**Impact of COVID-19 Lockdown on Anthropometric Variables, Blood Pressure, and Glucose and Lipid Profile in Healthy Adults: A before and after Pandemic Lockdown Longitudinal Study**

**José Ignacio Ramírez Manent 1,2, Bárbara Altisench Jané 1,\*, Pilar Sanchís Cortés 2,3, Carla Busquets-Cortés 4,5, Sebastiana Arroyo Bote 4,5, Luis Masmiquel Comas 1,2 and Ángel Arturo López González 2,4,5**


**Abstract:** In December 2019, 27 cases of pneumonia were reported in Wuhan. In 2020, the causative agent was identified as a virus called SARS-CoV-2. The disease was called "coronavirus disease 2019" (COVID-19) and was determined as a Public Health Emergency. The main measures taken to cope with this included a state of lockdown. The aim of this study was to assess how the unhealthy lifestyles that ensued influenced different parameters. A prospective study was carried out on 6236 workers in a Spanish population between March 2019 and March 2021. Anthropometric, clinical, and analytical measurements were performed, revealing differences in the mean values of anthropometric and clinical parameters before and after lockdown due to the pandemic, namely increased body weight (41.1 ± 9.9–43.1 ± 9.9), BMI (25.1 ± 4.7–25.9 ± 4.7), and percentage of body fat (24.5 ± 9.1–26.9 ± 8.8); higher total cholesterol levels, with a statistically significant increase in LDL levels and a reduction in HDL; and worse glucose levels (90.5 ± 16.4–95.4 ± 15.8). Lockdown can be concluded to have had a negative effect on health parameters in both sexes in all age ranges, causing a worsening of cardiovascular risk factors.

**Keywords:** COVID-19; cardiovascular risk factors; lockdown; disease

#### **1. Introduction**

In December 2019, 27 cases of severe pneumonia of unknown cause were reported in the city of Wuhan (Hubei, China), which had in common their appearance in a wholesale market for fish and live animals [1]. On 7 January 2020, the causative agent was identified as a new virus (Coronaviridae), called SARS-CoV-2 [2]. The disease caused by this virus became internationally known as "coronavirus disease 2019" (COVID-19). The most common clinical manifestations were fever, cough, fatigue, and gastrointestinal symptoms. Respiratory or gastrointestinal symptoms could coexist or be found in isolation [3–5]. Depending on individual genetics, ethnic origin, age and geographical location, it has been seen that the clinical manifestations and morbidity and mortality from COVID-19 are different [6–8]. On 30 January 2020, COVID-19 was determined as a Public Health Emergency of International Importance (ESPII) and later, on 11 March 2020, declared a global pandemic by the WHO [5].

**Citation:** Ramírez Manent, J.I.; Altisench Jané, B.; Sanchís Cortés, P.; Busquets-Cortés, C.; Arroyo Bote, S.; Masmiquel Comas, L.; López González, Á.A. Impact of COVID-19 Lockdown on Anthropometric Variables, Blood Pressure, and Glucose and Lipid Profile in Healthy Adults: A before and after Pandemic Lockdown Longitudinal Study. *Nutrients* **2022**, *14*, 1237. https:// doi.org/10.3390/nu14061237

Academic Editors: Dimitrios T. Karayiannis and Zafeiria Mastora

Received: 20 February 2022 Accepted: 11 March 2022 Published: 15 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/).

The rapid spread and severity of the COVID-19 pandemic became a threat to public health, with the lack of effective drugs or vaccines at that time leading governments of more than 100 countries to apply strict measures in their efforts to limit and control the spread of the disease [9,10]. Measures such as a lockdown, quarantine, or isolation of their populations were put in place, in such a way that in April 2020 more than a third of the world's population was under some type of lockdown [11]. In Spain this was established by Royal Decree 463/2020 of March 14, declaring a state of emergency [12].

This state of lockdown had a negative impact on the physical and mental health of the population, with a decrease in physical activity and a significant change in eating patterns at all ages [13–18]. Lifestyle modifications and withdrawal from work, university, or school are all related to boredom and were discovered to cause bingeing or loss of appetite [13,19,20]. A decrease in the consumption of fish, seafood, fruit, and vegetables was found, along with [21] a rise in the consumption of salty and sugary snacks (including desserts, sweets, chips, nuts, crackers, popcorn, peanuts, pistachios, sunflower seeds, etc.) [22,23]. There was also a high prevalence of sleep [24–26] and physical activity disorders [6], which are related to unbalanced nutritional patterns in adults and adolescents [27,28].

Consequently, there was an increase in weight in the world population range from 11.1% to 72.4% during the lockdown period. In Spain, the weight gain reported by patients themselves ranged between 12.8% and 44% [29]. People who put on weight during the lockdown also had a more sedentary lifestyle most of the time - watching television and doing on-screen leisure activities, using smartphones, the internet, or socializing online. This weight gain related to COVID-19 will cause an increased risk of developing metabolic disorders in the population with a previous diagnosis of disease [30], but also in the population who had not suffered from these disorders beforehand [31]. Moreover, it has been observed that the population with previous pathology has a higher risk of becoming severely ill if infected by the virus [31,32].

Our objective was to evaluate how these unhealthy lifestyles influenced different anthropometric parameters, blood glucose levels, lipid profile, and blood pressure in a sample of 6236 workers in Spain, with the aim that if at some future time a similar situation occurs, we would be able to take adequate preventive measures to reduce its side effects on people's health and the development of disease.

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

A prospective study was carried out on 6283 workers in the Balearic Islands and the Valencian Community in companies from different productive sectors, the most represented were working at hotels, construction, commerce, health and public administration, transport, education and the cleaning industry between March 2019 and March 2021. Employees were selected from among those who attended the periodic occupational medical check-ups during those years. Of these, 47 were excluded (19 since they did not agree to participate and 28 since they did not undergo the second medical examination), leaving 6236 finally included in the study (Figure 1).

Inclusion criteria


Anthropometric, clinical, and analytical measurements were performed by the health personnel of the different occupational health units participating in the study, after homogenizing the measurement techniques. To measure weight, expressed in kilograms, and height, expressed in cm, a scale with a measuring rod was used, namely model SECA 700 with a capacity of 200 kg and 50-g divisions, with a SECA 220 telescopic measuring rod with millimetric division and a 60–200 cm interval.

