*Article* **Predictors of SARS-CoV-2 IgG Spike Antibody Responses on Admission and Clinical Outcomes of COVID-19 Disease in Fully Vaccinated Inpatients: The CoVax Study**

**Eleni Livanou 1,†, Erasmia Rouka 1,\* ,† , Sotirios Sinis <sup>1</sup> , Ilias Dimeas <sup>1</sup> , Ioannis Pantazopoulos <sup>1</sup> , Dimitrios Papagiannis <sup>2</sup> , Foteini Malli <sup>2</sup> , Ourania Kotsiou <sup>2</sup> and Konstantinos I. Gourgoulianis <sup>1</sup>**


**Abstract:** Background: SARS-CoV-2 vaccines have shown high efficacy in protecting against COVID-19, although the determinants of vaccine effectiveness and breakthrough rates are yet to be determined. We aimed at investigating several factors affecting the SARS-CoV-2 IgG Spike (S) antibody responses on admission and clinical outcomes of COVID-19 disease in fully vaccinated, hospitalized patients. Methods: 102 subjects were enrolled in the study. Blood serum samples were collected from each patient upon admission for the semiquantitative determination of the SARS-CoV-2 IgG S levels with lateral flow assays. Factors influencing vaccine responses were documented. Results: 27 subjects had a negative antibody test upon hospital admission. Out of the 102 patients admitted to the hospital, 88 were discharged and 14 died. Both the absence of anti-S SARS-CoV-2 antibodies and poor clinical outcomes of COVID-19 disease were associated with older age, lower Ct values, and a shorter period between symptom onset and hospital admission. Ct values and time between symptom onset and hospitalization were independently associated with SARS-CoV-2 IgG S responses upon admission. The PaO2/FiO2 ratio was identified as an independent predictor of in-hospital mortality. Conclusions: Host- and disease-associated factors can predict SARS-CoV-2 IgG S responses and mortality in hospitalized patients with breakthrough SARS-CoV-2 Infection.

**Keywords:** breakthrough COVID-19 hospitalizations; clinical outcomes; SARS-CoV-2 IgG Spike responses; vaccine-induced immunity

#### **1. Introduction**

The ongoing COVID-19 pandemic, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has become a significant global public health issue [1]. As of 28 January 2022, there has been 364,191,494 confirmed cases of COVID-19 worldwide, with 5,631,457 deaths reported to the World Health Organization (https://covid19.who.int/, accessed: 29 January 2022). SARS-CoV-2 causes a variety of symptoms ranging from mild, flu-like symptoms to severe pulmonary damage with respiratory distress syndrome and death [2]. Subjects with pre-existing comorbidities including obesity, cardiovascular disease, type 2 diabetes mellitus (T2D), and chronic renal and lung disease are at an increased risk of developing acute respiratory distress syndrome (ARDS), requiring mechanical ventilation and admission to the intensive care unit (ICU) [3].

Vaccination is the most cost-effective medical intervention, preventing millions of deaths every year [4]. Vaccines have greatly reduced the burden of infectious diseases [5] and constitute an important tool for limiting epidemics caused by emerging pathogens [4].

**Citation:** Livanou, E.; Rouka, E.; Sinis, S.; Dimeas, I.; Pantazopoulos, I.; Papagiannis, D.; Malli, F.; Kotsiou, O.; Gourgoulianis, K.I. Predictors of SARS-CoV-2 IgG Spike Antibody Responses on Admission and Clinical Outcomes of COVID-19 Disease in Fully Vaccinated Inpatients: The CoVax Study. *J. Pers. Med.* **2022**, *12*, 640. https://doi.org/10.3390/ jpm12040640

Academic Editor: Luisa Brussino

Received: 6 March 2022 Accepted: 11 April 2022 Published: 15 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Vaccine-induced immunity is mediated by the complex interaction of innate, humoral, and cell-mediated immunity [6]. Vaccines operate by inducing an immune response and, as a result, an immunological memory, which protects against infection or disease [7].

The approved SARS-CoV-2 vaccines have been highly efficient in protecting against COVID-19 [8,9], although the determinants of vaccine effectiveness and breakthrough rates are yet to be determined, especially in light of the emergence of viral variants of concern [10]. Antibody responses to SARS-CoV-2 vaccines have been shown to be affected by a variety of factors, including age [11], sex [12], central obesity [13], hypertension [11,13], cancer [14], dyslipidemia [13], and smoking habits [11,13].

However, there is a plethora of factors that influence humoral and cellular vaccine responses in humans. These include intrinsic host factors as well as extrinsic, environmental, behavioral, nutritional, and vaccine factors [6]. Variables affecting the immune response to the SARS-CoV-2 vaccination have not been extensively investigated. In this study, we examine several factors that may have an impact on SARS-CoV-2 IgG Spike (S) antibody responses and the outcome of COVID-19 disease in fully vaccinated, hospitalized patients.

#### **2. Materials and Methods**

#### *2.1. Study Design*

Within two months we prospectively studied 102 fully vaccinated adult patients (71 men, 31 women) who were admitted to the COVID-19 Department of the University Hospital of Larissa, Greece. SARS-CoV-2 infection was verified by real-time reversetranscription polymerase chain reaction (RT-PCR). Several factors that influence vaccine responses were documented [6] (Table 1). The patients were monitored until hospital discharge or death. The study was approved by the Institutional Research Ethics Committee (46943/29.11.2021) and each participant provided written informed consent.

**Table 1.** The factors that were investigated in the present study in terms of their effect on SARS-CoV-2 IgG S antibody responses.


#### *2.2. Detection of the SARS-CoV-2 IgG S Protein-Specific Antibodies*

On the first day of hospitalization, blood serum samples were collected from each patient for the semiquantitative determination of the SARS-CoV-2 IgG S protein-specific antibodies with lateral flow immunochromatographic assays (Rapid Test 2019-nCoV IgG, ProGnosis Biotech, Larissa, Greece).

5 µL of serum per sample was injected into a test tube containing dilution buffer. A strip was then immersed in the tube for 15 min. Subsequently, the strips were scanned in the S-flow reader to interpret the results. The scanner could automatically calculate the ratio (T/C) by measuring the density of the test (T) and control (C) lines of the strip. Eight standards of recombinant antibodies were used in order to create the standard/ratio curve for the anti-S Ig semiquantification. A strip in which no colored line appeared in the control band was considered invalid.

In terms of diagnostic specificity, 468 samples of pre-pandemic COVID-19 patients were analyzed, with 100% specificity. A study was conducted with 122 patients who had clinical symptoms of COVID-19 and a positive PCR result for diagnostic sensitivity. The sensitivity was calculated to be 96.72% (Rapid Test 2019-nCoV IgG, V1430, Version 24 September 2021/rev.01, ProGnosis Biotech, Larissa, Greece).

Additional blood and serum samples were collected upon hospital admission for the evaluation of the following hematological and biochemical parameters: white blood cells (WBC), lymphocytes, platelets (PLT), C-reactive protein (CRP), creatinine, urea, aspartate transaminase (SGOT), alanine transaminase (SGPT), lactate dehydrogenase (LDH), ferritin, and creatine kinase (CPK).

#### *2.3. Statistical Analysis*

The SPSS v 19.0 software (IBM) was used to conduct the statistical analysis. Data distribution was assessed using the Kolmogorov–Smirnov normality test. The independent samples *T*-Test and the Mann–Whitney test were used to determine significant differences of parametric and non-parametric data, respectively, between two groups. Associations between categorical variables were determined with the Fisher's Exact Test. Correlations between quantitative variables were measured with the Pearson (r) or the Spearman (ρ) coefficients as appropriate. Logistic regression was used for the analysis of multiple variables influencing the presence of anti-S SARS-CoV-2 antibodies upon admission and the outcome of COVID-19 disease. All the variables with significant univariate associations were entered into the analysis in a single step (method selection: Enter). Statistical significance was set at the *p* < 0.05 level.

#### **3. Results**

#### *3.1. Baseline Characteristics of the Study Population*

The mean age of participants was 72.44 ± 1.22 years. Seventy-three subjects had received the BNT162b2/Pfizer vaccine, 22 the Vaxzevria, ChAdOx1-S/AstraZeneca vaccine, and 3 the Johnson & Johnson's Janssen COVID-19 Vaccine (information regarding the type of COVID-19 vaccine was unavailable for four subjects). The mean number of days since completion of vaccination was 159.03 ± 6.35. The mean real time PCR cycle threshold (Ct) value was 20.01 ± 0.54. The baseline laboratory characteristics of the study population are presented in Table 2.


**Table 2.** Baseline characteristics of the study population (*N* = 102).


**Table 2.** *Cont.*


**Table 2.** *Cont.*

\* SD; Standard deviation, \*\* Mann-Whitney U Test.

#### *3.2. Factors Influencing SARS-CoV-2 IgG S Antibody Responses*

Twenty-seven subjects had a negative antibody test upon hospital admission. In the remaining patients, the anti-S IgG antibodies ranged from 0.09AU to >12.48AU. A strong positive correlation was observed between the SARS-CoV-2 IgG S levels (when detectable) and Ct values upon admission (ρ = 0.592, *p* < 0.001). Compared to cases with detectable antibody levels, cases with a negative antibody test were older (*p* = 0.039) and had a higher creatinine level on admission (*p* = 0.001). The same group of patients was also observed to have lower Ct values (*p* < 0.001) and a shorter duration between symptom onset and

hospital admission (*p* < 0.001). The absence of anti-S SARS CoV-2-antibodies on the first day of hospitalization was also associated with the presence of diabetes (*p* = 0.006), PaO2/FiO2 (PF) ratio values <150 mm Hg (*p* = 0.014), and death (*p* = 0.019) (Table 2). The Ct values and time between symptom onset and hospitalization remained significant in the multiple regression analysis (*p* = 0.023 and *p* = 0.025, respectively) (Table 3).

**Table 3.** Results of the multiple regression analysis with respect to the variables affecting the presence of anti-S SARS-CoV-2 antibodies upon admission.


Dependent variable: detection of anti-S SARS-CoV-2 antibodies upon admission; Parameter coding (1): nondiabetic; PF ratio > 150 mm Hg. \* Days between symptom onset and hospitalization, \*\* B; the coefficient for the constant, S.E.; the standard error for B, Wald; the Wald chi-square test, df; the degrees of freedom for the Wald chi-square test, Exp(B); The exponentiation of the B coefficient, C.I; confidence interval.

#### *3.3. Factors Influencing the Outcome of COVID-19 Disease in Fully Vaccinated, Hospitalized Patients*

Out of the 102 patients admitted to the hospital, 88 were discharged and 14 died. All deceased subjects had received SARS-CoV-2 mRNA vaccines (*p* = 0.035). Poor disease outcome was associated with older age (*p* = 0.003), lower Ct values (*p* = 0.036), a shorter duration between symptom onset and hospital admission (*p* = 0.007), and lower BMI (*p* = 0.029). Non-deceased patients were more likely to have hypertension (*p* = 0.014) and PF ratio values > 150 mm Hg (*p* < 0.001) (Table 2). The PF ratio was identified by the multiple logistic regression model as an independent predictor of in-hospital mortality (*p* = 0.001) (Table 4). The "vaccine type" variable was not included in the multiple regression analysis since none of the deceased patients had received a viral vector COVID-19 vaccine.

**Table 4.** Results of the multiple regression analysis with respect to the variables affecting the outcome of COVID-19 disease in fully vaccinated, hospitalized patients.


Dependent variable: Mortality; Parameter coding (1): No detection of antibodies, non-hypertensive, PF ratio > 150 mm Hg. \* Days between symptom onset and hospitalization, \*\* B; the coefficient for the constant, S.E.; the standard error for B, Wald; the Wald chi-square test, df; the degrees of freedom for the Wald chi-square test, Exp(B); The exponentiation of the B coefficient, C.I; confidence interval.

#### **4. Discussion**

To our knowledge, this is the first study to assess several factors affecting SARS-CoV-2 IgG S antibody responses in fully vaccinated COVID-19 patients needing hospitalization due to severe COVID-19 disease. We found that older age, lower Ct values, and a shorter duration between symptom onset and hospital admission were associated with a lack of anti-S SARS-CoV-2 antibodies and poor clinical outcomes of COVID-19 disease.

The available evidence suggests that humoral and cellular immune responses are impaired in aged individuals, resulting in decreased vaccine responses [15]. Age has been reported to be inversely correlated with neutralizing antibody responses following the first immunization dose of BNT162b2, a finding that was particularly evident for individuals over 80 years [16]. The investigation of humoral immunity after two doses of BNT162b2 and mRNA-1273 vaccines has indicated that adults aged 18–55 years are more responsive to vaccination and maintain humoral immunity longer compared to individuals who are older than 70 years [17]. In addition, the anti-S SARS-CoV-2 immunoglobulin G antibody titers were found to be significantly lower in elderly vaccinees over the age of 80 years, with 31.3% of them having no detectable neutralizing antibodies after the second vaccine dose [18]. These observations and our findings underline the need for prioritizing booster COVID-19 vaccination in the elderly population.

Regarding the SARS-CoV-2 viral load, it has been shown that fully vaccinated subjects with breakthrough infections have a comparable peak viral load to those who are unvaccinated [19]. However, peak viral load increased with age, highlighting the importance of adjusting for age when comparing the two groups [20]. In our cohort of fully vaccinated inpatients, lower Ct values, which are indicative of higher viral loads, were associated with the absence of anti-S SARS-CoV-2 antibodies upon admission, both in the univariate and multiple regression analysis. The shorter number of days between symptom onset and hospital admission could account for the lower Ct values in the group of cases with a negative antibody test whose disease progressed faster, requiring earlier hospitalization. Lower Ct values were also observed in the group of deceased subjects, yet this finding did not remain significant in the multiple regression analysis. With respect to the positive correlation between anti-S SARS-CoV-2 IgG levels and Ct values upon admission, it has been reported that higher Ct values following BNT162b2 vaccination are associated with higher IgG concentrations [21].

The PF ratio was identified as an independent predictive variable of mortality in our cohort of fully vaccinated COVID-19 inpatients. Both the PF ratio and the ratio between standard PaO2 over FiO2 (STP/F) have been described as accurate predictors of acute respiratory failure outcome in COVID-19 patients [22].

Despite the fact that COVID-19 is characterized by atypical pneumonia followed by severe respiratory failure, about 10% of COVID-19 inpatients have been reported to endure acute kidney injury, which is linked to a poor prognosis [23]. It has been reported that changes in serum creatinine during the early stage of admission could predict mortality during hospitalization in COVID-19 patients [23,24]. In our study, serum creatinine levels upon admission were not predictive of in-hospital mortality, but subjects with a negative anti-S SARS-CoV-2 antibody test had higher creatinine levels on the first day of hospitalization compared to participants with detectable antibody levels, albeit not independently from other factors. Of interest, a multicenter cohort study of 543 subjects on hemodialysis and 75 healthy subjects found that both the humoral and cellular immune responses to SARS-CoV-2 vaccination were significantly impaired in the patients' group [25].

Findings with respect to diabetes were recently published as sub-study results for 92 patients of the CoVax study [26]. Diabetes mellitus, particularly T2D, is a prevalent comorbidity that considerably increases the risk of mortality in COVID-19 patients [27]. The immune system is thought to cause transitory alterations in systemic metabolism as a defense against viral infection. This mechanism is impaired in subjects with T2D, reducing the antiviral immune response [28].

