonate (HCO 3 À ) levels (≥27 mmol/L) on arterial blood gas measured at the time of discharge from ICU defined alveolar hypoventilation. 11

| Statistical analysis
The results are expressed as median (±standard deviation) for continuous variables and as frequency (percentage) for categorical variables. Different patient groups (OHS suspected vs. OHS not suspected, COVID-19 positive vs. negative patients, IV required vs. not required, survivors vs. non-survivors, different BMI and age groups) were compared with 2 Â 2 or 3 Â 2 Pearson's Chi-square. P < 0.05 value was considered significant. All statistical analyses were performed with SPSS (IBM Corp., Armonk, NY).

| RESULTS
The five participating ICUs treated a total of 904 patients during the study period. Overall mortality reported by the units was 35.40%. Overall COVID-19 prevalence of the ICUs was 58.96% during the study period. Out of the 904 ICU patients screened, 79 patients (8.74 ± 5.53%) were reported to have risk factors for OHS. Prevalence of suspected OHS varied by site, and a significant difference was observed between ICU sites (p < 0.001) (see Figure 1).
Daytime sleepiness scores were reported for 55 out of 79 patients and were abnormal for 45 patients (81.8%) ( Table 2). Snoring, nocturia and pitting oedema were the most frequent symptoms among the reported individuals (86.1%, 69.6% and 65.8%, respectively). Seventeen patients (21.5%) had witnessed apnoea by relatives; 69% of the F I G U R E 1 The overall prevalence of patients with suspected obesity hypoventilation syndrome (OHS) in intensive care unit (ICU). cohort displayed at least 3 symptoms related to OHS ( Figure 2).
Characteristics of the study patients are listed in Table 2 (Figure 3).

| DISCUSSION
Our multicentre, cross-sectional observational study aimed to assess the prevalence and mortality of suspected OHS in ICUs during the COVID-19 pandemic. The mean prevalence of suspected OHS was 8.74% in our examined sites, with patients showing frequent persisting symptoms, and potential signs of chronic hypoventilation at discharge. Mortality of patients with suspected OHS was similar to patients without OHS risk factors but increased significantly if COVID-19 infection was present.
The overall prevalence of OHS is 0.15-0.3% in the general population. 2 This data are based on the  As previous studies have shown, patients with obesity are admitted to the ICU frequently, and increased BMI categories are associated with comorbidities such as diabetes mellitus, cardiac failure and respiratory failure. 23 During the COVID-19 pandemic, the prevalence of obesity in ICUs reached almost 50%; moreover, increased mortality rate and frequent need for invasive mechanical ventilation were observed among obese and severely obese patients. 15,22 In our study, both pneumonia and COVID-19 infection were risk factors for mortality and the need for invasive mechanical ventilation. However, the actual mortality rate was independent of BMI categories. These findings suggest that the presence of the risk factors for OHS and chronic respiratory failure may carry an increased risk in COVID-19 independently of the actual body weight.
Among critically ill patients, obesity may be associated with greater survival. 23 This paradoxical protective effect was apparent in OHS-suspected patients as well 13 but was negatively affected by COVID-19. 22 In our study, OHS-suspected patients displayed higher mortality rates than was expected by the obesity paradox. This was mainly driven by COVID-19 infections, as suspected OHS patients without COVID-19 had similar mortality F I G U R E 2 Cumulative percentage of obesity hypoventilation syndrome (OHS) suspected intensive care unit (ICU) patients with OHS-related symptoms. OHS-related symptoms: Epworth Sleepiness Score ≥ 6; pitting oedema; nocturia; loud snoring; witnessed apnoea.
T A B L E 3 Combined mortality rate in the ICU sites and among the OHS suspected patients. rates to previously reported historical data (10-16.7%). 12, 23 We also found that COVID-19-positive status increased mortality independent of actual BMI. These data suggest that the presence of suspected OHS is a risk factor for mortality in COVID-19 positive patients independent of actual body weight. This is a strong indication that obese patients with risk factors of chronic respiratory failure may be more fragile during the pandemic.
As previous studies highlighted, OHS patients are predominantly diagnosed late, during an episode of acute hypercapnic respiratory failure. 11 It is also known that some of these patients, even in this stage, are falsely diagnosed with chronic obstructive pulmonary disease (COPD) rather than OHS. 9 Our study highlighted that more than 60% of OHS-suspected patients carry at least three symptoms connected to OHS besides obesity. It is also noteworthy that more than 60% of patients had at least one documented comorbidity related to OHS, meaning OHS could have been suspected and diagnosed before their current hospitalization. 2 There is an urgent need to draw attention to this population of patients who might benefit from advanced screening for OHS.
Additionally, even before the pandemic, the mortality of ICU survivors with subsequently untreated OHS was 24-46% in the years following ICU discharge. 2,11 According to the American Thoracic Society guideline, OHS patients with persisting hypercapnia should receive non-invasive ventilation (NIV) after critical care treatment of a hypercapnic respiratory failure episode until further evaluation in a sleep laboratory. 4 It has been shown that long-term NIV treatment initiated after acute hospitalization can improve the quality of life significantly in this patient group. 24 Almost one-third of survivors in our study had elevated p a CO 2 and/or HCO 3 À levels at the time of discharge from the ICU, potentially qualifying them for long-term respiratory support, 25,26 which can improve survival and quality of life. 24 This reinforces that ICUs are important sites for flagging patients with suspected OHS. In summary, we found that the overall ICU prevalence of suspected OHS during the COVID-19 pandemic was similar to historical data, despite the higher prevalence of the obese in critically ill patients. Several reasons may be suspected behind this apparent 'protective' effect, most obviously that multimorbid patients were more likely to take precautionary measures and avoid infection T A B L E 4 The impact of anthropometric data, admitting diagnosis and past medical history on outcome. during the pandemic. Mortality of COVID-19-positive OHS suspected patients was significantly higher than mortality of COVID-19-negative patients, independent of BMI categories, highlighting the fragility of this population during the pandemic. Despite these important findings, our study has some limitations. The relatively short observational period might have influenced our results, as seasonal fluctuation of patients with respiratory insufficiency in ICUs is well known. 27 The markedly different prevalence of suspected OHS at the different study sites might indicate different clinical practices during the pandemic, but our study did not focus on uncovering reasons for this occurrence. Additionally, the study was executed during the second and fourth wave of the COVID-19 pandemic in Hungary with different dominant strains, which might have influenced the prevalence of obesity, and mortality rates observed in our study.

| CONCLUSION
In conclusion, we found that the prevalence of suspected OHS in ICUs was similar during the COVID-19 pandemic compared with historical data, despite the well-known increased prevalence of obesity in critically ill patients during this time. The mortality rate for the COVID-19-positive suspected OHS population was high (40.5%) and independent of BMI. This suggests that suspected OHS was protective from hospitalization with critical illness during the pandemic but was a risk factor for death once COVID-19 infection occurred. Further studies are needed to confirm and explain the potential reasons for these observations.