**Figure 1.** Flowchart of participants.

Abdominal waist circumference was measured in cm with a measuring tape (SECA model 20, with an interval of 1–200 cm and millimeter division). The person stood feet together and trunk erect, abdomen relaxed, and upper limbs hanging on both sides of the body. The tape measure was placed parallel to the floor at the level of the last floating rib. Hip circumference was measured with a SECA model 200 tape with a measuring interval of 12–200 cm and millimeter division. The same position was adopted as for waist circumference and the measuring tape was passed horizontally at hip level. Waist/height and waist/hip indices were obtained by dividing waist circumference by height and hip circumference, respectively. The cut-off point for the former was 0.50 while for the latter it was 0.85 for females and 0.95 for males [33].

Blood pressure was measured in the supine position with a calibrated OMRON M3 automatic sphygmomanometer after 10 min of rest. Three measurements were taken at one-minute intervals and the mean of the three was calculated. Blood tests were obtained by peripheral venepuncture after a 12-h fast, sent to reference laboratories, and processed within 48–72 h. Automated enzymatic methods were used for blood glucose, total cholesterol, and triglycerides. Values are expressed in mg/dL. HDL was determined by precipitation with dextran sulphate Cl2Mg, and values are expressed in mg/dL. LDL was calculated using the Friedewald formula (provided triglycerides were less than 400 mg/dL). Values are expressed in mg/dL.

Friedewald's formula:

$$\text{LDL} = \text{total cholesterol} - \text{HDL} - \text{triglycine}/5$$

BMI was calculated by dividing weight by height in meters squared. Obesity was considered to be over 30 [33]. Body fat percentage was determined by bioimpedance using a Tanita model MC-780MA S.

A smoker was considered to be a person who had regularly consumed at least one cigarette/day (or the equivalent in other types of consumption) in the previous month or had stopped smoking less than a year before.

Physical activity was determined by means of the International Physical Activity Questionnaire (IPAQ) [34], a seven-question self-administered questionnaire that assesses the type of physical activity performed in daily life during the previous seven days which was performed at each medical check-up.

#### *2.1. Statistical Analysis*

A descriptive analysis of the categorical variables was carried out by calculating the frequency and distribution of responses for each one. For quantitative variables, the mean and standard deviation were calculated, whereas for qualitative variables, the percentage was calculated. Bivariate association analysis was performed using the X<sup>2</sup> test (with correction of Fisher's exact statistic when conditions so required) and Student's *t* test for independent samples. For multivariate analysis, binary logistic regression was used with the Wald method, with calculation of the odds ratio and the Hosmer-Lemeshow goodnessof-fit test. The statistical analysis was performed with the SPSS 27.0 (IBM, New York, USA) program, with an accepted statistical significance level of 0.05.

#### *2.2. Ethical Considerations and Aspects*

The study was approved by the Clinical Research Ethics Committee of the Balearic Islands Health Area no. IB 4383/20. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 2013 Declaration of Helsinki. All patients signed written informed consent documents before participating in the study.

#### **3. Results**

Lockdown began for all participants on 15 March 2020, and post-lockdown anthropometric measurements were carried out by the same health personnel from the different occupational health units. Of the 6283 workers, 51.9% were female and 48.1% were male, constituting a proportional representation of both sexes. Participant characteristics, including anthropometric characteristics, physical activity, and smoking before and after lockdown, are all summarized in Table 1. The number of participants was the same each year, being a total of 6236 Spanish workers.

Table 1 shows the statistically significant differences in the mean values of anthropometric and clinical parameters before and after lockdown due to the COVID-19 pandemic.

An increase in body weight and therefore also in BMI can be seen in the population studied; as well as an increase in the percentage of body fat, and hip and waist circumference; being this the one with the highest increase. Regarding clinical parameters, the elevation of total cholesterol levels stands out, with a statistically significant increase in LDL levels (going from mean values of 117.4 mg/dL to 131 mg/dL) and a reduction in HDL levels. Triglyceride values also rose.

Comparing blood pressure levels, during lockdown there was a tendency to higher diastolic blood pressure levels. Systolic blood pressure was also affected. Glucose levels increased during lockdown such as the previously analyzed parameters.

The percentage of people who increased their smoking habit during lockdown was 2%, while an 11% decrease their physical activity, causing an elevation of 4.1% in overweight and 2.5% in obesity.

Comparing the mean values of the anthropometric, clinical, and laboratory parameters in different ranges of years according to whether this change occurred before or during the COVID-19 pandemic lockdown, it is possible to observe a tendency to a worsening of the values of the different parameters analyzed, as shown in Table 2.


**Table 1.** Characteristics of the population.

SBP: systolic blood pressure; DBP: diastolic blood pressure; HDL: high density lipoproteins; LDL: low density lipoproteins.

Figure 2 shows the graphs of the parameters related to cardiovascular risk factors. It can be observed that over the years, the trend in all of them was towards an increase in mean values, with an exponential, statistically significant increase during the year of the pandemic due to COVID-19. It can be seen an increase in BMI levels, percentage of fat mass, laboratory values of glucose and total cholesterol, as well as higher blood pressure levels. The alteration encountered in mean blood pressure values was less pronounced compared to the rest of the parameters analyzed, which underwent greater changes, causing an increase in cardiovascular risk in the population studied.


**Table 2.** Changes in anthropometric, clinical, laboratory, and healthy habit variables in different pre-COVID and COVID years.

**Figure 2.** Changes in cardiovascular risk factors in 2018, 2019, and 2020: BMI, waist, lean mass, glucose, blood pressure, and total cholesterol.

Comparing the different cardiovascular risk parameters that had worse mean values after COVID (and considering an increased risk when the category changed to a worse one according to the corresponding risk factor), it can be observed in Figure 3 that by analyzing the parameters according to relative risk, with a 95% confidence interval, and odds ratio, the changes in the mean values of triglyceride levels were not related to the other parameters analyzed. In the other variables analyzed, blood pressure levels were those with the highest RR (1.261–1.306) along with OR (1.283). The rest of the parameters can be seen in the figure.

**Figure 3.** Relative Risk (RR) of patients who had worsened cardiovascular risk factors in the 2019–2020 season when compared with the 2018–2019 season. Patients were considered to have worsened when they changed to a worse category in the corresponding risk factor: BMI (normal, overweight, obesity); glycemic status (normal, prediabetes, diabetes); blood pressure status (normal, pre-AHT, AHT1, AHT2); waist (normal, high); fat mass (normal, high); total cholesterol (normal, high); LDL (normal, high, very high); HDL (normal, low); and TG (normal, high).

The parameters of BMI, glycemic status, blood pressure, waist circumference, and percentage of fat mass were analyzed in relation to age, sex, BMI, glycemic status, and blood pressure levels. It can be observed in Table 3 that upon analyzing by age variable, there is a statistically significant worsening of the values of BMI, glycemic status, blood pressure, and waist circumference, which does not occur in the percentage of fat mass, although this relationship is not statistically significant with age. In the group over 50 years of age, the RR is higher.