Comorbidities related to a metabolic syndrome such as T2D, obesity, and hypertension are also characterized by low-grade chronic inflammation, which leads to immune system dysregulation and increased susceptibility to severe COVID-19 disease [3]. Paradoxically, in our cohort, deceased patients were less likely to have hypertension and their mean BMI was lower compared to non-deceased participants. The "obesity paradox" has been described in patient cohorts with several diseases including, but not limited to, T2D, hypertension, and chronic kidney disease [29]. However, caution is needed in interpreting these data given that all possible confounding variables should be taken into account and measured prospectively [29].

The remaining factors investigated in our study were not predictive of either the SARS-CoV-2 IgG S antibody responses or the outcome of COVID-19 disease in fully vaccinated inpatients. Gender and sex-specific effects have been reported to induce different immunization and adverse events outcomes [4,30]. The recent implementation of a within-host mathematical model of vaccine dynamics from lipid nanoparticle-formulated COVID-19 mRNA vaccines found no difference between sexes in the long-term duration of humoral immunity [17]. Regarding the "place of residence" variable, it has been reported that individuals living in highly deprived areas have increased odds of post-vaccination SARS-CoV-2 infection following the first vaccine dose [31].

It would be of great importance to ensure that the positive antibody test is a resultant of immunity induced exclusively by SARS-CoV-2 vaccination. Anti-S SARS-CoV-2 antibodies are produced in response to vaccine administration and/or COVID-19 infection. Thus, our method could not distinguish between post-vaccine response and infection. We also acknowledge that this is a single center study with a relatively small sample. Future studies should evaluate the parameters that have an impact on the vaccine-induced immunity against SARS-CoV-2 in subjects with breakthrough infections not requiring hospitalization.

#### **5. Conclusions**

Host- (age) and disease-associated factors (Ct values, time between symptom onset and hospitalization, and PF ratio) can predict SARS-CoV-2 IgG S responses and clinical outcomes in hospitalized COVID-19 patients with breakthrough SARS-CoV-2 infection post vaccination.

**Author Contributions:** K.I.G.; Conceptualization and supervision, E.L. and E.R.; formal analysis, investigation, data curation, writing—original draft preparation, S.S. and I.D.; resources, data curation, I.P., D.P., F.M. and O.K.; writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University Hospital of Larissa, Greece (46943/29.11.2021).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data that support the findings of this study are available upon request from the corresponding author.

**Acknowledgments:** The authors would like to thank Prognosis Biotech (Larissa, Greece) for providing the equipment and materials for the SARS-CoV-2 IgG Spike measurements.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Brief Report* **The Comparative Superiority of SARS-CoV-2 Antibody Response in Different Immunization Scenarios**

**Ourania S. Kotsiou 1,2,\* , Nikolaos Karakousis <sup>3</sup> , Dimitrios Papagiannis <sup>4</sup> , Elena Matsiatsiou <sup>1</sup> , Dimitra Avgeri <sup>1</sup> , Evangelos C. Fradelos <sup>1</sup> , Dimitra I. Siachpazidou <sup>2</sup> , Garifallia Perlepe <sup>2</sup> , Angeliki Miziou <sup>2</sup> , Athanasios Kyritsis <sup>2</sup> , Eudoxia Gogou <sup>2</sup> , George D. Vavougios <sup>2</sup> , George Kalantzis <sup>2</sup> and Konstantinos I. Gourgoulianis <sup>2</sup>**


**Abstract:** Background: Both SARS-CoV-2 infection and/or vaccination result in the production of SARS-CoV-2 antibodies. We aimed to compare the antibody titers against SARS-CoV-2 in different scenarios for antibody production. Methods: A surveillance program was conducted in the municipality of Deskati in January 2022. Antibody titers were obtained from 145 participants while parallel recording their infection and/or vaccination history. The SARS-CoV-2 IgG II Quant method (Architect, Abbott, IL, USA) was used for antibody testing. Results: Advanced age (>56 years old) was associated with higher antibody titers. No significant differences were detected in antibody titers among genders, BMI, smoking status, comorbidities, vaccine brands, and months after the last dose. Hospitalization length and re-infection were predictors of antibody titers. The individuals who were fully or partially vaccinated and were also double infected had the highest antibody levels (25,017 ± 1500 AU/mL), followed by people who were fully vaccinated (20,647 ± 500 AU/mL) or/partially (15,808 ± 1800 AU/mL) vaccinated and were infected once. People who were only vaccinated had lower levels of antibodies (9946 ± 300 AU/mL), while the lowest levels among all groups were found in individuals who had only been infected (1124 ± 200 AU/mL). Conclusions: Every hit (infection or vaccination) gives an additional boost to immunization status.

**Keywords:** antibody; COVID-19; infection; immunization; vaccination

### **1. Introduction**

The Coronavirus disease pandemic 2019 remains an excellent concern for ethnicities. It is already well-established that the SARS-CoV-2 virus is rapidly evolving and spreading through mutagenesis, a quite threatening condition that lengthens the duration of the pandemic and might affect the efficacy of the existing vaccines and lead to the need to develop new ones in order to confront new variants of the specific viral infection [1,2].

There is a debate regarding the durability of antibody responses over time in patients infected by SARS-CoV-2, with several studies reporting stable, long-lasting antibody immunity and others showing rapidly waning antibody immunity or late appearances with low antibody levels and/or a complete lack of antibodies [3].

FDA decided on booster vaccines because the benefits of the COVID-19 vaccination far outweigh the potential risks. However, further studies are needed to demonstrate the efficacy of booster vaccinations to determine the best dosing and mix-and-match schedules of vaccinations [3]. Nevertheless, the result of the combination of infection and vaccination on the antibody levels is unknown and leads to a condition of questioning and concern.

**Citation:** Kotsiou, O.S.; Karakousis, N.; Papagiannis, D.; Matsiatsiou, E.; Avgeri, D.; Fradelos, E.C.; Siachpazidou, D.I.; Perlepe, G.; Miziou, A.; Kyritsis, A.; et al. The Comparative Superiority of SARS-CoV-2 Antibody Response in Different Immunization Scenarios. *J. Pers. Med.* **2022**, *12*, 1756. https:// doi.org/10.3390/jpm12111756

Academic Editor: Mariangela Manfredi

Received: 31 August 2022 Accepted: 21 October 2022 Published: 23 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

In this study, we aimed to compare the titers of antibodies against SARS-CoV-2 in different scenarios for antibody production, which is of great importance, especially in the era of the pandemic in which we possess certain preventive tools such as vaccines.

#### **2. Materials and Methods**

A surveillance program was conducted in the semi-closed municipality of Deskati in January 2022. To assess the different scenarios for antibody production, antibody titers were obtained from participants while recording their infection and/or vaccination history since the pandemic wave initiation in the community in October 2020.

All the residents of Deskati were invited to participate in this program by the local authority and were notified of the time and place. Participants were recruited by announcing the research in the media, while local officials organized a one-month recruitment campaign. There were no exclusion criteria. The participants were analyzed to evaluate seroprevalence and antibody-response longevity to the SARS-CoV-2 infection and/or vaccination.

All subjects provided written and oral informed consent. Following consent, demographic information and data regarding past PCR-confirmed COVID-19 infection and vaccination history were recorded on questionnaire forms for all participants.

The SARS-CoV-2 IgG II Quant method (Architect, Abbott, IL, USA) was used for antibody testing. This is an automated two-step chemiluminescent microparticle immunoassay that was used for the qualitative and quantitative determination of IgG antibodies against the spike receptor-binding domain (RBD) of SARS-CoV-2 in the serum specimens, with a sensitivity of 99.9% and specificity of 100% for detecting the IgG antibodies generated by prior infection or vaccination, as previously described [4,5]. The sequence used for the receptor-binding domain was taken from the WH-Human 1 coronavirus, GenBank accession number MN908947. The analytical measurement interval is stated as 21 to 40,000 AU/mL, and the positivity cutoff as ≥50 AU/mL (manufacturer defined) [6].

The Pearson correlation method was used for correlation analysis between the pairs of continuous variables. Stepwise multiple linear analysis was conducted with numerical and categorical variables turned into dummy variables. It was used to analyze the correlation between antibody titers and various factors affecting the population. The mean age, gender, mean BMI, smoking status, presence of comorbidities, previous infection, hospitalization, mean length of hospitalization, re-infection, vaccination status, brand name of the vaccine, number of vaccination doses, and months after the last vaccine dose were used as independent variables in the prediction of antibody titers. To identify differences the between two independent groups, an unpaired *t*-test was used. Parametric data comparing three or more groups were analyzed with a one-way ANOVA and Tukey's multiple comparisons test, while non-parametric data were analyzed with the Kruskal–Wallis test and Dunn's multiple comparison test. Pearson's chi-squared test was used to determine whether there was a statistically significant difference between the frequencies. A result was considered statistically significant when the *p*-value was <0.05. Data were analyzed and visualized using SPSS Statistics v.23 (Armonk, NY, USA: IBM Corp.) and Tableau (Tableau Software LLC, Seattle, WA, USA), respectively.

#### **3. Results**

In this study, 145 participants were recruited. The main characteristics of the study population are presented in Table 1. As shown, females had more comorbidities than males. None of the participants were immunocompromised. Half of the population had previously been infected by SARS-CoV-2 for one year. A total amount of 8.1% of the infected population (*n* = 12) had a recent double infection (in the last three months) during that year, from which ten were fully vaccinated while two were not vaccinated at all.

Most of the population (93.1%, *n* = 135) were vaccinated. A total of 82.2% (*n* = 111) were fully vaccinated (with three doses), and the rest were partly vaccinated. A total of 70.3% (*n* = 95) of the population has been vaccinated with Pfizer/BioNTech, 29% (*n* = 39) by Moderna and 0.7% by Johnson & Johnson, with no difference between genders. We found

no difference in antibody titers between the different brands of vaccines (Pfizer/BioNTech vs. Moderna, 14,644 ± 11,567 vs. 10,793 ± 11,596, *p* = 0.084). There was no correlation between the months after the last vaccine dose and antibody titers (r = −0.027, *p* = 0.761).


**Table 1.** Characteristics of the study population stratified by gender (N = 145).

No difference in antibody production was observed among the genders after 27 months (*p* = 0.193). No correlation was found between BMI and antibody titers (r = 0.92, *p* = 0.293). No significant differences were detected in antibody titers by tobacco use (current and ex-smokers vs. nonsmokers, *p* = 0.522) and comorbidities (*p* = 0.073). Advanced age (>56 years old) was associated with higher antibody titers compared to younger adults (14,595 ± 12,869 AU/mL vs. 10,517 ± 9735 AU/mL, *p* = 0.039).

SARS-CoV-2 seropositivity was 93.1% in the study population. Seronegative (*n* = 10) were only infected but unvaccinated. More specifically, the infected and not vaccinated people had no seropositivity one year after the SARS-CoV-2 infection. The winners in antibody production were the patients who were fully or partly vaccinated and had also been infected twice (Table 2), followed by people who were fully vaccinated or partially vaccinated and were infected once, with no significant difference between the last two groups. People who were vaccinated had lower antibody levels, while individuals who had only been infected had the lowest antibody titers. A statistical analysis of the different immunization scenarios revealed a significant difference in antibody titers between the groups.


**Table 2.** Antibody response in different immunization scenarios during a year period in immunocompetent population (N = 145) and statistical analysis of different immunization scenarios.

Note: \* One-way ANOVA compares the means of the antibody titers of fully or partially vaccinated and double infected people with all the other independent immunization scenarios; \*\* One-way ANOVA compares the means of the antibody titers of fully or partially vaccinated and once infected people with all the other independent immunization scenarios; \*\*\* One-way ANOVA compares the means of the antibody titers of partially vaccinated and once infected people with all the other independent immunization scenarios; # One-way ANOVA compares the means of the antibody titers of only vaccinated people with all the other independent immunization scenarios.

Stepwise multiple linear analysis was used to analyze the correlation between antibody titers and the various factors affecting the population (Table 3). The mean age, gender, mean BMI, smoking status, presence of comorbidities, previous infection, hospitalization, length of hospitalization, re-infection, vaccination status, the brand name of the vaccine, number of vaccination doses, and months after the last vaccine dose were used as independent variables in the prediction of antibody titers. The hospitalization period and re-infection were independent predictor variables of antibody titers, explaining 52% of the total variance in this regression model. There was no multicollinearity between the explanatory variables.


**Table 3.** Stepwise multiple linear analysis between antibody titers and significant predictors.

<sup>a</sup> Dependent variable: antibody titers (AU/mL), R = 52.2%, R<sup>2</sup> = 27%, R<sup>2</sup> (adjusted) = 23%.

#### **4. Discussion**

In this study, for the first time, we investigated the different scenarios for antibody production among immunocompetent participants by recording their infection and/or vaccination history during a one-year period. In particular, we found that advanced age (>56 years old) was associated with higher antibody titers. No significant differences were detected in antibody titers among genders, sex, BMI, smoking status, comorbidities, vaccine brands, and months after the last shot. Hospitalization periods and re-infection were independent predictor variables of antibody titers. Individuals who were fully or partially vaccinated and were also double infected had the highest antibody levels, followed by people who were fully vaccinated or/partially vaccinated and were infected once. People who were only vaccinated had lower levels of antibodies, while the lowest levels among all the groups were found in individuals who had only been infected. A significant difference was detected between all the groups.

Interestingly, SARS-CoV-2 seropositivity was 93% in the study population. Seronegative people were only infected with the virus and remained unvaccinated. The longevity of the antibody response to the SARS-CoV-2 infection are not well defined. We have recently reported that antibody responses to the SARS-CoV-2 infection were maintained nine months after the pandemic and especially in those with severe disease leading to hospitalization [4,5]. A recent study identified the over one year duration of SARS-CoV-2 antibodies in 82.90% of 538 convalescent COVID-19 patients [7,8]. Similarly, other studies supported a long-lasting immunological memory against SARS-CoV-2 one year after mild COVID-19 [9,10]. Conversely, in this study all the people who were infected (*n* = 11) but not vaccinated had no seropositivity after one year [9]. One large study showed that 13% of individuals lost detectable IgG titers 10 months post-infection. Yan et al. documented that SARS-CoV-2-specific IgG persistence and titer depended on COVID-19 severity, as 74.4% of recovered asymptomatic carriers had negative anti-SARS-CoV-2 IgG test results, while many others had very low virus-specific IgG antibody titers, among a population of 473 previously infected patients [11]. Hence, further studies are needed to clarify this field.

Multiple vaccine constructs have been quite promising, with an approximately 95% protective efficacy against COVID-19 [12]. Since identifying the Omicron variant, many countries have made modifications to their vaccination programs by including the recommendation of a third and fourth injection of boosting vaccination dosages in large populations to reduce the risk of adverse effects. However, all three vaccine producers (Johnson et Johnson, BioNTech, Pfizer, and Moderna) have published statements claiming vaccines

would protect against severe sickness, as well as the fact that variant-specific vaccinations and boosters are in the works [13]. Nevertheless, it is unknown how long the immunity following COVID-19 vaccinations last, and this is a quite provoking situation that leads to an increased feeling of uncertainty and disbelief. It has been supported that the antibody persistence time of the mRNA vaccine is about 180 days (six months), following the adenovirus vaccine with 90 days [14]. Moreover, a short duration of antibody persistence of about 2-months has been found after the second dose [14].

Roy et al. reported significant differences in neutralizing antibody titers after 180 days in age, sex, COVID-19 infection, tobacco use, and asthma patients [15]. Swartz et al. measured antibody titers in 4553 participants over 11 months and documented that individuals may remain antibody positive from natural infection beyond 500 days, depending on age and smoking or vaping use [15]. Conversely, in our study, no significant differences were detected in antibody titers by sex, BMI, smoking status, and comorbidities. No significant differences were also detected in antibody titers among different brands of vaccines and months after the last shot. There are supporting data that prove there is a significantly higher humoral immunogenicity of the SARS-CoV-2 mRNA-1273 vaccine (Moderna) compared with the BNT162b2 vaccine (Pfizer-BioNTech) in infected as well as uninfected participants and across age categories [16].

An interesting finding was that advanced age (>56 years old) was associated with higher antibody titers. However, age was not a predictor of antibody titers in the stepwise multiple linear analysis. Yang et al. investigated the antibody test results among 31,426 patients from a wide range of age groups and supported an age-dependent variation in antibody titers, with children having higher antibody-binding avidity compared with young adults, but the difference was not significant [17]. However, contradictory data also exist [18,19]. Although there is an expectation that COVID-19 will become endemic, the pandemic will not end with the virus disappearing, and many questions remain unanswered. Further studies are needed to clarify the arising issues.

Moreover, in this study, the hospitalization period and re-infection were independent predictor variables of antibody titers. Similarly, Klein et al. found that hospitalization for severe COVID-19 could predict greater antibody responses against SARS-CoV-2. [18]. At present, it is unclear how long serum antibodies persist after reinfection [20]. Townsend et al. supported the fact that reinfection by SARS-CoV-2 under endemic conditions would likely occur between 3months and 5.1 years after the peak antibody response, with a median of 16 months [20]. However, to identify the correlates of protection, the relationship between in vitro neutralization levels of anti-SARS-CoV-2 antibodies and protection from severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection by large convalescent cohorts should be tested. Khoury et al. documented that neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection and estimated the neutralization level for 50% protection against detectable SARS-CoV-2 infection to be 20.2% of the mean convalescent level, predicting that over the first 250 days after immunization a significant loss in protection from SARS-CoV-2 infection will occur, although protection from severe disease should be largely retained [21]. However, how high a titer is protective of further infection remains unclear, and we cannot provide conclusive evidence that these antibody responses protect from reinfection. However, we believe it is very likely that higher titers will decrease the odds ratio of reinfection and may attenuate disease in the case of breakthrough infection. Undoubtedly, it is imperative to swiftly perform studies to investigate and establish a correlate of protection from SARS-CoV-2 infection.

Higher antibody titers were found in cases of vaccination in previously infected subjects, according to a previous study by our scientific team [4,5]. This finding is also supported by many other studies which report that a low concentration of SARS-CoV-2 spike protein antibodies after 9–12 months indicates that re-exposure to the virus or vaccination is required to use the B-cell immunity to full capacity [22]. In the current study, we found a 15 times higher titer of anti-SARS-CoV-2 antibodies in individuals who were fully or

partially vaccinated and who were double infected than previously infected patients, while fully vaccinated patients who were infected once had 25 times higher titers of anti-SARS-CoV-2 antibodies than previously infected patients. Interestingly, patients who were only vaccinated had nine times greater antibody titers than the only-infected patients. These results reflect those of Teresa Vietri et al., who also found that a booster dose resulted in a marked increase in antibody response, which then subsequently decreased over time [23].

Our study's findings should be interpreted within the context of its limitations and strengths. As such, when considering absolute numbers, our study's population is smaller compared to other studies [20,21]. It does, however, represent a specific epidemiological framework in rural Greece, reflecting remote populations differentially affected by the pandemic. Within these parameters, albeit nested, our study reports on real-world data representative of the geographical, cultural, and healthcare settings from which they stem. Furthermore, as previously mentioned, the corroboration of our findings in larger cohorts reflects that these data may be generalizable in similar settings. Another limitation was that the sample group was rather uniformly young and very lean, which limits the generalizability of our findings. It would be appealing to apply this search among individuals of a more significant number and in different geographical areas. This could give us the unique opportunity to evaluate and more profoundly assess the potential fluctuation between the titers of antibodies against SARS-CoV-2 in different scenarios in terms of antibody production and environmental conditions. In addition, it would be intriguing to study the titers of antibodies against SARS-CoV-2 in different statuses concerning the production of antibodies in various eating habits and lifestyles.

Last but not least, it would be particularly thrilling if we could develop a score or index concerning the antibody titers, the viral infection's different statuses related to the antibodies produced, and the patient's clinical image in order to have a potential prognostic tool, especially in individuals living with many comorbidities. Using scores or indices such as these, it might be possible to detect early the need for further and more specialized medical intervention and care in subjects infected with SARS-CoV-2 and vaccinated, fully or not, against the virus which has invaded our everyday routine. This could probably be beneficial not only for the infected individuals and their families but also for the healthcare system that has sustained a tremendous burden, both economically concerning every country worldwide and psychologically, especially for healthcare workers, due to the pandemic. We seem to have a long road to cross for understanding and decoding the mechanisms concerning this viral infection and its effect on human body systems.

#### **5. Conclusions**

The winners in anti-SARS-CoV-2 titers were the individuals who were fully or partially vaccinated and who were also double infected, followed by people who were fully vaccinated or/partially vaccinated and only infected once. In addition, subjects who were only vaccinated had lower levels of antibodies, whilst the lowest levels among all the groups were found in individuals who had only been infected. Overall, these results are quite promising, but the SARS-CoV-2 variant seems to be the dragon in this medical issue.

Our findings indicate that every hit (infection or vaccination) gives an additional boost in immunization status. However, the antibody response raised by vaccines is roughly affected by not only the time but also the emergence of new SARS-CoV-2 variants. The spread of new variants is associated with an escape from antibodies; therefore, to mitigate the spread of this infection in the long run, a more effective longitudinal observation of the immune response is needed.

**Author Contributions:** Conceptualization, K.I.G., O.S.K., D.P. and E.C.F.; methodology, K.I.G., O.S.K., D.P. and E.C.F.; formal analysis, O.S.K.; investigation, O.S.K., D.P., E.G., E.C.F., G.P., A.M., D.I.S., A.K., G.D.V., E.M., D.A., O.S.K., N.K. and G.K.; resources, D.P. and K.I.G.; writing—original draft preparation, O.S.K., G.D.V.; writing—review and editing, O.S.K.; supervision, O.S.K.; visualization, K.I.G.; project administration, K.I.G., O.S.K., D.P. and E.C.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of the Ethics Committee of the University of Thessaly (No. 2800-01/11/2020; approved on 1 November 2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data that support the findings of this study are available on request from the corresponding author, O.S.K.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


### *Systematic Review* **Therapeutic Vitamin D Supplementation Following COVID-19 Diagnosis: Where Do We Stand?—A Systematic Review**

**Angelina Bania <sup>1</sup> , Konstantinos Pitsikakis <sup>2</sup> , Georgios Mavrovounis 3,\*, Maria Mermiri <sup>4</sup> , Eleftherios T. Beltsios <sup>3</sup> , Antonis Adamou <sup>5</sup> , Vasiliki Konstantaki <sup>6</sup> , Demosthenes Makris <sup>7</sup> , Vasiliki Tsolaki <sup>7</sup> , Konstantinos Gourgoulianis <sup>8</sup> and Ioannis Pantazopoulos <sup>3</sup>**


**Abstract:** Vitamin D has known immunomodulatory activity and multiple indications exist supporting its potential use against SARS-CoV-2 infection in the setting of the current pandemic. The purpose of this systematic review is to examine the efficacy of vitamin D administered to adult patients following COVID-19 diagnosis in terms of length of hospital stay, intubation, ICU admission and mortality rates. Therefore, PubMed and Scopus databases were searched for original articles referring to the aforementioned parameters. Of the 1376 identified studies, eleven were finally included. Vitamin D supplements, and especially calcifediol, were shown to be useful in significantly reducing ICU admissions and/or mortality in four of the studies, but not in diminishing the duration of hospitalization of COVID-19 patients. Due to the large variation in vitamin D supplementation schemes no absolute conclusions can be drawn until larger randomized controlled trials are completed. However, calcifediol administered to COVID-19 patients upon diagnosis represents by far the most promising agent and should be the focus of upcoming research efforts.

**Keywords:** COVID-19; vitamin D; hospitalization; ICU admission; intubation; mortality

### **1. Introduction**

The ongoing COVID-19 pandemic proven a major challenge both for the scientific community and society in general, resulting in millions of deaths worldwide [1]. Despite the immunization of a large percentage of the world population [1], predominantly in firstworld countries, SARS-CoV-2 and its variants remain a significant cause of morbidity and mortality. In the absence of SARS-CoV-2-specific pharmacological agents, drug repurposing has emerged as the only available treatment strategy. Remdesivir plus dexamethasone, immunomodulatory agents and, more recently, monoclonal antibodies are approved under Emergency Use Authorization for various severity stages of COVID-19 [2], but efforts for more, largely available and safe drugs are continuous.

Vitamin D is a fat-soluble vitamin, regulating circulating calcium and phosphate levels with an important role in bone homeostasis. The active form of vitamin D is 1,25(OH)2D3

**Citation:** Bania, A.; Pitsikakis, K.; Mavrovounis, G.; Mermiri, M.; Beltsios, E.T.; Adamou, A.; Konstantaki, V.; Makris, D.; Tsolaki, V.; Gourgoulianis, K.; et al. Therapeutic Vitamin D Supplementation Following COVID-19 Diagnosis: Where Do We Stand?—A Systematic Review. *J. Pers. Med.* **2022**, *12*, 419. https://doi.org/ 10.3390/jpm12030419

Academic Editor: Bruno Mégarbané

Received: 18 February 2022 Accepted: 5 March 2022 Published: 8 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(calcitriol) and its biosynthesis includes the conversion of skin 7-dehydrocholesterol to pre-vitamin D3 and then vitamin D3 (cholecalciferol) in the presence of ultraviolet sun radiation [3,4], followed by two steps of hydroxylation to 25(OH)D3 (calcifediol) by the liver and finally to 1,25(OH)2D3 by the kidney. The vitamin D receptor (VDR) acts as a transcription factor and alongside the retinoid X receptor (RXR) binds on a DNA motif on a variety of human tissues [5], regulating the epigenome and expression of thousands of genes and gene networks [6], involved in mineral, bile acid and exogenous compound metabolism, cell differentiation and immune response [7].

Vitamin D deficiency, defined by the Endocrine Society [8] as 25(OH)D3 below 20 ng/mL and vitamin D insufficiency, defined as 25(OH)D3 of 21–29 ng/mL are highly prevalent findings among the general population, linked to rickets in children and osteomalacia and osteoporosis in adults, as well as diabetes, cardiovascular disease, auto-immune disorders, cancer, hepatitis B and C, allergies, asthma and respiratory tract infections [4,9].

In the current setting, vitamin D has been shown to exert immunomodulatory actions in SARS-CoV-2 infection [10,11]. More specifically, it increases the expression of defensins and cathelicidin (LL-37), an endogenous antimicrobial [12], as well as other antiviral agents involved in the TNF-a [13], IFN-γ [14] and NF-κB [15] pathways. It also reduces inflammation, and thus the risk to develop the potentially fatal Cytokine Storm Syndrome, by inhibiting the Th1 response and the production of inflammatory cytokines [14], while enhancing the production of anti-inflammatory cytokines [14]. Its role as a potential immunomodulatory agent is further supported by its capacity to increase regulatory T lymphocytes [16], which are significantly decreased in the setting of COVID-19 [17].

Vitamin D has been hypothesized to intervene in the mechanism by which COVID-19 infection induces a hypercoagulative state [18,19], thus increasing the risk for thrombosis, as well as results to the Acute Respiratory Distress Syndrome (ARDS). It is known that SARS-CoV-2 utilizes the angiotensin converting enzyme 2 (ACE2) receptor [20], thus downregulating it. This results in excessive accumulation of angiotensin II, the substrate of ACE2, which can lead to ARDS [21]. Serum vitamin D levels have been found to be inversely correlated with the Renin-Angiotensin-Aldosterone System activation [22,23], meaning that in COVID-19 patients with vitamin D deficiency, the increase of angiotensin may facilitate progress to ARDS. Conversely, vitamin D can protect from ARDS by lowering renin and increasing ACE2 expression [24].

Based on these data and thanks to their safety profile, availability and low cost, vitamin D supplements are currently used as an off-label pharmacological agent for the treatment of SARS-CoV-2 infection, while their efficacy has been examined in multiple studies with varying results. In this systematic review, we aim to summarize the most recent evidence regarding the therapeutic role of vitamin D on severe COVID-19 outcomes (length of hospital stay, mechanical ventilation, mortality) in adult populations.

#### **2. Materials and Methods**

#### *2.1. Protocol*

The protocol for this systematic review is registered in the International Prospective Registry of Systematic Reviews, PROSPERO, under the ID: PROSPERO2021 CRD42021281646 and is fully available online at https://www.crd.york.ac.uk/prospero/display\_record.php? ID=CRD42021281646 (Last accessed: 18 February 2022; 19:01:38 EET).

#### *2.2. Literature Search*

Two investigators (A.B. and K.P.) individually performed an electronic search of the PubMed (MEDLINE) and Scopus databases to identify relevant studies, based on the predetermined inclusion and exclusion criteria. Disagreements between the two authors were resolved by discussion between them or with the help of a third investigator (G.M.), when necessary. The search algorithms, fully available in the Supplementary File Document S1, consisted of the terms 'vitamin D' and 'COVID-19' and their derivatives, as well as the Boolean operators 'AND' and 'OR'. The references of previous systematic reviews

and meta-analyses were also screened for additional original studies. Only articles fully available in the English language were included in this review. The last literature search was performed on 28 September 2021.

#### *2.3. Inclusion and Exclusion Criteria*

Original articles, restricted to randomized controlled trials, prospective and retrospective observational studies, case-control studies and case series with at least ten participants were included in this systematic review. No restriction on publication date was imposed. The studies pertained to the post-diagnosis administration of any form of vitamin D to adult (>18 years of age) patients diagnosed with COVID-19. Our studied outcomes were: duration of hospitalization, need for mechanical ventilation/intubation, ICU admission and all-cause mortality.

Studies on the chronic supplementation with vitamin D and studies focusing on paediatric populations were excluded from this systematic review. Congress abstracts, letters to the editor, case reports, case series of less than ten patients, ecological studies, reviews and meta-analyses were also excluded.

Jevalikar et al. [25] included a small number of children in their cohort. Here we only report findings based on the data relevant to vitamin D administration in a sub-population of the initial cohort, but the presence of children in this sub-group is not specified.

#### *2.4. Data Extraction*

Using a pre-determined data table, two of the authors (A.B. and K.P.) performed the data extraction. The following data were extracted: First Author's Name, Month and Year of Publication, Study Design, Vitamin D Administration Scheme, Control Method, Population Size and Number of Participants in each group, Male to Female Ratio, Mean Age, Presence Of Comorbidities (Hypertension, Cancer, Myocardial Infarction, Diabetes Mellitus, Chronic Obstructive Pulmonary Disease, Chronic Kidney Disease, Obesity), Baseline And Post-Intervention Serum Vitamin D Levels in each group, Mortality, Length Of Hospital Stay, ICU Admissions and Intubation events in each group, Mortality Time Point and Length of Follow-Up.