When analyzing by sex, the BMI values obtained are not considered statistically significant, although there is statistical significance for the rest of the variables studied. It should be noted that, in males, there is no causal relationship with an increase in percentage of fat mass, while in females there is.

There is no clear association between overweight or obesity and glucose levels, with a RR < 1 and a *p*-value of 0.241. In terms of BMI, despite having a causal relationship with the percentage of fat mass with a RR > 1, the values obtained have a *p*-value of 0.704, therefore this relationship and the levels obtained could be caused by other factors as they are not statistically significant. Changes in BMI are not always related to a higher percentage of fat mass; patients with high muscle mass and a low percentage of fat mass, for instance, could also have altered BMI values since this parameter does not differentiate between muscle or fat.

By analyzing glycemic status, workers with prediabetes did reveal a direct relationship with alterations in percentage of fat mass, but not with blood pressure levels or waist circumference. When relating baseline blood glucose levels to BMI, despite observing a causal relationship with a RR > 1, the values obtained were not statistically significant (*p*-value 0.065) and would therefore not be interpretable for this reason.

Blood pressure levels do not have a statistically significant relationship with glycemic status. Although the relationship between blood pressure and waist circumference and percentage of fat mass have been statistically related to a *p*-value < 0.001, in the group of workers with type 1 hypertension, there was no increase in blood glucose levels as there was no direct association with the anthropometric parameters analyzed.

When analyzing according to age of participants, a statistically significant increase in total cholesterol levels was observed, with a greater association in workers aged 40–50 years) with a *p*-value of < 0.001. There are also statistically significant differences with *p*-value < 0.001 in LDL cholesterol levels with a lower association in workers over 50 years old), as can be seen in Table 4.

When the relationship between age, HDL cholesterol, and triglyceride levels is analyzed, the differences are not statistically significant with a *p*-value of > 0.05, so the changes in the parameters could be due to other factors and not by age.

According to the sex of the patient studied, a statistically significant worsening stands out with a direct association for the clinical parameters of total cholesterol, HDL cholesterol, and LDL cholesterol in both males and females, but not for triglyceride levels where the differences between males) and females are not statistically significant.

In the population studied that was overweight or obese, the increased analytical values of total cholesterol and the decrease in HDL cholesterol were statistically significant (*p*-value < 0.05) even though in the group of overweight patients, there was no direct association with HDL cholesterol values. No statistically significant relationship was found with LDL cholesterol or triglyceride levels in the overweight and obese population.

The glycemic level of the workers studied was found to have a direct, statistically significant association with a *p*-value of < 0.05 for HDL cholesterol and LDL cholesterol, but not with total cholesterol or triglyceride levels. In the group of patients with prediabetes, no direct relationship with changes in LDL cholesterol was observed.

If we analyze according to blood pressure levels, patients with normal blood pressure or type II hypertension are associated with a higher risk (RR: 1.106; 95% CI 1.081–1.132) and (RR: 1.053; 95% CI 0.620–1.787) for total cholesterol and a RR: 1.005; 95% CI 0.987–1.024 and RR: 1.003; 95% CI 0.955–1.053 for HDL cholesterol, respectively; but with a non-statistically significant association for LDL cholesterol levels with a *p*-value of 0.072.

Regarding triglyceride levels, according to blood pressure levels, no direct relationship was found in patients with hypertension, but was found in normotensive patients with a *p*-value of 0.036.

In relation to physical activity, the changes caused by lockdown, measured through the IPAQ questionnaire, can be seen in Table 5. A decrease in physical activity is observed in both sexes for the group that exercised before lockdown, as well as an increase in sedentary lifestyle in groups that did not perform physical activity before lockdown on a regular basis.

The decrease in physical activity is statistically significant for both sexes, with a *p*-value < 0.0001 compared to the time before the pandemic.

*Nutrients* **2022**, *14*, 1237

**Table 3.** Relative risk of patients with increased BMI, waist, fat mass, glycemic status, and hypertension status divided into categories of age, sex, BMI, glycemic status, and hypertension status. Patients were considered to have worsened when they changed to an upper category in the corresponding risk factor: BMI (normal, overweight, obesity); glycemic status (normal, prediabetes, diabetes); blood pressure status (normal, pre-AHT, AHT1, AHT2); waist (normal, high); fat mass (normal, high, very high); total cholesterol (normal, high); LDL (normal, high); HDL (normal, low); TG (normal, high).


**Table 4.** Relative Risk of patients with increased total cholesterol, LDL, TG and a decrease in HDL levels, divided by categories of age, sex, BMI, glycemic status and hypertension status. Patients were deemed to have worsened when they changed to an upper category in the corresponding risk factor: BMI (normal, overweight, obesity); glycemic status (normal, prediabetes, diabetes); blood pressure status (normal, pre-AHT, AHT1, AHT2); waist (normal, high); fat mass (normal, high, very high); total cholesterol (normal, high); LDL (normal, high); HDL (normal, low); and TG (normal, high).



**Table**

**4.**

*Cont.*



**Table 5.** Percentage of physical activity, before and during lockdown in men and women separated into groups according to whether or not they previously performed physical activity.

#### **4. Discussion**

The global pandemic caused by COVID-19 has had a great impact on health population [35]. Not only due to the infection caused by the virus, which has left complications and consequences of the disease from which some people are yet to fully recover, but also due to the pathology derived from lockdown, social distancing, and isolation, in which chronic diseases have worsened [36–39]. These measures taken by governments to protect public health have produced a psychological impact on the population that has caused overeating, a more sedentary lifestyle, and modification of several anthropometric, clinical, and laboratory health parameters affecting all body systems [40,41].

In this study, we objectify the changes produced in the population due to lockdown, in a population of Spanish workers. Although it is true that, in recent decades, the population has had a tendency to obesity and overweight due to a sedentary lifestyle [13,16], what can be seen in this study is that, during lockdown, many of the parameters that influence cardiovascular risk were affected, such as obesity, alcoholism, and a sedentary lifestyle, changing their values, leading to a greater risk of suffering from cardiovascular diseases. Our results are similar to those published by other authors from different countries such as Lithuania, China, Korea, Israel, UK, amongst others [40–45].

In the study carried out by Paltrienteri et al., the changes produced during lockdown in relation to physical activity, diet, alcohol consumption, and tobacco were studied through a self-administered survey. The results showed there had been a reduction in physical activity without a change in diet [46]. In our study, these modifications were studied through anthropometric, clinical, and laboratory parameters, and their changes could be observed with a statistically significant increase in both obesity and overweight, as seen in our results.