#### *2.5. Quality Assesment*

Quality scoring was performed using the Cochrane Risk of Bias (RoB) [26] tool for the Randomized Controlled Trials and the Methodological Index For Non-Randomized Studies (MINORS) [27] for the observational studies. The RoB tool calculates the risk of bias accounting for the randomization process, the deviations from the intended interventions, potential missing data, the outcome measuring methods and the selection of the reported result. The MINORS tool evaluates twelve factors, relevant to the aim and design of the study, patient selection and grouping, follow-up, potential size calculation and result assessment and analysis on a scale of 0–24.

#### **3. Results**

#### *3.1. Search Results*

The literature search yielded a total of 1376 articles (832 on PubMed and 544 on Scopus), among which 189 were selected to be evaluated as full texts. Finally, a total of 11 [25,28–37] articles fully met our inclusion criteria and were included in this systematic review (Figure 1).

**Figure 1.** PRISMA flow diagram. **Figure 1.** PRISMA flow diagram.

#### *3.2. Study Characteristics 3.2. Study Characteristics*

The studies were published from October 2020 to September 2021 and were conducted in four different continents. Three studies took place in Spain [31,33,34] and one in each of the following countries: France [28], USA [32], Brazil [35] Turkey [30], Singapore [36], Saudi Arabia [29], India [25] and Egypt [37]. The majority [25,28,30,31,33,36,37] were single-center, while in four studies patients from two [35], three [29,32] or five [34] centers were recruited. Our study collection includes four randomized controlled trials [29,32,33,35], one non-randomized controlled trial [28] and six observational cohort stud-The studies were published from October 2020 to September 2021 and were conducted in four different continents. Three studies took place in Spain [31,33,34] and one in each of the following countries: France [28], USA [32], Brazil [35] Turkey [30], Singapore [36], Saudi Arabia [29], India [25] and Egypt [37]. The majority [25,28,30,31,33,36,37] were single-center, while in four studies patients from two [35], three [29,32] or five [34] centers were recruited. Our study collection includes four randomized controlled trials [29,32,33,35], one nonrandomized controlled trial [28] and six observational cohort studies [25,30,31,34,36,37], among which two [25,31] were reported as prospective and three [34,36,37] as retrospective.

ies [25,30,31,34,36,37], among which two [25,31] were reported as prospective and three [34,36,37] as retrospective. All patients examined in the aforementioned studies were hospitalized for COVID-19 infection. Vitamin D deficiency was not always a prerequisite for inclusion in the studies. A few studies focused on specific subpopulations of COVID-19 patients. More specifically, Güven et al. [30] reported only on vitamin D deficient (25(OH)D3 < 12 ng/mL) patients who had already been admitted to the ICU. In terms of age and comorbidities, Nogues et al. [31] studied high risk patients, i.e., with severe COVID-19 and/or comorbidities, Tan et al. [36] included only patients of age 50 or older, Soliman et al. [37] selected All patients examined in the aforementioned studies were hospitalized for COVID-19 infection. Vitamin D deficiency was not always a prerequisite for inclusion in the studies. A few studies focused on specific subpopulations of COVID-19 patients. More specifically, Güven et al. [30] reported only on vitamin D deficient (25(OH)D3 < 12 ng/mL) patients who had already been admitted to the ICU. In terms of age and comorbidities, Nogues et al. [31] studied high risk patients, i.e., with severe COVID-19 and/or comorbidities, Tan et al. [36] included only patients of age 50 or older, Soliman et al. [37] selected elderly (>60 years of age) vitamin D deficient (<20 ng/mL) Type II diabetics, while Annweiler et al. [28] focused on frail elderly inpatients at a geriatric acute care unit.

elderly (>60 years of age) vitamin D deficient (<20 ng/mL) Type II diabetics, while The studies and their characteristics are presented in Table 1.

The studies and their characteristics are presented in Table 1.

Annweiler et al. [28] focused on frail elderly inpatients at a geriatric acute care unit.


#### **Table 1.** Study Characteristics.

IU: International Units, IM: intramuscular, d: day, mg: milligrams, µg: micrograms, ng/mL: nanograms per milliliter, IQR: Interquartile Range, NA: not available. \* Originally given at nmol/L, but converted here to ng/mL for consistency.

#### *3.3. Interventions*

The administered substance, its administration route and dosing scheme varied significantly among studies. Seven of them investigated the effect of vitamin D<sup>3</sup> (cholecalciferol), administered daily per os [29,36], as a high single oral dose of 60,000 IU [25], 80,000 IU [28] or 200,000 IU [35] or as an intramuscular injection of 200,000 [37] or 300,000 IU [30]. Three studies [31,33,34] applied various regimens of oral 25(OH)D3 (calcifediol) and one trial [32] used oral 1,25(OH)2D<sup>3</sup> (calcitriol) at 0.5µg/day for 14 days.

With two exceptions [29,36], the intervention and control groups did not receive any further treatments other than the appropriate standard of care of their centers or placebo. However, in a retrospective study by Tan et al. [36], the intervention group received 1,000 IU/day oral vitamin D3, 150 mg/day oral magnesium and 500 mcg/day oral vitamin B12 for a median interval of 5 days. Finally, in a randomized controlled trial by Sabico et al. [29] both groups received oral vitamin D3, but at different doses (5000 IU vs. 1000 IU).

No severe adverse effects related to this treatment were observed in any of the studies.

#### *3.4. Length of Hospital Stay*

Out of 10 studies of vitamin D-supplemented vs. vitamin D-non-supplemented patients, three reported on the length of hospitalization. Neither a single dose of 300,000 IU of intramuscular vitamin D<sup>3</sup> [30], a single dose of 200,000 IU of oral D3 [35] or a 14-day regimen of 0.5 µg 1,25(OH)2D3 per day [32] managed to affect the duration of hospital stay for the intervention group [9(6–16) vs. 9(5–17), *p*-value = 0.649, 7.0(4.0–10.0) vs. 7.0(5.0– 13.0) days, *p*-value = 0.59, 5.5 ± 3.9 vs. 9.24 ± 9.4 days, *p*-value = 0.14 respectively]. Additionally, no difference in hospitalization duration was observed between the 5000 IU and the 1000 IU oral D3 group in the randomized controlled trial by Sabico et al. [29] [6(5–8) vs. 7(0–10), *p*-value = 0.14] (Table 2).


**Table 2.** Patient Outcomes.

ICU: Intensive Care Unit, IQR: Interquartile Range, NA: not available.

#### *3.5. Need for Intubation and ICU Admission*

Ten out of eleven studies provided data regarding either the need for intubation and mechanical ventilation (three studies) or intensive care admission (four studies) or both (two studies). Entrenas-Castillo et al. [33], explored the effect of a regimen comprised of 0.532 mg oral 25(OH)D3 on the day of admission followed by 0.266 mg on the 3rd and 7th

day and then weekly until discharge or ICU admission in a randomized controlled trial. Of 50 patients in the intervention arm, only one required ICU admission, in contrast to the 13/26 patients from the control group (*p*-value < 0.001).

A similar dosing scheme (0.532 mg oral 25(OH)D3 on day 1 plus 0.266 mg on days 3, 7, 15, and 30) was later investigated by Nogues et al. [31], in a large prospective study of 838 high-risk COVID-19 patients. ICU admission was necessary for 21% of the patients in the control group, compared to 4.5% in the intervention group (OR = 0.18 (0.11–0.29), *p*-value < 0.001), showing an 87% risk reduction following adjustment for age, gender, baseline vitamin D levels and comorbidities [OR = 0.13, (0.07–0.23), *p*-value < 0.001]. A statistically significant difference in vitamin D levels between ICU and non-ICU patients was also noted [10 (7–14) ng/mL vs. 13 (8–23) ng/mL, *p*-value < 0.001].

Tan et al. [36], evaluated the combination of vitamin D, vitamin B12 and magnesium in a retrospective study of 43 patients over 50 years of age. The combination therapy was shown to significantly (*p*-value = 0.006) reduce the need for any form of oxygenation therapy. Specifically, 3/17 treated patients required oxygen therapy (including 1 in the ICU), compared to 16/26 non-treated ones (including 8 in the ICU). A subgroup analysis focusing on 30 non-diabetic patients aged 50–60 years was later performed and failed to show a statistically significant difference in oxygenation needs [25% vs. 58.3%, *p*-value = 0.197 and 12.5% vs. 41.7% in regard to ICU admission].

The remaining eight studies, among which the trial of 5000 IU vs. 1000 IU of oral vitamin D3 by Sabico et al. [29] did not show a statistically significant difference in ICU admission [25,29,32,35] and need for intubation [30,32,34,35,37] between study groups (Table 2).

#### *3.6. Mortality*

All eleven studies reported on the in-hospital mortality of COVID-19 patients. In a multi-center retrospective analysis of 537 patients by Alcala-Diaz et al. [34], 79 patients had received 0.532 mg of oral 25(OH)D3 on day 1 followed by 0.266 mg on day 3 and 7 and then weekly until hospital discharge or ICU admission. Mortality rates among these patients were significantly lower than those of the control group [5% versus. 20%, *p*-value < 0.001, OR = 0.22 (0.08–0.61), *p*-value < 0.01]. Given that all intervention group patients were from the same center and the patients in the control group had a greater comorbidity burden and worse clinical image upon admission, an analysis adjusted for age, center, CURB-65, ARDS at admission, neutrophil/lymphocytes ratio and comorbidities followed and still demonstrated the favorable position of the intervention group in terms of mortality [OR = 0.16 (95%CI = 0.03–0.80), *p*-value = 0.02]. The elderly (>65 years) subgroup with oxygen saturation <96% also greatly benefited from calcifediol administration [OR 0.06 (0.04–0.8), *p*-value = 0.04].

Nogues et al. [31] also attributed a reduction of death rates to 25(OH)D3 administration both in the initial [4.7% vs. 15.9%, OR: 0.26 (0.15–0.43), *p*-value < 0.001] and the adjusted analysis for age, gender, vitamin D levels and comorbidities [OR = 0.21; (95%CI, 0.10–0.43)], which translates into a 70% mortality risk reduction. Baseline vitamin D levels were greater in survivors compared to non-survivors [13 (8–22.7) ng/mL vs. 9 (6–13.5) ng/mL, *p*-value < 0.001].

In this study, 53 of 82 patients from the control group who required intensive care were started on the 25(OH)D3 regimen upon ICU admission. A sub-analysis of a total of 102 ICU COVID-19 patients was then performed. Interestingly in these patients, administration of vitamin D upon initial hospital admission was associated with lower mortality than initiation of supplementation upon ICU admission, while never receiving vitamin D at any point of the disease course had the worst prognosis. However, these differences in mortality were considered statistically insignificant (10.0% vs. 28.3% vs. 31% respectively).

No other study observed significantly different death rates among study groups (Table 2).

#### *3.7. Quality Assessment and Risk of Bias* The quality of the comparative observational studies ranged from 17 to 22 out of 24,

*3.7. Quality Assessment and Risk of Bias* 

respectively).

ble 2).

*J. Pers. Med.* **2022**, *12*, x FOR PEER REVIEW 8 of 14

The bias risk for the randomized controlled trials varied significantly, as seen in Figure 2. One study [35] is marked as low-risk, one [29] as moderate risk and two [32,33] as high risk, with concerns arising mainly form the randomization process bias. based on the MINORS tool, which translates into moderate and high quality in five [28,30,34,36,37] and two [25,31] studies respectively. MINORS scores for each individual study are reported in Table 3.

high risk, with concerns arising mainly form the randomization process bias.

administration of vitamin D upon initial hospital admission was associated with lower mortality than initiation of supplementation upon ICU admission, while never receiving vitamin D at any point of the disease course had the worst prognosis. However, these differences in mortality were considered statistically insignificant (10.0% vs. 28.3% vs. 31%

No other study observed significantly different death rates among study groups (Ta-

The bias risk for the randomized controlled trials varied significantly, as seen in Figure 2. One study [35] is marked as low-risk, one [29] as moderate risk and two [32,33] as

**Figure 2.** Risk of Bias of Randomized Controlled Trials. **Figure 2.** Risk of Bias of Randomized Controlled Trials.

**Table 3.** MINORS Score for non-randomized trials. **Author MINORS Score (Out of 24)**  Annweiler 18 The quality of the comparative observational studies ranged from 17 to 22 out of 24, based on the MINORS tool, which translates into moderate and high quality in five [28,30,34,36,37] and two [25,31] studies respectively. MINORS scores for each individual study are reported in Table 3.

Guven 18

Nogues 19 **Table 3.** MINORS Score for non-randomized trials.


#### peutic administration of vitamin D in hospitalized patients following COVID-19 diagnosis **4. Discussion**

from chronic vitamin D supplementation for unrelated purposes. The aim of this systematic review was to explore the impact of vitamin D administration on important parameters of COVID-19 disease course, such as length of hospital stay, ICU admissions and mortality. Of the four studies mentioning vitamin D and hospitalization duration, none managed to prove an association. Moreover, the majority of studies To our knowledge, this is the largest and most updated systematic review focusing exclusively on post-COVID-19 diagnosis administration of vitamin D, having included more recent articles compared to previous work. Thus, we have distinguished the therapeutic administration of vitamin D in hospitalized patients following COVID-19 diagnosis from chronic vitamin D supplementation for unrelated purposes.

did not observe significant differences in the need for intubation, ICU admission or mortality, since only four out of eleven studies finally support vitamin D administration to prevent one or multiple among these unfavorable outcomes. The aim of this systematic review was to explore the impact of vitamin D administration on important parameters of COVID-19 disease course, such as length of hospital stay, ICU admissions and mortality. Of the four studies mentioning vitamin D and hospitalization duration, none managed to prove an association. Moreover, the majority of studies did not observe significant differences in the need for intubation, ICU admission or mortality, since only four out of eleven studies finally support vitamin D administration to prevent one or multiple among these unfavorable outcomes.

Evidence in favor of the use of vitamin D were identified in one randomized controlled trial of 76 patients [33], one large multi-center prospective study of 838 participants [31] and two retrospective studies [34,36] of 537 and 43 patients respectively. The observational studies, which lacked randomization, performed linear regression analyses adjusted for the confounders that were of statistical significance between the two groups and their results remained consistent with the initial findings.

It is possible that the active substance used in each study could have determined its results. Administration of vitamin D3 or 1,25(OH)2D3 alone in any form or dose failed to improve any of the outcomes. On the contrary, three out of eleven studies used 25(OH)D3 and all of them reached statistical significance regarding ICU admission, mortality or both. They all took place in Spain and employed a very similar intervention scheme, comprised of an initial oral dose of 0.532 mg 25(OH)D3 followed by 0.266 mg on days 3, 7, 15 and then weekly [33,34] or on days 3, 7, 15 and 30 in the case of Nogues et al. [31]. The fourth study [36] supporting the use of vitamin D supplements to reduce oxygenation and ICU admission used a triple combination of 1000 IU/day oral vitamin D3, 150 mg/day oral magnesium and 500 mcg/day oral vitamin B12 for a median duration of 5 days.