A simultaneous to the increase in cardiovascular risk, as detailed in Table 1, the 4.1% increase in the rate of overweight and obesity, as well as the 11% decrease in physical activity and 2% increase in smoking brought about the development of other diseases, both acute and chronic, which have modified the health status of the population and increased morbidity and mortality from other causes, not only due to infection with COVID-19. These consequences were also pointed out by Palmer et al. [47,48].

The results of our study reveal statistically significant differences when comparing clinical, laboratory, and anthropometric parameters in a population of workers due to a lockdown. The increase in body weight and therefore BMI is a consequence of the dietary habits and sedentary lifestyle of the population during this period [49,50]. An increase in percentage of fat mass and waist and abdominal perimeter was also observed [42,44,45]. There are studies in the literature that compare body weight during and after lockdown, such as the study by Blautani S, et al. which shows that it was not the entire population that suffered an increase in body weight, with 18.2% of the population in their study losing weight during lockdown. At the end of lockdown, those who had gained weight continued to gain, so it is likely that their health effects associated to body changes will persist over time [51] if lifestyle modifications are not made [52].

Not only were anthropometric parameters affected, but alterations were also found in biochemical parameters. In the lipid profile, for instance, an increase in total cholesterol levels was detected, which corresponded to an additional increase in LDL cholesterol levels and decrease in HDL cholesterol levels, with statistically significant differences. These results are similar to those published in other studies, although their sample sizes were much smaller than ours [40,53].

Plasma glucose levels deteriorated, probably in connection with the increased rate of obesity, overweight, and decrease in physical exercise [54]. In patients with blood glucose levels in the range of diabetes mellitus, a statistically significant decrease of LDL cholesterol levels was detected, although this was not statistically significant in patients with prediabetes values, in whom there was no clear relationship with changes in LDL cholesterol values. At a general level, there was an increase of triglyceride levels with statistically significant results. The combined increase in serum triglycerides points to the role of a variation in eating habits and reinforces the need not to attenuate attention to an adequate lifestyle program, with regular physical activity and a correct dietary approach, even when successful pharmacological treatment is ongoing [54]. A high-carbohydrate diet is known to raise fasting triglyceride levels more than a high-fat diet, which is also related to greater mortality [55]. Our results are similar to those obtained in previous studies [40,53].

Regarding blood pressure levels, lockdown also caused a deterioration in people who were not previously hypertensive, probably due to their lifestyle during these months and the worsening of the population's health status owing to a change in dietary habits and physical activity. In the literature consulted, we have found very few studies that refer to changes in blood pressure during lockdown due to COVID-19. However, the studies published are similar to our results [40,56].

The lockdown adversely affected multiple risk factors for disease, especially cardiovascular disease. Plasma concentrations for LDL and HDL cholesterol, respectively increased and decreased. Concurrently, blood glucose concentrations and blood pressures increased. According to the increases in ratio of waist to hip circumferences, these effects were associated with increased central obesity [57]. Notably, low HDL cholesterol, a large waist circumference, hyperglycemia, hypertension, and hypertriglyceridemia are components of metabolic syndrome, a global measure of risk for cardiovascular disease and developing type 2 diabetes mellitus [58]. According to the latest guidelines, the metabolic syndrome is described as a set of analytical and anthropometric alterations, in which the patient must have at least three altered parameters in order to be diagnosed of metabolic syndrome. These parameters are: waist circumference in men ≥102 cm and ≥88 cm in women, triglyceride values ≥150 mg/dL or being on pharmacological treatment, HDL levels <40 mg/dL in men and <50 mg/dL in women. Blood pressure values ≥130/85 mmHg or being on pharmacological treatment with antihypertensives and fasting blood glucose levels ≥100 mg/dL or being on antihyperglycemic treatment [59].

According to the results obtained in our study, which have been explained previously, a global worsening of these parameters could be detected, secondary to the change in lifestyle caused by lockdown. The increased metabolic syndrome has been able to develop the appearance of different cardiovascular and metabolic complications that have caused an increase in morbidity and mortality, as well as an increased risk of COVID-19 infection with greater potential for severity as has been seen in recent studies [60–62] as shown in the study carried out by Li B et al., which shows how the population with cardiovascular risk factors has a higher risk of severe infection by COVID-19 [63].

Periodic medical check-ups, in the case of our study, of Spanish workers, has allowed the detection of these alterations and the possibility of applying preventive measures to avoid the development of diseases in the future as well as complications in the event of infection by COVID-19 [60,61].

In the different tables of results, the worsening of blood pressure levels, glycaemia, waistabdominal perimeter as well as analytical levels of triglycerides and HDL cholesterol are observed. Early application of preventive measures in the different altered parameters could prevent the development of metabolic syndrome and its possible complications [59,62,64].

Despite being known as a syndrome, there is no single approach, since the objective is to apply preventive or therapeutic measures individually according to the altered parameters [65,66] in each individual, always starting with lifestyle modifications, a factor that has been greatly influenced by lockdown due to the increase in sedentary lifestyle, with a decrease in physical activity [67–69] and the complications driven by COVID-19.

The aim of treating or preventing these altered analytical and anthropometric parameters produced by the state of lockdown would be to prevent the development of cardiovascular diseases and therefore reduce cardiovascular risk at the population and its potential morbidity and mortality. In the study conducted by Yangjing X, et al., people with established cardiovascular disease or altered cardiovascular risk parameters are shown to have a worse prognosis when faced with COVID-19 infection [70].

In our study, many of these cardiovascular risks are seen to have modified their levels during lockdown, leading to an increase in cardiovascular risk. Also, the fact of decreasing physical activity and increasing sedentary lifestyle is associated with an increase and worsening in the different parameters studied and also with a tendency to obesity and therefore to the complications derived from it, causing an increase in cardiovascular risk factors in the population [71], as has also been seen in other studies such as the ones from Hendren et al. or Hu L et al. [72,73].

In our study, we found a significant decrease in physical activity both in men and women, being higher in women. Our results are consistent with other published studies [74,75], and differ from those obtained by Castañeda-Barbarro et al.'s. in which a greater decrease in physical exercise in men is found [76].

Regular physical activity helps reduce cardiovascular risk [68], by reducing the percentage of fat mass and improving laboratory and clinical values such as blood pressure, insulin resistance ... helping in the prevention of developing diseases derived from an unhealthy lifestyle and preventing cardiovascular diseases [69,77].

With all this, it can be stated that the COVID-19 pandemic has increased the risk of developing pathologies derived from lifestyle modifications, in addition to raising the risk of COVID-19 infection by altering parameters that increase the risk of illness [54,69,78].

#### **5. Strengths and Limitations**

Several studies compare the effects of the pandemic and COVID-19 infection with changes in obesity and overweight parameters, as well as cardiovascular risk factors and other pathologies, but we have not found any study in which so many parameters are compared in the same population as in our study.

Further, none of the studies that present the evaluation of the different parameters separately have a sample size such as ours, with 6283 patients.

The limitations found in this study are the fact that it was carried out in a specific geographic area, with a Caucasian working population, over a certain period of time, which could limit the generalization of the results to other areas where lifestyles may be different.

Selection bias is another limitation of our study, since it is limited to workers who voluntarily attended company medical examinations during those years.

Therefore, the results do not apply to other populations, and specific studies would have to be carried out.

#### **6. Conclusions**

Health behaviors have been negatively affected during lockdown, leading to an increase in sedentary behavior in all age groups, an unhealthy diet, and, therefore, associated with weight gain, as well as an increased consumption of tobacco.

With the parameters and results described above, it can be concluded that the months of lockdown caused a statistically significant deterioration of several health parameters due to increased sedentary behavior in a similar way in males and females in all age ranges, although the over 40-year-old group is the one where the worst values of the variables analyzed were observed, causing an increase in the magnitude of multiple risk factors for

cardiovascular disease and the appearance of new pathologies that have resulted in an increase in morbidity and mortality due to all causes.

**Author Contributions:** Conceptualization: J.I.R.M., B.A.J. and Á.A.L.G. Data collection and analysis: P.S.C., C.B.-C., S.A.B. and L.M.C. Methodology: J.I.R.M., B.A.J. and Á.A.L.G. Draft: B.A.J., Á.A.L.G. Revision: J.I.R.M., B.A.J., P.S.C., C.B.-C., S.A.B. and L.M.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

**Institutional Review Board Statement:** The study was carried out after the authorization of the Ethical Committee of the Balearic Islands, with the prior informed consent of the study subjects and following the norms of the Helsinki Declaration. The confidentiality of the subjects included will be guaranteed at all times in accordance with the provisions of the Organic Law 3/2018, of December 5, on the Protection of Personal Data and guarantee of digital rights and Regulation (EU) 2016/679 of the European Parliament and the Council of 27 April 2016 on Data Protection (RGPD).

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

**Data Availability Statement:** Data available on request due to restrictions (privacy or ethical).

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

#### **References**


## *Article* **Vitamin A Plasma Levels in COVID-19 Patients: A Prospective Multicenter Study and Hypothesis**

**Phil-Robin Tepasse 1,\*,†, Richard Vollenberg 1,†, Manfred Fobker 2, Iyad Kabar 1, Hartmut Schmidt 1, Jörn Arne Meier 1, Tobias Nowacki <sup>1</sup> and Anna Hüsing-Kabar <sup>1</sup>**


**Abstract:** COVID-19 is a pandemic disease that causes severe pulmonary damage and hyperinflammation. Vitamin A is a crucial factor in the development of immune functions and is known to be reduced in cases of acute inflammation. This prospective, multicenter observational cross-sectional study analyzed vitamin A plasma levels in SARS-CoV-2 infected individuals, and 40 hospitalized patients were included. Of these, 22 developed critical disease (Acute Respiratory Distress Syndrome [ARDS]/Extracorporeal membrane oxygenation [ECMO]), 9 developed severe disease (oxygen supplementation), and 9 developed moderate disease (no oxygen supplementation). A total of 47 age-matched convalescent persons that had been earlier infected with SARS-CoV-2 were included as the control group. Vitamin A plasma levels were determined by high-performance liquid chromatography. Reduced vitamin A plasma levels correlated significantly with increased levels of inflammatory markers (CRP, ferritin) and with markers of acute SARS-CoV-2 infection (reduced lymphocyte count, LDH). Vitamin A levels were significantly lower in hospitalized patients than in convalescent persons (*p* < 0.01). Of the hospitalized patients, those who were critically ill showed significantly lower vitamin A levels than those who were moderately ill (*p* < 0.05). Vitamin A plasma levels below 0.2 mg/L were significantly associated with the development of ARDS (OR = 5.54 [1.01–30.26]; *p* = 0.048) and mortality (OR 5.21 [1.06–25.5], *p* = 0.042). Taken together, we conclude that vitamin A plasma levels in COVID-19 patients are reduced during acute inflammation and that severely reduced plasma levels of vitamin A are significantly associated with ARDS and mortality.

**Keywords:** COVID-19; vitamin A; retinol; retinoic acid; ARDS; pneumonia; pandemic; SARS-CoV-2; inflammation

#### **1. Introduction**

Coronavirus disease 2019 (COVID-19) is a novel infectious disease that has been spreading worldwide [1]. The clinical manifestation of COVID-19 can range from asymptomatic infection to critical illness with severe pneumonia, respiratory failure, and death [2]. Worse clinical outcomes are related to dysregulated immune responses in the host, leading to the uncontrolled release of proinflammatory cytokines such as interleukin-6 (IL-6). This cytokine storm mediates the progression of lung damage and respiratory failure in a relevant number of cases [3]. The understanding of host parameters leading to the susceptibility of immune dysregulation is incomplete. A standard therapy has not yet been established. To date, the RNA-polymerase inhibitor remdesivir and the immunosuppressive corticosteroid dexamethasone are the only Food and Drug Administration (FDA) approved drugs for COVID-19 therapy with demonstrated effects on mortality and

**Citation:** Tepasse, P.-R.; Vollenberg, R.; Fobker, M.; Kabar, I.; Schmidt, H.; Meier, J.A.; Nowacki, T.; Hüsing-Kabar, A. Vitamin A Plasma Levels in COVID-19 Patients: A Prospective Multicenter Study and Hypothesis. *Nutrients* **2021**, *13*, 2173. https://doi.org/10.3390/nu13072173

Academic Editors: Dimitrios T. Karayiannis and Zafeiria Mastora

Received: 27 May 2021 Accepted: 20 June 2021 Published: 24 June 2021

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

**Copyright:** © 2021 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/).

disease outcome [4,5]. Despite certain therapeutic approaches, the lethality of hospitalized, mechanically ventilated patients remains high [6].