The active vitamin D substance chosen for administration might be of special importance in the setting of renal or liver disease. As expected from the fact that the activation of vitamin D takes place in these tissues, there is a high prevalence for vitamin D deficiency among patients with renal and liver disease [38,39]. Future studies should thus consider administering the fully activated 1,25(OH)2D3 to these subgroups or even 25(OH)D3 in the case of liver failure, to bypass the possibly inadequate intrinsic hydroxylation stages.

A general micronutrient sufficiency was shown to reduce SARS-CoV-2 infection and severe illness in a large meta-analysis [40]. Although most heated discussions revolve around vitamin D, other dietary supplements have also been administrated by clinicians in an off-label basis, thanks to their broad role in immune system function and minimal adverse effect burden. Vitamin C [41] and zinc [42] offered no benefit regarding disease outcomes. Vitamin B12, which was co-administered with vitamin D and magnesium in one of our included studies, might facilitate symptom alleviation in COVID-19 [43]. Curcumin, on the other hand, seems to be a more promising agent, associated with faster recovery and lower mortality in a systematic review of six trials [44].

Since the beginning of the COVID-19 pandemic, the use of vitamin D as a prognostic marker and a therapeutic agent has been debatable. This hypothesis was based on pre-existing knowledge from studies on its association with other respiratory tract infections, summarized in recent systematic reviews and meta-analyses, where vitamin D deficiency was found to increase susceptibility to infection [45], while vitamin D seemed to prevent [46,47] or improve [46] the disease course. Similar to our systematic review, a major source of concern on the reliability of these conclusions is the highly variable form of vitamin D analog, its dose and route of administration employed in each of the analyzed studies.

As far as COVID-19 is concerned, vitamin D status is regularly proven to attain a prognostic value in large recent meta-analyses. Lower vitamin D levels are measured in COVID-19 patients than in healthy individuals [42,48], indicating a possible link with susceptibility to infection. Indeed, vitamin D deficiency increased the odds of contracting SARS-CoV-2 by 80% [49]. When it comes to outcomes, a lower vitamin D status was observed in severe disease cases [48], while deficient COVID-19 patients were at an increased risk for prolonged hospitalization [50], ICU admission [51] and death [50,51], although its effect on mortality is quite debatable [48,52].

In the studies presented in this systematic review, no association was observed between baseline levels of vitamin D and the benefits of vitamin D administration. More specifically, two studies [30,37] recruited vitamin D deficient patients only, but no differences were observed between the intervention and control groups. Among the studies demonstrating significant improvements following vitamin D administration, Nogues et al. [31] was the only one providing data on baseline vitamin D levels and these were similar between groups. In any case, vitamin D status should be taken into consideration in the design of future trials.

When it comes to vitamin D as a treatment option, the effect on outcomes other than length of hospital stay, intubation and mortality have also been investigated. High (60,000–80,000 IU) total doses of vitamin D3 failed to reduce the incidence of severe COVID-19, defined as Ordinal Scale for Clinical Improvement (OSCI) score equal to or greater than 5 both in frail elderly [28] and vitamin-D-deficient patients [25]. Furthermore, the effect of vitamin D on inflammatory markers varied across studies. In the aforementioned

prospective study by Jevalikar et al. [25], no difference was observed in the fluctuation of any of the inflammatory markers (D-dimers, CRP, LDH, IL6, Ferritin) between the intervention and control groups. The same was reported by Sánchez-Zuno et al. [53], regarding vitamin D3 treated outpatients in regards to transferrin, ferritin and D-dimers. Among a small group of high-dose D3 supplemented and non-supplemented asymptomatic or mildly symptomatic patients the only significantly different decrease was observed in the fibrinogen levels [54]. On the contrary, a similar trial with mildly-moderately affected patients with vitamin D insufficiency reached statistical significance in all (N/L ratio, CRP, LDH, IL6, Ferritin) measured markers [55]. Vitamin D also facilitated symptom alleviation [53] and viral clearance [54] in one of two studies that reported these outcomes.

The role of vitamin D in the treatment plan against COVID-19 had been discussed in previous systematic reviews and meta-analyses. Our conclusions differ from a metaanalysis published by Pal et al. [56], who considered vitamin D supplementation to be beneficial with regards to COVID-19-related ICU admissions and mortality. This is the largest meta-analysis so far, including 13 studies, five of which are common with the ones presented in this systematic review. The difference in our conclusions may be attributed to the study selection. Pal et al. were able to include additional studies compared to our systematic review after contacting their respective authors for data which were not available in the original studies. However, studies associating COVID-19 outcomes with regular vitamin D supplementation, which were excluded in our methodology, were taken into consideration by Pal et al., who subsequently concluded that it is inferior to vitamin D administration after COVID-19 diagnosis. Finally, that meta-analysis does not take into consideration six of the eleven studies presented here, including the five most recent ones.

Previous systematic reviews and meta-analyses containing smaller subsets of articles have reached varying conclusions. Da Rocha et al. [57] were the first to publish a systematic review including three randomized controlled trials on November 2020. These three trials were also the basis for another systematic review by Stroehlein et al. [58] and a metaanalysis by Bassatne et al. [59]. The general conclusion was that vitamin D may have a therapeutic potential, but due to the insufficient, then available, evidence, the need for more, higher quality trials was highlighted.

Other systematic reviews have used subsets of the aforementioned studies and have reached conflicting conclusions supporting or disregarding the therapeutic value of vitamin D in COVID-19. An early meta-analysis by Shah et al. [60] observed the potential of vitamin D to reduce ICU admissions only. A meta-analysis of five studies [41] reported no statistically significant improvements in acute inflammatory markers, ventilation/ICU needs and mortality among patients receiving a variety of different supplementation regimens. This totally contradicts the conclusions of Dramé et al. [61] and Petrelli et al. [62], who express themselves in favor of vitamin D administration to improve all major outcomes. The co-presence of both regular supplementation regimens and post-diagnosis administration as interventions in the included studies is common among many of the systematic reviews and meta-analyses. The common denominator among all these, some of which date back to the very beginning of the pandemic, is the call for large randomized controlled trials. Indeed, the inconsistencies in population selection and more importantly in vitamin D form, dosage and route of administration among the existing studies prevents the extraction of definite conclusions, even two years into the pandemic. Therefore, as we highlight again the necessity for further research, we distinguish calcifediol from all other agents, identifying it as the most promising to be evaluated in upcoming trials.

#### **5. Limitations**

It has to be noted that, with one exception [35], the clinical trials presented in this review recruited a relatively small number of participants (<100, usually around 50) and this might be a reason for their failure to reach statistical significance.

This systematic review focuses only on the administration of vitamin D following COVID-19 diagnosis to improve important outcomes, such as length of hospital stay, intubation and ICU admission and mortality. Studies discussing the effect of vitamin D as a pre-existing regular supplementation were excluded.

It was also noticed that the included studies employed a highly variable intervention scheme, which consisted of different forms, doses and administration routes of vitamin D which, on one occasion, was co-administered with other agents. It was therefore hypothesized that it could affect the study results making data unsuitable to be pooled or processed in a meta-analysis. Indeed, the form of vitamin D analog seemed to affect outcomes, with 25(OH)D3 being associated with lower ICU admission and mortality, as opposed to vitamin D3 and 1,25(OH)2D3.

As about 80% of vitamin D reserves are derived from its biosynthesis in the skin, differences in exposure to UV radiation could influence the results of the included studies. Finally, cases of liver and kidney disease in the studied cohorts might underlie the lack of response to non-activated vitamin D compounds.

#### **6. Conclusions**

In this systematic review we have summarized existing knowledge regarding the role of vitamin D on important COVID-19 outcomes indicative of disease severity (length of hospital stay, ICU admission, mortality). Despite the conflicting evidence surrounding the effect of vitamin D across the reviewed studies, we observed 25(OH)D3 (calcifediol) to be by far the most successful agent in reducing intensive care needs and mortality. Therefore, given the insufficient level of evidence of these studies, we are looking forward to larger randomized controlled trials to evaluate calcifediol's role as an adjuvant to the existing treatment regimens. Finally, given that different SARS-CoV-2 variants are currently spreading worldwide, it could be interesting and useful for further studies to include data on the effect of vitamin D on different variants as well as the patients' viral load.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jpm12030419/s1, Document S1: Literature search algorithms.

**Author Contributions:** Conceptualization, A.B., G.M., D.M., V.T., K.G. and I.P.; methodology, A.B., G.M., E.T.B., A.A., D.M., V.T., K.G. and I.P.; investigation, A.B., K.P., G.M., M.M. and V.K.; data curation, A.B., K.P., G.M., M.M. and I.P.; writing—original draft preparation, A.B., K.P., G.M., M.M., E.T.B., A.A., V.K., D.M., V.T., K.G. and I.P.; writing—review and editing, A.B., K.P., G.M., M.M., E.T.B., A.A., V.K., D.M., V.T., K.G. and I.P.; visualization, A.B., K.P. and G.M.; supervision, G.M. and I.P.; project administration, I.P.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data are available upon request.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Prevalence of Hemorrhagic Complications in Hospitalized Patients with Pulmonary Embolism**

**Nikolaos Pagkratis <sup>1</sup> , Miltiadis Matsagas <sup>2</sup> , Foteini Malli <sup>3</sup> , Konstantinos I. Gourgoulianis <sup>4</sup> and Ourania S. Kotsiou 3,\***


**Abstract:** Background: The prevalence of anticoagulant therapy-associated hemorrhagic complications in hospitalized patients with pulmonary embolism (PE) has been scarcely investigated. Aim: To evaluate the prevalence of hemorrhages in hospitalized PE patients. Methods: The Information System "ASKLIPIOS™ HOSPITAL" implemented in the Respiratory Medicine Department, University of Thessaly, was used to collect demographic, clinical and outcome data from January 2013 to April 2021. Results: 326 patients were included. Males outnumbered females. The population's mean age was 68.7 ± 17.0 years. The majority received low molecular weight heparin (LMWH). Only 5% received direct oral anticoagulants. 15% of the population were complicated with hemorrhage, of whom 18.4% experienced a major event. Major hemorrhages were fewer than minor (29.8% vs. 70.2%, *p* = 0.001). Nadroparin related to 83.3% of the major events. Hematuria was the most common hemorrhagic event. 22% of patients with major events received a transfusion, and 11% were admitted to intensive care unit (ICU). The events lasted for 3 ± 2 days. No death was recorded. Conclusions: 1/5 of the patients hospitalized for PE complicated with hemorrhage without a fatal outcome. The hemorrhages were mainly minor and lasted for 3 ± 2 days. Among LMWHs, nadroparin was related to a higher percentage of hemorrhages.

**Keywords:** pulmonary embolism; venous thromboembolism; bleeding complications; anticoagulant treatment; prediction of bleeding; in-hospital bleeding

### **1. Introduction**

Pulmonary embolism (PE) is defined as a blockage in the pulmonary artery and its branches. It is caused by detached blood clots that move through the large veins to the pulmonary arteries. Embolism is usually caused by blood clots in the deep network of veins of the lower limbs—mainly in their proximal parts—such as by blood clots in the pelvic network, the upper limbs, and the right part of the heart. Rarely PE is caused by nonthrombotic sources, such as amniotic fluid, tumors, fat, large amounts of air and foreign bodies. In every patient suffering from PE, there is a degree of pulmonary obstruction. The effects of the mechanical obstruction depend on the percentage of the pulmonary circulation that is obstructed, the existence or non-existence of a cardio-respiratory disease and on time taken for the obstruction to occur [1]. If the amount of obstruction is higher than 30%, then the pressure in the pulmonary artery is increased well beyond normal, and consequently, the right part of the heart is beaten. A serious obstruction cannot be compensated by pulmonary capillaries, thus leading to increased pulmonary vascular resistance. This, in turn, provokes an increase in the right ventricular afterload, which results in increased

**Citation:** Pagkratis, N.; Matsagas, M.; Malli, F.; Gourgoulianis, K.I.; Kotsiou, O.S. Prevalence of Hemorrhagic Complications in Hospitalized Patients with Pulmonary Embolism. *J. Pers. Med.* **2022**, *12*, 1133. https:// doi.org/10.3390/jpm12071133

Academic Editor: Giovanni Squadrito

Received: 11 June 2022 Accepted: 12 July 2022 Published: 13 July 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

parietal tension and finally, ischemia. Respiratory effects include tachypnea in 92% of patients and serious hypoxemia (PaO<sup>2</sup> < 70%) in a 63% [2].

PE presents a wide range of hemodynamic effects, from asymptomatic and undiagnosed disease to life-threatening emergencies. It is the third most frequent cause of death in hospitalized patients and a major cause of morbidity and mortality, with a total annual effect of 62 to 112 cases per 100,000 inhabitants [3]. Prognosis may worsen in PE patients, during intrahospital treatment, by experiencing hemorrhagic complications, which are mainly attributed to anticoagulant therapy [4].

Even with the best-coordinated care, hemorrhagic complications may occur. A minor hemorrhage could predict a major one and lead to modification of the anticoagulant therapy, underlying its importance for the prognosis and the efficient management of the major hemorrhagic episodes [1]. Hemorrhage is the most frequent complication caused by any anticoagulant [5].

Only a few studies investigated the in-hospital hemorrhage cases in patients with PE. Data regarding the in-hospital hemorrhagic complications in patients with PE presented with hemodynamic instability have also been scarcely noted, while the percentage of hemorrhagic complications has not been clarified in those receiving thrombolytic therapy. It is also important that there are no references regarding minor hemorrhages in patients hospitalized for PE.

In that context, this study aimed to evaluate the prevalence of hemorrhagic events in hospitalized PE patients and investigate the correlation of hemorrhagic events with the type of anticoagulant treatment, patients' demographic and clinical parameters, clinical burden, and outcome.

#### **2. Materials and Methods**

#### *2.1. Study Participants*

This was a retrospective study recording the hemorrhagic complications of patients with confirmed PE who were hospitalized in the Department of Respiratory Medicine of the University of Thessaly from January 2013 to April 2021. This research included all hospitalized patients in the Department of Respiratory Medicine, University of Thessaly with a discharge diagnosis I-26 Pulmonary Embolism (coding in ICD-10).

#### *2.2. Data Collection*

Demographic, clinical data, the type of anticoagulant treatment, the burden of disease, hemorrhagic events and outcomes were recorded by the Health Information System "ASKLIPIOS™ HOSPITAL" of the University Hospital of Larissa. Overview of all parameters extracted from the recordings are presented in Table 1.


**Table 1.** The parameters were extracted from the e-recordings of the patients hospitalized with PE.

#### *2.3. Statistical Analysis*

The chi-square test was used to make comparisons between frequencies. Unpaired t-tes was used for comparing parametric data between two groups, while non-parametric data were analyzed with the Mann–Whitney U test. Parametric data comparing three or more groups were analyzed with one-way ANOVA and Tukey's multiple comparisons test, while non-parametric were analyzed with the Kruskal–Wallis test and Dunn's multiple comparison test. Spearman's correlation was used for correlation analysis. Multiple logistic regression was used to examine a series of predictor variables to determine those that best

predict a hemorrhagic event. Statistical analyses were performed with IBM SPSS Statistics for Windows, version 23.0, IBM Corp., Armonk, NY, USA.

#### **3. Results**

The study included 326 patients with a mean age of 68.7 ± 17.0 years. 57.7% (188) of them were men and much younger than the women. 97.5% of patients were Greek, 1.2% were refugees, and the rest 1.3% were of other nationalities. 86.2% of the population had at least one comorbidity, with arterial hypertension being the most frequent one (52.5% of the patients).