Vitamin A is of special interest in the field of infectious diseases, especially for pulmonary infections. It is crucial for the development of normal lung tissue and tissue repair after injury due to infection [7]; therefore, it may play a role in recovery after severe COVID-19 pneumonia. Vitamin A has immune regulatory functions [8] and positively affects both the innate and adaptive immune cell response [9,10]. Malnutrition leads to relevant incidences of vitamin A deficiency worldwide. Vitamin A deficiency can disrupt vaccine-induced antibody-forming cells and negatively influences immunoglobulin development in the upper and lower respiratory tract [11]. Several studies revealed increased risks of severe illness due to respiratory tract infections in vitamin A-deficient individuals, whereas vitamin A supplementation can reduce the risks of severe illness and death, as was shown for children with measles and in influenza pneumonia in mice models [12–14]. The occurrence of severe infections and inflammation can also negatively affect vitamin A status, and several mechanisms such as urinary loss of vitamin A [15], decreased vitamin A hepatic mobilization [16], and reduced intestinal vitamin A absorption [17] during infection have been described.

Data concerning vitamin A plasma levels in COVID-19 patients are lacking. Therefore, this study aimed at characterizing vitamin A plasma levels in acute COVID-19 and analyzed the association of plasma levels with disease severity and outcome. Vitamin A plasma levels were found to be significantly reduced in COVID-19 patients during the acute phase of disease compared to plasma levels in convalescent patients, and a reduction in vitamin A levels correlated significantly with an increase in inflammatory parameters. Critically ill patients showed lower vitamin A levels compared to moderately ill patients and severely reduced vitamin A levels were associated with ARDS and mortality. Further research is necessary to investigate the role of vitamin A as a possible therapeutic agent for COVID-19.

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

#### *2.1. Participant Selection and Patient Samples*

This multicenter, prospective observational cross-sectional study included 40 hospitalized patients with laboratory-confirmed SARS-CoV-2 infection (nasopharyngeal swab and test by polymerase-chain reaction), admitted to the University Hospital Muenster and Marien-Hospital Steinfurt in Germany between March and June 2020. Details of medical history and laboratory data were collected. Blood from hospitalized patients was collected during acute phase of disease. Disease severity was defined as critical (presence of acute respiratory distress syndrome [ARDS]; *n* = 22), severe (requiring oxygen supplementation; *n* = 9) or moderate (neither ARDS was present nor oxygen supplementation required; *n* = 9). ARDS was diagnosed according to the Berlin definition (bilateral opacities on chest radiograph, exclusion of other causes of respiratory failure) [18]. COVID-19 patients were categorized according to their condition at the time of blood collection. One blood sample per patient was taken within the first 24 h after admission.

Additionally, 91 individuals with laboratory-confirmed SARS-CoV-2 infection who had recovered from infection were contacted for donation of convalescent plasma in our outpatient clinic. The blood sample was taken after recovery from disease. Of these, 47 were manually selected to match the age and gender of inpatients (Supplementary Figure S1). This group of patients had only moderate symptoms, and hospitalization was not necessary (Supplementary Figure S2).

Plasma samples were obtained from each participant after they provided informed consent. The Ethics Committee of Muenster University approved the current study (local ethics committee approval AZ 2020-220-f-S and AZ 2020-210-f-S), and the procedures were in accordance with the Helsinki Declaration of 1975 as revised in 1983. None of the patients received antiviral, experimental, or immunosuppressive therapies. Plasma samples were protected from light and frozen (−80 ◦C) until measurement.

#### *2.2. Vitamin A Measurement*

Vitamin A (both free/unbound vitamin A and vitamin A bound to retinol binding protein [RBD]) was assayed on EDTA-plasma using a commercially available highperformance liquid chromatography kit (Chromsystems, Munich, Germany) following the manufacturer's instructions. Vitamin A levels were given in mg/L. For further analysis, participants were divided in to two groups (Group 1: vitamin A plasma levels ≥ 0.2 mg/L; Group 2: vitamin A plasma levels < 0.2 mg/L) following international guidelines on vitamin A deficiency from the World Health Organization (WHO, clinically relevant nutritional Vitamin A deficiency defined as Vitamin A plasma levels < 0.2 mg/L) [19].

#### *2.3. Laboratory Measurements and Validation of the Clinical Status*

Clinical laboratory assessment included complete blood count and levels of D-dimer, creatinine, C-reactive protein (CRP), albumin, lactate dehydrogenase (LDH), pseudocholinesterase (PCHe), and alanine aminotransferase (ALT). SAPS II (Simplified Acute Physiology Score [20]) was determined on the day of laboratory measurement and used to characterize the physiological conditions of hospitalized patients.

#### *2.4. Data Analysis/Statistics*

For continuous variables, we report the median with the interquartile range, and values were compared using the Mann–Whitney U (Wilcoxon) test. For categorical variables, we report absolute numbers and percentages, and values were compared with Chi-square tests of association or Fisher's exact tests. A Kruskal–Wallis test was conducted to compare more than two groups. To compare subgroups, the Bonferroni correction post hoc test was performed when variance was equal (Levene's test), and the Games-Howell test was performed when variance was different. The Pearson correlation coefficient was determined to analyze the correlation of vitamin A levels with clinical laboratory parameters. Univariable logistic regression analysis was conducted to determine the association of severely reduced vitamin A plasma levels with the risk of acute respiratory distress syndrome (ARDS) and mortality due to COVID-19 in hospitalized patients (*n* = 40). All the tests were two-tailed, and *p* < 0.05 was considered to indicate a statistically significant difference. All the statistical analyses were performed using SPSS (Version 26IBM Corp., Armonk, NY, USA).

#### **3. Results**

#### *3.1. Cohort Characteristics*

Most participants were male (77.8–100%) in all groups. The median interval between symptom onset and sample collection in hospitalized patients was 12 days (IQR 8, 3–17), in outpatients 52 days (IQR 40–75). The median age was 50–58 years and did not significantly differ between groups, nor too did the interval from the first symptom to the collection of the blood sample. Preexisting diseases were prevalent almost exclusively in the group of critically ill patients. SAPS II differed significantly between hospitalized patients, as did inflammatory and blood count parameters (for example C-reactive protein and lymphocyte count) (Table 1). The correlations of the subgroups are shown in Supplementary Figure S3.

#### *3.2. Correlation of Vitamin A Plasma Levels with Laboratory Parameters*

In the overall study group (*n* = 87), reduced vitamin A plasma levels correlated significantly with increased levels of the inflammation markers C-reactive protein (*r* = −0.54, *p* < 0.001) and ferritin (*r* = −0.45, *p* < 0.001). Reduced absolute (*r* = 0.4, *p* < 0.001) and relative lymphocyte counts (*r* = 0.43, *p* < 0.001), reduced albumin levels (*r* = 0.65, *p* < 0.001), and elevated lactate dehydrogenase levels (*r* = −0.53, *p* < 0.001) correlated significantly with reduced vitamin A plasma levels. Elevated levels of liver markers (AST: *r* = −0.22, *p* < 0.05; gamma-GT: *r* = −0.29, *p* < 0.01) also correlated significantly with reduced vitamin A levels. Reduced vitamin A levels were associated with reduced pseudocholinesterase (PCHe) levels as a parameter of the liver synthetic capacity (*r* = 0.53, *p* < 0.001) (Figure 1).

**Table 1.** Cohort characteristics; differences were calculated via the Kruskal–Wallis test; BMI = body mass index, SAPS II = Simplified Acute Physiology Score II; IQR = interquartile range; LDH = lactate dehydrogenase; PCHe = pseudocholinesterase, ALT = alanine aminotransferase; abs. = absolute; n.d. = not defined.


*3.3. Vitamin A Plasma Levels in COVID-19 Patients*

Vitamin A plasma levels differed significantly depending on disease severity. Hospitalized patients of all groups (mild, moderate, severe, and critical disease) revealed significantly reduced vitamin A plasma levels compared to convalescent outpatients (*p* < 0.01 to *p* < 0.001). In hospitalized patients, the vitamin A plasma levels of critically ill patients were significantly reduced compared to patients with moderate disease (*p* < 0.05). Patients with severe disease also had reduced plasma levels compared to patients with moderate disease, but this finding did not reach statistical significance (median: 0.32 vs. 0.48 mg/L) (Figure 2).

**Figure 1.** Correlation of vitamin A plasma levels with laboratory parameters.

**Figure 2.** Vitamin A plasma levels in patients with moderate, severe, and critical disease, and in convalescent patients; the Wilcoxon rank test showed significant differences (*p* < 0.001). Subgroups were tested for differences in the Bonferroni correction (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001). Mann– Whitney U test was used to compare groups (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001).

#### *3.4. Association of Severely Reduced Vitamin A Plasma Levels with the Development of ARDS and Mortality*

As described in the methods section, Vitamin A plasma levels <0.2 mg/L were considered clinically relevant reduced. In the overall study group, 14% (*n* = 11/87) of participants revealed clinically relevant reduced vitamin A plasma levels. Looking at different patient

groups, 41% (*n* = 9/22) of critically ill, 11% (*n* = 1/9) of moderately ill, and 11% (*n* = 1/9) of severely ill patients had vitamin A plasma levels <2 mg/L, while none of the participants in the convalescent group had clinically relevant reduced vitamin A levels (0/47) (Figure 3).

**Figure 3.** Percentages of patients with vitamin A plasma levels <2 mg/L in different patient groups: moderate disease (*n* = 9), severe disease (*n* = 9), critical disease (*n* = 22), and convalescent patients (*n* = 47). Representation of significances according to Bonferroni correction (\* *p* < 0.05).

After dividing the hospitalized patients into two groups (Group 1: Vitamin A < 2 mg/L; Group 2: Vitamin A ≥ 2 mg/L), Group 1 patients (*n* = 11) revealed significantly higher C-reactive protein levels (median 13.9 vs. 6 mg/dL, *p* = 0.03), lower absolute lymphocyte counts (median 0.69 Tsd/μL vs. 1.13 Tsd/μL, *p* = 0.01), and lower relative lymphocyte counts (median 9.4% vs. 15.2%, *p* = 0.049) compared to Group 2 (*n* = 29). Patients in Group 1 developed ARDS much more frequently (*p* = 0.038), and mortality was significantly higher (*p* = 0.047) (Table 2). Logistic regression analysis revealed significant associations of clinically relevant reduced vitamin A plasma levels with the development of ARDS (OR = 5.54 [1.01–30.26] *p* = 0.048) and with mortality (OR 5.21 [1.06–25.5], *p* = 0.042) in the hospitalized patient group (*n* = 40).

**Table 2.** Clinical outcomes and inflammation parameters based on vitamin A plasma levels; differences were calculated using the Mann–Whitney U test or Fisher's exact test (both 2-tailed); SAPS II = Simplified Acute Physiology Score II; BMI = body mass index; ARDS = acute respiratory distress syndrome; IQR = interquartile range; LDH = lactate dehydrogenase; PCHe = pseudocholinesterase, ALT = alanine aminotransferase.



**Table 2.** *Cont.*

#### **4. Discussion**

This study analyzed associations of vitamin A plasma levels with inflammatory parameters, disease severity, and outcomes in moderately, severely, and critically ill COVID-19 patients during the acute phase of the disease and in comparison to convalescent patients.

Age, gender, and time from disease onset to collection of blood sample adequately matched and did not differ significantly between hospitalized groups in the acute phase of the disease. Our cohort characteristics showed significantly higher levels of inflammatory markers (CRP, IL-6, ferritin) and lower lymphocyte counts in patients with critical disease cases compared to moderately and severely ill and convalescent patients. These data are in line with other studies as both increased inflammatory parameters and decreased lymphocyte counts are well-defined markers of active disease and predictive of a severe course of COVID-19 [21]. D-Dimer is also an established predictive marker [22] and its levels were found to be significantly increased along with disease severity in our study cohort.

This study revealed significantly decreased vitamin A plasma levels in the acute phase of moderately, severely, and critically ill patients compared to convalescent patients after recovery from acute disease. In the acute phase, critically ill patients had significantly lower vitamin A plasma levels compared to moderately ill patients. In the overall study cohort, correlation analysis provided evidence that higher inflammatory parameters such as C-reactive protein, ferritin, and albumin significantly correlated with reduced vitamin A plasma levels. This finding supports existing data showing reduced vitamin A plasma levels due to several mechanisms such as urinary loss [15], decreased hepatic mobilization [16], and reduced absorption [17] during infection and acute inflammatory conditions. Moreover, lymphopenia, an established disease activity marker and predictor of worse outcomes in COVID-19 patients [21], significantly correlated with reduced vitamin A plasma levels. Reduced serum albumin levels [23] and elevated LDH levels [24], also established predictors of severe disease and worse outcome, were both significantly correlated with lower vitamin A plasma levels. Liver damage is often found in severely ill COVID-19 patients [25]. Elevated levels of liver enzymes also correlated significantly with reduced vitamin A levels. Interestingly, a reduction in the levels of pseudocholinesterase (PChE), a parameter for liver synthetic capacity, correlated with reduced vitamin A levels. The

association with decreased hepatic vitamin A mobilization during acute inflammation remains speculative but conceivable and needs further analysis.