Demographics and comorbidities are presented in Table 2. 8.9% of the total population had no prior medical history. 61.1% of men had had recent surgery in the last three months. Three of these operations had been performed on the vertebral column. In females, three cases of PE were noted in the postnatal period, three cases noted after a recent fracture and immobilization, and two cases after a recent fracture.


**Table 2.** Demographics and comorbidities of the sample, *n* = 326.

Note: Data are expressed as mean ± SD or as frequencies (percentages).

17.1% had a history of malignancy and 7.3% of them had a gender-related active disease. An absolute predominance of men (3.7% of the total population) was observed in the most frequent malignancy which is lung cancer. Surprisingly, in the present study, cancer was firstly diagnosed in 2.8% of patients, and more specifically, PE was the first sign of malignancy. 25.2% of the patients (82 people) had a history of thrombosis. 17.2% of the population received antiplatelet agents without any difference in the gender noticed. 5.8% (19) of the patients mentioned a previous episode of hemorrhage, and 21.1% presented a new hemorrhage during the hospitalization because of the PE. 9.2% of the population had a history of thrombophilia.

During the hospital admission, 44.2% presented dyspnea, 32% presented thoracic pain, 26.7% presented fever, and 7.7% had bloody sputum, while 4% of the population was asymptomatic. 18% presented tachycardia in the electrocardiogram (ECG).

Wells scores and Geneva scores, as well as the rates of the laboratory on patients' admission are presented in Table 3. 26% of the population had respiratory failure and 54% had hypocapnia on admission.

3.3% of the population presented with thrombocytopenia on admission. 10% were complicated by a fall in the number of platelets and thrombocytopenia during the hospitalization. 3.3% had an abnormal international normalized ratio (INR) >1.50, and 19.9% presented uremia on admission. 50% of the patients had a proximal deep vein thrombosis (DVT). 1.8% of the population had a paradox embolism.


**Table 3.** Wells scores, Geneva scores, clinical and laboratory data on patients' admission, *n* = 326.

Note: Data are expressed as mean ± SD; Abbreviations: ALT, Alanine Aminotransferase; AST, Aspartate Aminotransferase; BNP, Brain; Natriuretic Peptide; CRP, C-reactive protein; HCT, hematocrit; PO2, partial pressure of oxygen; PCO2, partial pressure of carbon dioxide.

Most of the patients admitted (92.4%) received LMWH, as shown in Table 3. 74.7% of them received 12-h action LMWH, and the rest received one subcutaneous dosage daily.

The anticoagulant therapy administered during patients' hospitalization is presented in Table 4. In 57.3% of the population, the treatment was modified during hospitalization. Specifically, 51.8% shifted to direct oral anticoagulants (DOACs), with which they were discharged. In 9.2% of the population, there had been a shift from 12-h to 24-h action Low-Molecular-Weight Heparin (LMWH), while in 0.9%, there had been a shift from 24-h to 12-h action LMWH. There was only one case of switching from DOAC to LMWH after an episode of gastric bleeding. The patients more frequently received rivaroxaban (75%) and less frequently dabigatran (12.5%) and apixaban (12.5%).



Data are expressed as frequencies (percentages).

15% of the hospitalized patients (49 people) experienced an episode of hemorrhage without any gender difference (12.2% of men vs. 17.4% of women, *p* = 0.240). 18.4% of them experienced a major hemorrhage, without any difference regarding the gender noticed. Major hemorrhages were much fewer than the minor ones (18.4% vs. 81.6%, *p* = 0.001), while the average duration of hemorrhage was 3 ± 2 days. The sites of the hemorrhage are presented in Table 5.



Data are expressed as frequencies (percentages).

16% of the patients with hemorrhagic complications (2.1% of the sum) needed a transfusion. The patients who had been transfused were the ones that presented major hemorrhages. 2.1% of the patients with hemorrhagic complications (0.3% of the total population) needed to be transferred to an ICU because of the bleeding.

4.2% of patients with hemorrhagic complications had to interrupt the anticoagulant therapy by missing doses, and 19.1% had to shift to 12-h action LMWH, especially enoxaparin. One out of the 49 patients with hemorrhagic complications who interrupted the therapy experienced a thrombotic event (2.1%). The average duration of hospitalization was 8 ± 5 days. 5.2% of the patients died. No death due to hemorrhagic complications was recorded.

The highest Wells score and the highest rate of creatinine (1.3 vs. 12 + 0.2, *p* = 0.029) were positively correlated with the risk of hospital bleeding. An accounting regression model was used to search for dependent variables (age, gender, comorbidities, the presence of cancer, Wells score, INR on admission, uremia on admission, location of PE, PESI score, ICU, platelet count on admission, right heart failure, instability, antithrombotics) to identify the parameters that could predict hospital bleeding, but no clinical or laboratory predictors were identified.

#### **4. Discussion**

The frequency of hemorrhagic complications during the hospitalization of patients with PE has not been previously determined in Greece. The present study was the first to investigate this issue. We found that 15% of the population hospitalized due to PE were complicated with hemorrhage, of whom 18.4% experienced a major event. Major hemorrhages were fewer than minor. Nadroparin related to 83.3% of the major events. Hematuria was the most common hemorrhagic event. 22% of patients with major events received a transfusion, and 11% were admitted to ICU. The events lasted for 3 ± 2 days. No death was recorded.

We found that among the hospitalized patients due to PE, males outnumbered females, a finding following the literature supporting that the risk of PE is higher in men than in women [6]. In some studies, the frequency of unprovoked PE varies between 16.5% and 51%, up to 69–76% [7–9]. We considered that unprovoked PE should be accepted when there are no comorbidities or provocations with proven PE hazards. Based on this definition, it was found that the frequency of unprovoked PE was 8.9% in our study. However, major predisposing factors were detected in the majority population, such as major surgery, fractures, and postnatal period [10].

The delayed diagnosis of PE was a finding of great interest that accords with previous reports commenting that PE has no typical symptomatology, thus, confirming the difficulty in PE diagnosis [11]. Patient delay of an average of 4.2 days and delay in primary care of an average of 3.9 days were the major contributors to this delay [12]. However, diagnostic delay of PE of more than seven days is common in primary care, especially in the elderly, and if chest symptoms, like pain on inspiration, are absent [13,14].

Surprisingly, in the present study, cancer was firstly diagnosed in 2.8% of patients, and more specifically, PE was the first sign of malignancy. In the case of cancer, the venous thromboembolism (VTE) risk is increased from 7 up to 28 times [15]. Neoplasia is caused when the tumor secretes substances with a prothrombotic effect, such as adhesion molecules that activate the macrophages and the platelets [16]. It has been reported that cancer is

usually diagnosed within the first months after a VTE episode, with an overall incident rate of 4.1% in 1 month and 6.3% in 1 year [17].

25.2% of the patients had a previous history of thrombosis. The location and manifestation of thrombosis are of great predictive value for the risk of re-thrombosis. In a meta-analysis of patients with PE or/and DVT, the re-thrombosis percentages were 22% for PE and 26.4% for DVT [18]. The risk for a new PE was 3.1 times higher in patients with symptomatic PE than in those with proximal DVT. The patients with proximal DVT had a 4.8 times higher percentage of relapse than those with peripheral DVT [19].

In the present study, 3.1% of the patients were suffering from chronic renal failure (CRF), a disease associated with an increased risk for hemorrhage, because of the platelet dysfunction and uremic toxins in the blood, which harms the primary hemostasis [20]. Also, patients with moderate or severe CRF present higher rates of major hemorrhage than those with mild to non-CRF during the next 12 days after VTE diagnosis, despite the administration of anticoagulant therapy [20,21].

Moreover, 1.5% of the patients had asthma, and 4.3% had chronic obstructive pulmonary disease (COPD). It has been shown that asthma increases the risk for PE. In comparison with the non-asthmatic people, asthmatic patients of all age groups run an increased risk for PE, which is even more increased depending on the age and the severity of the respiratory disease [22]. Even in a stable phase, COPD is considered an independent risk factor for PE. At the same time, a meta-analysis suggests that one out of four patients with a COPD exacerbation who need hospitalization may suffer from PE [18,22].

Diabetes mellitus (DM) appeared in 12.2% of the patients. Clinically, patients with PE who suffer from DM have a higher risk of mortality than those who do not suffer from DM, while it seems that elevated glucose rates increase the risk of VTE [23]. Also, a study on the Asian population considers insulin-independent diabetes as an independent risk factor for the development of DVT and PE [24].

15% of the hospitalized population were complicated with hemorrhage, of whom 18.4% experienced a major event. Major hemorrhages were fewer than minor (29.8% vs. 70.2%, *p* = 0.001). Hematuria was the most common hemorrhagic event. The overall prevalence of bleeding in acute PE cohorts is approximately 10/100 patient-years [25–28].

Specifically, in the MAPPET registry, among 1001 cases of PE, 92 (9.2%) presented a major hemorrhage required a transfusion of blood units or discontinuation of the anticoagulant therapy [29]. In the EMPEROR registry, 10.3% of the patients with massive pulmonary embolism (MPE) and 3.5% of those without MPE had hemorrhagic complications. 3 out of the 63 patients in the second group died because of the hemorrhage [7].

In the IPER registry, among 1716 patients with PE, a loss of hemoglobin > 4 g/dL was reported in 53 patients (3.1%), while 6 out of 10 patients with intracranial hemorrhage died [30]. In the ZATPOL registry, hemorrhagic complications were reported in 6% (67 out of 1112) of the patients with PE. Major hemorrhage was reported in 3.6% of the patients, while 0.5% had a fatal hemorrhage. Among the patients receiving anticoagulant therapy, 24% (29 patients) presented hemorrhagic complications. Specifically, 19% (23 cases) of the hemorrhages were major and 5% were (6 cases) minor. 38 hemorrhagic cases were reported in patients who had not received thrombolysis. 17 of them were major and 21 were minor. Among the 67 cases that presented hemorrhagic complications, 17 were presented after oral anticoagulant therapy was initiated [31].

Recently, a higher risk of bleeding (RR: 2.53, 95% CI: 1.60–4.00; I 2: 65%) has been reported in ICU patients receiving an anticoagulant therapeutic regimen [26,27]. In the elderly population, in which the risk of acute PE is increased due to advanced age, bleeding is even more pronounced, with the risk of major bleeding including intracerebral bleeding doubling in patients aged above 80 years and the risk of hemorrhagic complications is highest in the early days of treatment [32–35]. Interestingly, the risk of bleeding resulting in hospitalization or death within 3 and 12 months after the index PE admission increased over the last years [36].

Several bleeding risk prediction scores have been proposed, including the VTE-BLEED, RIETE, HASBLED, and HEMORR2HAGES scores [31–35], that have several limitations as most are retrospective, few focus on real-life cohorts, and patients in the stable (not acute) phase of anticoagulation are mainly included [37].

Most patients hospitalized due to PE received LMWH related to 6 major and 39 minor hemorrhagic episodes. Fondaparinux was only related to minor episodes of hemorrhage. According to studies that have compared it with enoxaparin, the percentage of major hemorrhage in 9 days is much lower when fondaparinux is used rather than enoxaparin [38]. On the other hand, we found that nadroparin related to 83.3% of the major events. Generally, it has been reported that absolute major bleeding rates are low for all LMWH agents [38]. Nevertheless, twice-daily dosing with nadroparin appeared to be associated with a 1.77 times greater bleeding risk as compared with once-daily dosing, as also suggested in a meta-analysis of controlled clinical trials [38,39].

As initial therapy, low-risk patients can receive DOACs, specifically rivaroxaban or apixaban [29]. Rivaroxaban and apixaban can be given in a higher initial dose without previous heparin therapy [29]. In the present study, 4.1% were receiving DOACs, and they underwent one major and one minor hemorrhagic episode. Multiple clinical studies support the safer bleeding profile of DOACs over Vitamin K antagonists [38]. However, it has been supported that DOACs at standard dose, except apixaban, had a higher risk of major gastrointestinal bleeding compared to warfarin. Apixaban had a lower rate of major gastrointestinal bleeding compared to dabigatran and rivaroxaban [40].

2% of patients with major events received a transfusion, and 11% were admitted to ICU. The bleeding events lasted for 3 ± 2 days. No death was recorded. Hemorrhagic complications were associated with an average hospitalization of 10.7 days, with higher risk of hospital-acquired infection and higher healthcare cost, compared to 7.4 days of hospitalization for those without bleeding, In-hospital major bleeding has been identified as strong predictor of in-hospital (OR 7.7, 95% CI 2.3–25.8) and 1-year mortality (HR 3.6, 95% CI 2.0–6.6), especially in normotensive patients [41]. Generally, an improvement in mortality has been reported over years attributed to both a real improvement in patient care and "over-diagnosis" of incidental and sub-segmental PE [36].

According to a recent meta-analysis of 14 randomized controlled trials and 13 cohort studies, including 9982 patients who received a vitamin K antagonist and 7220 received a DOAC, it has been supported that the incidence of major bleeding was statistically significantly higher among those who had creatinine clearance less than 50 mL/min [42]. Accordingly, in the present study, we found a correlation between high serum creatinine levels and hemorrhagic complications, but the regression model did not prove that this variable was an independent predictor of hemorrhage. A few limitations need to be noted regarding the present study. A major limitation of this study was its retrospective design that it might generate a great deal of missed data. There was also absence of data on potential confounding factors.

#### **5. Conclusions**

15% of the hospitalized patients of the study (49 patients) presented an episode of hemorrhage, while 18.4% of them presented an episode of major hemorrhage. Hemorrhages were mainly minor and there was no hemorrhage leading to death. 16.2% of the patients with hemorrhagic complication (2.1% of the total population) needed transfusion. The average duration of hemorrhage was 3 ± 2 days. 2.1% of the patients with major hemorrhage (0.3% of the total population) needed to be transferred to an ICU, because of the hemorrhagic complication. 83.3% of the cases that presented major hemorrhage and received LMWH were given nadroparin. There was not any independent predictor of hemorrhage, but there was a correlation between high Wells score or high levels of serum creatinine and hemorrhagic complication.

Only a few studies investigated the in-hospital hemorrhage cases in patients with PE, as detecting these rare events in large datasets remains difficult. The present study evaluating data throughout an 8-year period highlights a significant likelihood of bleeding and a small, but not negligible, possibility of major hemorrhage in patients hospitalized for PE. We found that nadroparin administration was associated with major hemorrhagic events; thus, it should probably not be the first therapeutic choice among other LMWH during the in-hospital treatment of patients with PE. Until now, there are no clear guidelines and scientific evidence available for physicians in this field for early diagnosis and tools to avoid hemorrhagic complications in patients hospitalized for PE. The optimal management of bleeding involves the application of predictive scores in combination with anticoagulant reversal strategies. However, risk assessment tools are relevant in managing patients with atrial fibrillation but are not widely validated in PE patients. Hence, the performance of existing prediction models in patients with PE should be further assessed. More comprehensively, the combination of clinical, biological, and genetic markers should be incorporated to build predictive scores to estimate the risk of bleeding and help the decision process about the proper type of anticoagulant treatment.

**Author Contributions:** Conceptualization, O.S.K., M.M., K.I.G. and N.P.; methodology, O.S.K. and N.P.; validation, M.M., K.I.G., F.M. and O.S.K.; formal analysis, O.S.K.; investigation, N.P.; data curation, N.P.; writing—original draft preparation, N.P. and O.S.K.; writing—review and editing, N.P., O.S.K. and M.M.; supervision, O.S.K.; project administration, O.S.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University of Thessaly (No. 2800- 01/11/2020; approved on 1 November 2020).