Clinically relevant decreased vitamin A plasma levels (defined as <2 mg/L following the WHO definition for vitamin A deficiency [19]) were almost exclusively found in the acute phase of disease in critically ill patients. To analyze the correlation of severely reduced vitamin A plasma levels with disease outcome, we divided the hospitalized patient cohort in two groups (patients with vitamin A levels ≥2 and <2 mg/L). This subgroup analysis showed that patients with vitamin A plasma levels <2 mg/L developed ARDS at a significantly higher rate (*p* < 0.05), and mortality was also much higher in this group (*p* < 0.05). Levels of C-reactive protein were higher and lymphopenia more distinctive, indicating a more severe disease course [21]. These results provide evidence that the reduction of vitamin A plasma levels in COVID-19 is dependent on disease severity and that plasma levels are reduced due to acute infection and inflammation rather than preexisting vitamin A deficiency and malnutrition, as convalescent patients had normal levels of vitamin A in plasma. Furthermore, malnutrition and vitamin A deficiency are rare in Germany. Nonetheless, vitamin A deficiency due to malnutrition is still one of the most prevalent micronutrient deficiencies worldwide, affecting approximately one-third of preschool-age children, especially in underdeveloped countries. Vitamin A deficiency is associated with increased mortality in children and pregnant women due to severe gastrointestinal and respiratory infections [19,26]. Several studies have highlighted the importance of vitamin A for lung function and development [27]. Other studies confirmed that vitamin A has immune-modulating properties and plays an important role in the immune response to infections, mainly through retinoic acid, its main metabolite. Early randomized clinical trials studying the effect of vitamin A supplementation showed it resulted in a significant reduction of mortality and less severe manifestations of several infectious diseases in cases of vitamin A deficiency due to malnutrition [12,28]. On the contrary, the consequences of reduced vitamin A levels that are due to acute inflammation are not well understood.

Vitamin A influences cellular immunity in a wide variety. A number of studies have shown that vitamin A has a central function in the development and differentiation of dendritic cells (DCs), the most important antigen-presenting cells for activating naive T-cells. DCs express three isotypes of RA receptor and, therefore, directly respond to vitamin A [29]. After receptor activation, DCs drive T-cell differentiation into either antiinflammatory regulating T-cells (Treg) or proinflammatory effector T-cells, and through this, they maintain homeostasis between anti-inflammatory and proinflammatory stimuli [30]. To resolve infection, vitamin A leads to the migration of effector T-cells to the inflammatory site via the induction of leukocyte-homing receptors such as CCR9 and α4β7 integrin [31]. Vitamin A initiates the production of proinflammatory cytokines such as interferon-gamma to resolve viral infection [32]. Vitamin A also drives the humoral immune response, as it crucially promotes B-cell maturation and antibody responses in viral clearance [33,34]. To limit proinflammatory stimuli, vitamin A promotes the differentiation and extravasation of anti-inflammatory Treg cells to the site of inflammation [35]. Taken together, vitamin A can promote both proinflammatory and anti-inflammatory cellular immune responses.

In COVID-19, immune imbalance and the disruption of T-cell responses may be important drivers of severe disease. Real-world data suggest a strong antiviral T-cell response after infection with SARS-CoV-2. The majority of SARS-CoV-2-specific T-cells show an effector phenotype with the dominant production of antiviral proteins, such as interferon-gamma, to terminate infection [36]. However, in patients with severe and critical disease, T-cells exhibit lower levels of antiviral interferon-gamma compared to patients with mild disease [37]. This is known as COVID-19-associated "T-cell exhaustion" and may lead to impaired viral clearance. The resulting massive replication of SARS-CoV-2 and viral infection of cells and organs and subsequent viral release from dying cells is considered as an initial driver of cytokine storming, leading to uncontrolled immune reactions and often fatal outcomes due to multiple organ failure [38]. Furthermore, levels of anti-inflammatory Treg cells in critically ill COVID-19 patients are lower than those in patients with mild disease [39], possibly leading to further uncontrolled immune reactions. As described above, vitamin A influences the production of interferon-gamma through effector T-cells as well as the differentiation of anti-inflammatory Treg cells. Our study results show a significant reduction of vitamin A levels in critically ill COVID-19 patients. Owing to these study results, vitamin A should be considered a relevant agent in maintaining immune balance in mild and moderate COVID-19. Immune imbalance and disruption of antiviral Tcell responses in cases of severe and critical COVID-19 may, in part, be driven by decreased vitamin A plasma levels during the acute phase.

This study has some limitations. First, this study can only reveal descriptive data and does not prove causality. Further studies are needed to clarify whether COVID-19 disease course worsens as a result of Vitamin A reduction in patients serum, or Vitamin A levels decrease as a consequence of severe COVID-19 and whether this reduction itself has an impact on disease course. Second, it has to be pointed out that in this study both unbound vitamin A and vitamin A bound to retinol binding protein (RBP) were measured. It is known that hepatic synthesis of RBP is reduced following acute phase reactions in terms of prioritization of the liver for synthesis of acute phase proteins such as CRP and ferritin [40]. Further studies including analysis of free (unbound) vitamin A in plasma are needed to clarify whether vitamin A reduction in plasma in COVID-19 patients is a consequence of reduced RBD synthesis or free (unbound) vitamin A is reduced as well. Third, the sample size in cohort subgroups is small, which can lead to bias especially in analyses for associations between Vitamin A levels and ARDS/mortality. The small number of events in univariate analyses could possibly lead to confounding factors and lead to misinterpretation. Taking these limitations into account the consequences of reduced vitamin A plasma levels in COVID-19 patients require further research to investigate possible therapeutic approaches of vitamin A supplementation during acute infection. Finally, controlled prospective studies are needed to investigate the therapeutic effect of vitamin A supplementation in cases of acute COVID-19.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/nu13072173/s1, Supplementary Figure S1: Study flowchart Vitamin A plasma levels in COVID-19 patients: A prospective multicenter study and hypothesis. COVID-19, coronavirus disease 2019, Supplementary Figure S2: Representation of the percentage symptom expression of the convalescent COVID-19 patient cohort during acute phase of disease. At the time of blood sampling after recovery from disease, the patients were all free of symptoms, Supplementary Figure S3: Presentation of important laboratory parameters. Classification into moderate, severe and ciritical disease and convalescent patients (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001).

**Author Contributions:** Conceptualization, P.-R.T., R.V. and A.H.-K.; Investigation, P.-R.T., R.V. and T.N.; Methodology, M.F. and A.H.-K.; Software, J.A.M.; Supervision, I.K. and H.S.; Validation, J.A.M.; Visualization, P.-R.T. and J.A.M.; Writing—original draft, P.-R.T. and R.V.; Writing—review & editing, P.-R.T., I.K. and A.H.-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 according to the guidelines of the Declaration of Helsinki and approved by the local Ethics Committee of the University of Münster (AZ 2020-220-f-S and AZ 2020-210-f-S).

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

**Data Availability Statement:** Data cannot be made public as personal patient data are included.

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

#### **References**