**Informed Consent Statement:** Patient consent was waived due to the retrospective nature of this study and the analysis used anonymous clinical data.

**Data Availability Statement:** The data that support the findings of this study are available on request from the corresponding author, O.S.K.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **Exercise Preferences and Benefits in Patients Hospitalized with COVID-19**

**Sevasti Kontopoulou <sup>1</sup> , Zoe Daniil <sup>1</sup> , Konstantinos I. Gourgoulianis <sup>1</sup> and Ourania S. Kotsiou 2,\***

> <sup>1</sup> Department of Respiratory Medicine, University of Thessaly, 41110 Larissa, Greece; sevi\_kon@hotmail.com (S.K.); zdaniil@uth.gr (Z.D.); kgourg@med.uth.gr (K.I.G.)

<sup>2</sup> Faculty of Nursing, University of Thessaly, 41110 Larissa, Greece

**\*** Correspondence: raniakotsiou@gmail.com

**Abstract:** Background: Obese people are at risk of becoming severely ill due to SARS-CoV-2. The exercise benefits on health have been emphasized. Aim: To investigate the correlation of obesity with the length of hospitalization, the pre- and post-hospitalization exercise preferences of COVID-19 patients, and the impact of pre-admission or post-hospitalization physical activity on dyspnea one month after hospitalization and recovery time. Methods: A telephone survey was conducted in patients hospitalized at the Respiratory Medicine Department, University of Thessaly, Greece, from November to December 2020. Results: Two-thirds of the patients were obese. Obesity was not associated with the hospitalization time. Two-thirds of the patients used to engage in physical activity before hospitalization. Males exercised in a higher percentage and more frequently than women before and after hospitalization. The methodical pre-hospitalization exercise was associated with lower levels of dyspnea one month after hospitalization. In-hospital weight loss, comorbidities, and dyspnea on admission independently predicted longer recovery time. Lockdown had boosted men's desire to exercise than females who were negatively affected. Conclusions: Obesity is common in COVID-19 hospitalized patients. In-hospital weight loss, comorbidities, and dyspnea on admission predicted a longer post-hospitalization recovery time. The pre-hospitalization exercise was associated with less post-hospitalization dyspnea and recovery time.

**Keywords:** dyspnea; exercise; hospitalization; recovery

#### **1. Introduction**

COVID-19 is a multisystemic and multivessel disease that involves the respiratory, cardiovascular, renal, gastrointestinal, and central nervous systems. The presence of comorbidities increases the risk for severe illness due to SARS-CoV-2. More often, people with underlying medical conditions display respiratory failure that requires admission to the ICU, multiorgan failure, or even loss of their lives [1].

Obesity exposes infected individuals to peril, increasing the required days of hospitalization and recovery. This connection occurs because the chronic storage of body fat is directly associated with a chronic pro-inflammatory state, weakening the immune system and creating an ideal environment for the virus to grow inside the fat cells. Obesity, combined with comorbidities, aggravates the symptoms of the COVID-19 disease by extending the time of needed hospitalization and eventually raising the mortality rate. Regular body fat storage and a sedentary lifestyle adopted due to the mandatory quarantine can create the perfect conditions for infection and growth of contagious pathogens, such as the SARS-CoV-2 virus among the more vulnerable population [2].

Consequently, a risk factor for severe illness and needed admission to the Intensive Care Unit (ICU) is the increased body mass index (BMI), namely the increased storage of body fat [3]. A recently conducted study pointed out that the need for intensive mechanical ventilation for COVID-19 patients under 60 was seven times higher for those with a BMI over 35 kg/m<sup>2</sup> compared to those with a BMI under or equal to 25 kg/m<sup>2</sup> [4].

**Citation:** Kontopoulou, S.; Daniil, Z.; Gourgoulianis, K.I.; Kotsiou, O.S. Exercise Preferences and Benefits in Patients Hospitalized with COVID-19. *J. Pers. Med.* **2022**, *12*, 645. https://doi.org/10.3390/jpm12040645

Academic Editor: Bruno Mégarbané

Received: 13 February 2022 Accepted: 14 April 2022 Published: 17 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Starting on 4 May 2020, a 42-day strict lockdown was implemented in Greece. Movements of individuals to serve their needs outside the house were permitted only for seven categories of reasons: (i) transition to the workplace during work hours; (ii) going to the pharmacy or visiting a doctor; (iii) going to a food store; (iv) going to the bank for services not possible online; (v) helping a person; (vi) going to a significant ritual (funeral, marriage, baptism) or movement, for divorced parents, which was essential for contact with their children; and (vii) moving outdoors for exercising or taking one's pet out, individually or in pairs. Again, from 7 November 2020, Greece implemented new measures and restrictions on movement and business activity. Kindergartens, primary and special schools initially remained open, and from 18 November 2020, they switched to distance learning. On 14 December 2020, shops, hairdressers, and other facilities were allowed to open, while schools and restaurants remained closed [5].

At the beginning of the quarantine, people generally adopted a sedentary lifestyle with decreased physical activity, ultimately harming their physical and psychological health and their quality and quantity of sleep [6]. Multiple vulnerabilities and an interplay leading from simple anxiety to clinical depression and suicidality through distress were revealed among the Greek population [7].

Moderate to intense exercise entrains important positive adjustments to the cardiorespiratory ability, reduces the levels of chronic inflammation that may have preexisted due to related diseases, and improves the immune system's function for a faster reaction against viral infection such as COVID-19 [8], improves lipid profile, reduces the BMI, and can have a positive effect on our psychological health [9]. Moreover, physical exercise can help individuals maintain their muscle mass, good respiratory function, and boost the immune system to keep respiratory function levels high [10].

An international online survey including 41 research institutions from Europe, Western-Asia, North-Africa, and the Americas, documented that COVID-19 lockdown deleteriously affected physical activity and sleep patterns [11]. A recently conducted Greek study supported that during a pandemic, compared with a typical week, physical activity of a high and moderate intensity decreased for 43.0% and 37.0% of participants, did not change in 32.9% and 36.1% of participants, and increased only in 24.1% and 26.9%, respectively, whereas walking time decreased in 31.3%, did not change in 27.3%, and increased in 41.5% of participants [12]. In fact, after consecutive lockdown periods were examined, a decline in overall physical activity was evident in all age and gender groups during each lockdown phase [13].

During the pandemic, health regulators constantly point out, mainly to the elderly confined by the quarantine, to avoid a sedentary way of life and engage in any form of physical activity [8]. Furthermore, recent studies mention that exercise reduces hospitalizations due to the COVID-19 disease [14].

This study aimed to investigate the correlation of obesity with the duration of hospitalization, the pre–and post-hospitalization exercise preferences of COVID-19 patients, and the impact of pre-admission or post-hospitalization physical activity on dyspnea one month after hospitalization.

#### **2. Materials and Methods**

#### *2.1. Procedure*

This retrospective study was conducted via a telephone survey from February 2021 to March 2021. One specialized trainer was the researcher of this study and contacted previously hospitalized patients via telephone to interview by asking them a list of predetermined questions.

#### *2.2. Participants*

In the study were included all patients regardless of their age who were infected by the SARS-CoV-2 virus and were hospitalized at the Infection Diseases Unit (COVID-19) of the Department of Respiratory Medicine of the University of Thessaly from November to December 2020. An exclusion criterion was the inability to acquire information due to their non-consent or lack of good cooperation.

#### *2.3. Study Tools*

In this study, we used an original self-reported questionnaire with 26 questions regarding:


#### *2.4. Statistical Analysis*

The statistical analysis was conducted with the IBM SPSS v23. The quantitative variables were presented as mean value ± standard deviation (SD), and the qualitative variables were presented as an absolute value (frequency). The frequencies were compared with the chi-square statistical test. The *t*-test was used to test the difference between two mean values from independent samples. The nonparametric data were analyzed with the Mann–Whitney U test. The parametric data that compare three or more groups were analyzed with the ANOVA unidirectional Variance Analysis and the post hoc Bonferroni multiple comparison test. In contrast, the nonparametric data were analyzed with the Kruskal–Wallis test and the Dunn multiple comparison test. Spearman's correlation was used for the correlation analysis. A multiple linear regression model was utilized to examine a series of prediction variables to find those that can better predict faster recovery time in days.

#### **3. Results**

#### *3.1. Demographics, Clinical Parameters, and Symptomatology of the Hospitalized COVID-19 Population*

In total, 42 men (66%) and 22 women (34%) were included in the study, with a mean age of 62.2 ± 13.2 years old (min = 21 years, max = 91 years). The demographic and clinical parameters of the study's population and comparisons according to sex are presented in Table 1.

The men were significantly taller and heavier than the women, as expected (Table 1). The study's patients were overweight on average with no BMI difference between the genders. A total of 45.3% of the patients (29 patients) were overweight, while 23.4% (15 patients) were obese. Only 31% of the patients had normal weight. The comorbidities of the sample hospitalized due to COVID-19 and their comparison based on sex can be found in Table 2.


**Table 1.** Demographic and clinical data of the study's population and their comparison according to sex.

Note: The data are presented as mean value ± SD; \* Student's *t*-test.

**Table 2.** Comorbidities of the study's population and their comparison based on sex, n = 64.


Note: The data are presented as frequencies (percentages) or mean values ± SD; \* Chi-square Test of independence; Abbreviations: CD, coronary disease; CVA, cardiovascular accidents; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease; HLD, hyperlipidaemia.

The most common comorbidity of the sample was hypertension. The women suffered from DM in a greater frequency than men. The patients with comorbidities were significantly older than those without (65 ± 12 vs. 57 ± 13 years old, *p* = 0.019), as expected.

Older age was positively associated with the level of mMRC dyspnea on admission (r = 0.402, *p* = 0.001) and after hospitalization (r = 0.280, *p* = 0.025). The number of comorbidities was positively associated with the level of mMRC dyspnea before (r = 0.499, *p* = 0.001) and after (r = 0.298, *p* = 0.050) hospitalization. Patients hospitalized due to COVID-19 with high blood pressure, or another cardiovascular disease displayed higher levels of mMRC dyspnea at the time before their admission to the hospital by comparison with those who did not have any comorbidities (2 ± 1 vs. 1 ± 1, *p* < 0.001). Patients with COPD hospitalized due to COVID-19 also displayed higher levels of mMRC dyspnea before their hospital admission than those who had no comorbidities (4 ± 1 vs. 1 ± 1, *p* < 0.001).

The symptomatology of the study's population on admission and the comparisons based on the presence or absence of comorbidities are presented in Table 3.

The most frequently displayed symptom during hospital admissions among the study's population was fever. There was no differentiation on the symptoms during admission between genders. Patients with at least one comorbidity more often had dyspnea and fever than the healthy patients prior to infection (Table 3). There was no significant difference in length of stay between patients with and without comorbidities.


**Table 3.** Population's symptomatology during admission and their comparison based on the presence or absence of comorbidities n = 64.

Note: The data are presented as frequencies (percentages) or on average ± SD; \* Chi-square Test of independence.

#### *3.2. Physical Activity Preferences of the Population before Hospitalization*

The physical activity of the study's population before hospitalization and their comparison based on sex can be found in Table 4.

**Table 4.** Physical activity of the study's population before hospitalization and comparisons based on sex, n = 64.


Note: The data are presented as frequencies (percentages) or mean values ± SD; \* Chi-square Test of independence.

In total, 65% of the sample mentioned their engagement with at least one form of physical activity before being hospitalized. The men were exercising at a higher percentage than the women before their hospitalization and with greater frequency (Table 4). Overall, 51.6% of the study's population engaged in physical activity mentions walking as their primary exercise before hospitalization despite gender (Table 4).

The patients with at least one chronic disease were used to engage less in aerobic exercise (!8.2% vs. 2.4%, *p* = 0.044) and resistance training (18.2% vs. 0%, *p* = 0.010) compared to individuals without comorbidities.

#### *3.3. Physical Activity Preferences of the Population after Hospitalization*

The post-hospitalization physical activity of the study's population and their comparison based on sex is presented in Table 5.


**Table 5.** Post-hospitalization physical activity of the study's population and their comparison based on sex n = 64.

Note: The data are presented as frequencies (percentages); \* Chi-square Test of independence.

Overall, 68.8% of the study's population started or continued to exercise after their hospitalization. Men continued to exercise on a larger scale and at a greater frequency than women after their hospitalization (Table 5). In 57.8% of the abovementioned population engaged in any form of physical activity after their hospitalization mentioned walking as their primary physical activity, with the male percentages being significantly higher compared to those of females (Table 5). In 36 out of the 42 individuals (85.7%) exercising before their hospitalization continued to exercise after it. In 8 out of the 44 individuals (18.2%) exercising after their hospitalization had just started, and they were not before.

### *3.4. Changes of Body Weight before and after Hospitalization, and Views about the Effects of Hospitalization or Lockdown on the Frequency and the Desire of Exercise after Discharge*

The change of body weight before and after hospitalization, the various views about the effects of hospitalization or lockdown on the frequency and the desire of exercise after discharge and comparisons based on sex are shown on Table 6.

The BMI was not associated with the duration of hospitalization. However, the days of hospitalization were positively associated with more extensive changes in the patients' weight (r = 0.809, *p* < 0.001). Patients who experienced in-hospital weight loss were hospitalized more days than those who had gained weight (19 ± 14 vs. 9 ± 4, *p* < 0.001).

At a significantly higher percentage than the women, the men stated that their hospitalization or lockdown measures did not affect their exercise frequency after discharge from the hospital. In total, 50% of the men (a more considerable percentage than the women at 9%) supported that the restrictive measures did not affect the frequency of their desire to exercise. Actually, among men, 38.1% that were not previously exercising mentioned that the lockdown increased their desire to begin. Approximately half of women (45.5%) were negatively affected by the lockdown regarding their frequency and desire to exercise, a significantly higher percentage compared to the men.

Both the men and the women that did not exercise (before or after hospitalization) were those individuals who supported the harmful effects of the lockdown to their desire for exercise in comparison to those physically active before their hospitalization. The men and the women that believed in the positive effects of their hospitalization to their desire for exercise were also those who supported the positive effects of the restrictive measures to that desire.

The frequency of exercise before hospitalization was positively associated with the frequency of exercise after hospitalization (r = 0.645, *p* < 0.001). The patients previously engaged in physical activity needed significantly lesser time to recover (22 ± 14 vs. 65 ± 32 days, *p* < 0.001) and displayed significantly lower levels of dyspnea on the mMRC scale (1 ± 1 vs. 3 ± 1, *p* < 0.001) after their hospitalization compared to the patients with no history of physical activity. In fact, the frequency of exercise (days per week) was negatively associated with the levels of mMRC dyspnea after hospitalization (r = −0.342, *p* = 0.026). On the contrary, the recovery time in days was positively associated with the

time of hospitalization (r = 0.408, *p* = 0.001). Of all forms of exercise, walking was associated with a shorter recovery time (23 ± 15 vs. 55 ± 34, *p* < 0.001) and a lower score on the mMRC dyspnea scale (1 ± 1 vs. 2 ± 1, *p* = 0.04) in comparison to the absence of any previous physical activity.

**Table 6.** Changes in body weight before and after hospitalization, views about the effects of hospitalization on the frequency of exercise after discharge and during recovery, effects of the lockdown to the frequency and the desire of the study's population to exercise and their comparison based on sex, n = 64.


Note: The data are presented as frequencies (percentages) or mean values ± SD; \* Chi-square Test of independence; \*\* Bonferonni method; # Weight change refers to weight gain or loss.

The patients engaged in physical activity after their hospitalization displayed significantly lower levels of dyspnea on the mMRC scale (1 ± 1 vs. 3 ± 1, *p* = 0.001) compared to those who did not engage in any physical activity after their hospitalization. Of all the different kinds of post-hospitalization exercise, walking was positively associated with the most remarkable improvement of dyspnea after hospitalization (1 ± 1 vs. 2 ± 1, *p* = 0.004).

A multiple linear regression model was used to research those parameters that can independently predict a quicker recovery. The weight loss in Kg, the presence of chronic disease, and dyspnea on admission, were found to be independent predictors of a faster recovery time (in days) of patients hospitalized due to COVID-19 (R<sup>2</sup> = 92.0, adjusted R 2 : 89.4) (Table 7).


**Table 7.** Multiple linear regression model to predict a faster recovery time (a).

a. Dependent Variable: recovery time in days; b. Predictors: (Constant), weight loss, chronic disease, dyspnea.

#### **4. Discussion**

In this study, we investigated the correlation of obesity with the duration of hospitalization, the pre- and post-hospitalization exercise preferences of COVID-19 patients, and the impact of pre-admission or post-hospitalization physical activity on dyspnea one month after hospitalization or recovery time after discharge. We found that two-thirds of the patients were obese; however, obesity was not associated with the hospitalization time. Two-thirds of the patients used to engage in physical activity before hospitalization. The men exercised in a higher percentage and more frequently than women before and after hospitalization. The methodical pre-hospitalization exercise was associated with lower levels of dyspnea one month after hospitalization and less recovery time in days. Weight loss in Kg, preexisting comorbidity, and dyspnea on admission were independent predictors for faster recovery time in days. Most males used to exercise before infection supported that the lockdown had boosted their desire to exercise, compared to females who were negatively affected.

In our study, the average age of the sample of hospitalized patients in November and December 2020 was 62.2 ± 13.2 years old. It is distinctively mentioned that the frequency of infection is increased for men over 60 years old, who also display higher mortality rates than the women of the study up to 50%. According to the World Health Organization, there were expected differences between the two genders regarding the body metric data, while the population studied was overweight on average, without any statistically significant difference among genders.

Comorbidities were positively associated with age, as expected. It has been reported that 60 to 90% of patients who need hospitalization due to COVID-19, display at least one comorbidity [15,16]. The most common comorbidities mentioned in the literature are hypertension (57% of the patients), respiratory diseases (10% of the patients), and malignancy (8% of the patients) [15,16]. In a recent meta-analysis of 10 studies that included 76,993 patients, the prevalence of high blood pressure was 17.37% (95%CI, 10.15–23.65%), of cardiovascular disease 12.11% (95%CI 4.40–22.75%) and of DM 7.87% (95%CI 6.57–9.28%) [17]. Accordingly, the present study revealed that two-thirds of the population had at least one of the following comorbidities in descending order, high blood pressure (29.7%), DM (21.9%), cardiovascular diseases (CD, CVA) (18.7%), hypercholesterolemia/dyslipidemia (15.6%), malignancy (9.3%), coexistent autoimmune disease (4.7%), COPD (4.7%), asthma (3.1%), chronic hepatitis (3.1%), and thyroid disease (3.1%).

It has been reported that patients with more than one comorbidity experience more dyspnea [18]. In the present study, there was a positive association of dyspnea before and after hospitalization with the presence of comorbidities. Furthermore, advanced age was positively associated with the level of mMRC dyspnea on admission and after discharge. It has been reported that at least 87% of the infected by the virus still displayed at least one of the common symptoms after their recovery, more often dyspnea and fatigue, while 15% of the examined patients displayed increased breathing difficulty as a complication of the virus [19].

Patients with increased BMI are more likely to be severely infected by the new coronavirus. A large percentage of patients that need intensive care are overweight or obese [20]. We found that increased BMI was not associated with hospitalization time. However, it is essential to point out that the patients with more weight included in this study were exercising more often before their hospitalization, and the benefits of their exercise could possibly counterbalance their increased BMI. Physical activity is an important means of promoting health. In addition to improving functions related to cardiovascular and respiratory function, as well as avoiding the deposition of body fat, physical activity has many benefits for avoiding infectious diseases, or in the case of hospitalization, for faster recovery [10,21].

More specifically, physical activity improves inflammation associated with COVID-19 independent of body fat, explaining why bodyweight was not an independent contributor to hospitalization. There are data demonstrating that regular bouts of short-lasting (i.e., 45–60 min), moderate-intensity exercise (50–70% VO2max), performed at least three times per week is beneficial for the host immune defense, particularly in older adults and people with comorbidities [22,23], compared to prolonged and/or intense bout of endurance exercise that makes humans more susceptible to infection. Moderate-intensity exercise has been linked to increased leukocyte function in humans [24]. It has been found to enhance chemotaxis, degranulation, phagocytosis, cytotoxic activity, and the oxidative activity of macrophages and neutrophils in rats [25]. Increased cytolytic activity of NK cells and NK cell-activating lymphokine during 60 min of moderate-intensity exercise by healthy cyclists was also reported [11]. On the contrary, the long-duration/intense exercise-induced immunomodulation is associated with markers of immunosuppression, such as increased production of proinflammatory cytokines [26] reduced activity of NK cells, T and B lymphocytes, and neutrophils; reduced production of salivary IgA and plasma IgM and IgG; and a low expression of major histocompatibility complex II in macrophages [27,28]. These changes can be detected hours to days after the end of a prolonged and/or intense endurance exercise. Physical activity controls the viral gateway, modulates inflammation, stimulates NO production pathways, and establishes control over oxidative stress. Adaptation to usual exercise appears to affect immune function, particularly innate and adaptive immunity, and improve humoral immunity with increased vaccination responses. Exercise may at least partially counteract the detrimental effect of SARS-CoV-2 binding to the angiotensin-converting enzyme receptor [22]. Physical can activate anti-inflammatory signaling pathways. In this regard, the release of antiinflammatory cytokines from skeletal muscle contraction, cortisol elevations, prostaglandin E2, and soluble receptors against tumor necrosis factor and interleukin 2, and increased mobilization of immunoregulatory leukocyte subtypes may be relevant in attenuating the cytokine release in COVID-19. Exercise may enhance alternative routes of NO production, stimulating eNO with antiviral effects and post-infection lung recovery of COVID-19. The control of oxidative stress, which produce cell damage, is modulated by the practice of physical activity by two mechanisms, the inhibition of NF-κB, and the stimulation of Nrf2 pathways [22].

A total of 65% of the studied sample mentioned that they engaged at least in one form of physical activity before their hospitalization, with walking being the most common. The men, compared to the women, were exercising more frequently both before and after their hospitalization. Our results are in agreement with previous studies documenting that woman were less active than men [29] and that levels of physical activity decreased progressively with age especially among women [30]. Another study reported that women were less likely than men to prefer activities that require skill and practice or done outdoors [31].

Conversely, a recent Italian study reported that women, who previously had a lower level of physical activity than men, showed a lower tendency to reduce it during lockdown, revealing greater resilience than men. During that period, women were motivated by weight loss and toning more than men [32,33], being concerned with controlling their weight, improving their physical appearance, or counteracting the effects of aging. In the present study, the factor that affected their pursuit of physical activity differs between the two sexes, and the leading cause of this phenomenon may be that Greek women tend to spend more time handling family matters and everyday family needs [34], despite the

fact that females reported higher motivation for appearance and physical condition than males [33]. A study that was conducted before the COVID-19 pandemic documented that half of the studied population were physically inactive, indicating that sedentary lifestyles have become a serious epidemic in Greece [35].

An interesting finding of this study was that the patients methodically exercising had lower levels of dyspnea one month after hospitalization, needed less time to recover from the infection and return to their previous physical condition. Out of all kinds of exercise we included, walking was positively associated with a faster recovery time of the hospitalized patients. Given that in the present study the age group is over 60, we assume that it is not the walking itself in regard to intensity and how it affects the body but the fact that we had a high number of individuals in that category. Walking is a type of physical activity that is relevant to older adults as the walking ability is of primary importance for older adults [36]. On the other hand, there are data demonstrating that regular bouts of moderate-intensity exercise performed at least three times per week is more beneficial in immunomodulation compared to high intensity exercise, as mentioned above.

Exercise positively affects the immune system contributing considerably to its improvement, bearing in mind the kind, the intensity, and the duration of exercise [37]. Overall, it is a fact that mild exercises stimulate cellular immunity, increase the anti-infective activity of the macrophages and the effect of the inflammatory cytokines, contributing to the faster cure from the infection [37]. There are no data regarding the improvement of the immune response to COVID-19 infection through exercise; however, there are indications, from past viral infections of the respiratory system, of physical activity decreasing the duration and the severity of the symptoms, as well as the mortality rates of the viral disease. Physical activity of mild intensity could be considered nonmedicinal means to the fight against respiratory infections [37].

Regarding the physical activity after hospitalization, walking was again the main preferable exercise by 57.8%. It is essential to mention that this percent value increased (up to 68.8%) because most patients continued to exercise, and at the same time, some patients had begun to exercise after their recovery. Similarly, in that situation, there were more men who continued to exercise than women. In addition, the men stated that the hospitalization and restrictive measures did not affect their frequency or desire to exercise and they continued to work out the same after being cured, some even more, however, the women were affected negatively and reduced their exercise frequency.

A percentage of men and women believed that their hospitalization and the restrictive measures increased their desire to exercise. Mainly, the patients who exercised before infection claimed that their desire decreased, and their exercise after hospitalization became even more intense. On the other hand, the patients that were not exercising at all continued to keep their distance from physical activity and demonstrated that the confinement and the hospitalization affected their desire to exercise negatively. A large proportion of patients (20%) with COVID-19 will continue to have clinical manifestations of the disease, such as fatigue, weakness, severe dyspnea, and headaches for a period that may exceed one month. There is considerable evidence that physical activity has long-term health benefits that mitigate or even prevent the development of chronic non-communicable diseases (lung disease, heart disease, neurocognitive problems, musculoskeletal problems). On the other hand, physical inactivity has been associated with serious COVID-19 problems, including dyspnea [38]. Accordingly, the Centers for Disease Control and Prevention (CDC) advise not only to engage the inactive population in physical activity, but also to establish it as a tool in the management of patients with post-COVID-19 syndrome. Since exercise has been shown to be beneficial for many viral infections such as COVID-19, it is worth highlighting and further examining the extent of the favorable impacts of exercise [38]. It is also important to mention that moderate physical activity significantly increases the anti-pathogenic activity of macrophages, increases the circulation of immune cells, immunoglobulins, and anti-inflammatory cytokines while reducing the possibility of organ damage (such as the lung) due to COVID-induced inflammation [38]. Therefore,

physical exercise is shown to be a non-pharmacological intervention that achieves immune enhancement and reduces the negative effects of the disease [36].

Nevertheless, it is crucial to direct our attention to that significant part of the examined sample (36.4%) that did not exercise before the lockdown and started to, not so much, of course, as those who were already used to exercise regularly. A lower frequency of exercise is imperative to eventually adopt a healthier way of life in general than no physical activity at all. This mainly happened due to the shutdown of almost all businesses which left more free time for people. At the same time, while physical activity is the only way of transportation outside the house, even for an hour, a motive has been provided to a vast part of the population to engage in physical activities and follow a healthier way of living [38].

The recovery time was positively associated with the days of hospitalization, as the more days a patient was hospitalized, the more time they needed to return to their physical condition before infection. The data of many studies support that those patients hospitalized for a long time or subjected to invasive ventilation for a prolonged period displayed respiratory and muscle difficulties, a key factor for their recovery time and the restoration of their previous physical condition to how it was before hospitalization [39].

Together, in the present study we found that preexisting comorbidities, dyspnea on admission, and weight loss in Kg during hospitalization were independent predictors for a longer recovery time after the hospitalization. The days of hospitalization were associated with more significant weight fluctuations, either weight gain or weight loss. In total, 56.3% of the patients displayed weight loss. This study showed that weight loss in Kg is an independent indicating factor for greater needed time for recovery. The severe inflammation caused by the virus and results in the release of much more acute phase proteins disorganizes the metabolism and causes weight loss. Additionally, the decreased food intake mostly connected with loss of appetite due to the disease's symptoms is one more main factor. Another important cause of weight loss is the anxiety brought on the surface due to the disease and bad sleep quality during hospitalization. In addition, immobility due to hospitalization undoubtedly contributes to muscle atrophy, a decrease in adipose tissue, sarcopenia, and, eventually, weight loss. The decreased weight and cachexia due to hospitalization increase the time patients need to return to their state prior to infection [39].

Several limitations need to be noted regarding the present study. A main limitation is that the data-stream provided by self-reporting is not shielded from potential acquiescence response bias. For this reason, the self-reported administered questionnaires were used, measured physical activities that were relevant to older adults over a relatively short period of time before and after hospitalization to minimize reporting errors [40]. In addition, reliability measures were not used in the study, but only self-report questionnaires were used to collect data. Moreover, the physical activity is not the only predictor for obesity but there are other factors as well that should be evaluated in future studies. Furthermore, we do recognize that our study was obviously limited by the small sample size. Even though we aimed to have a larger sample size, the actual response rate was much lower. Nevertheless, this study for the first time evaluated the frequency and type of physical activity among adults previously hospitalized due to COVID-19 in Greece.

#### **5. Conclusions**

We found that two-thirds of the hospitalized patients were overweight or obese. The increased BMI was not associated with the hospitalization time. This study also showed that weight loss in Kg, a pre-existing chronic disease, and dyspnea as a symptom during hospital admission could independently predict a longer recovery time. Two-thirds of the patients used to engage in some form of physical activity before infection. The men were exercising in a higher percentage and more frequently than the women before their hospitalization. Most of the patients that used to exercise before infection supported that the lockdown had boosted their desire to exercise. At a significantly higher percentage than women, men supported that their hospitalization and the restrictive measures did not affect their frequency or desire to exercise after discharge from the hospital. Women in their majority were negatively affected by the lockdown, at a higher percentage than men, regarding their frequency and desire to exercise.

Hence, obesity is a common comorbidity in patients with COVID-19 that was not proven to be associated with recovery time. Physical activity has long-term health benefits in COVID-19 patients given that those with methodical contact with exercise before infection had low levels of dyspnea after their hospitalization and less recovery time. Avoiding a sedentary life and adopting a healthier way of living by engaging in any form of physical activity are proven to positively affect the rehabilitation from the immensely severe COVID-19 disease that requires hospitalization.

**Author Contributions:** Conceptualization, K.I.G., Z.D. and O.S.K.; methodology, K.I.G., O.S.K. and S.K.; formal analysis, S.K.; investigation, S.K. and O.S.K.; writing—original draft preparation, O.S.K.; writing—review and editing, O.S.K.; supervision, K.I.G., Z.D. and O.S.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of University of Thessaly (No. 2800-01/11/2020; approved on 1 November 2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data that support the findings of this study are available on request from the corresponding author, OSK.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

