Introduction

In December 2019, a global outbreak of acute systemic illness with symptoms including cough, dyspnea, fever and loss of smell began in China. Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) has been identified as the responsible pathogen, with the disease in question being COVID-19. To date, 770 million cases have been diagnosed worldwide, of which nearly 7 million were fatal1.

At the same time, one of the most prevalent metabolic diseases, diabetes (affecting approximately half a billion people worldwide), continued to rise in frequency2. Diabetes is the first non-communicable disease recognized by the United Nations as an epidemic of the twenty-first century. Thus, patients suffering from both diabetes and COVID-19 constituted a significant proportion of COVID-19 patients. Diabetes has already been shown to be a predictor of a more severe course and worse prognosis during previous global pandemics, such as Severe Acute Respiratory Syndrome (SARS) and Middle East Respiratory Syndrome (MERS)3.

Additionally, during the COVID-19 pandemic, patients with diabetes were reported to have a worse prognosis4. They are burdened with more severe symptoms, a higher likelihood of developing complications, and a higher mortality rate. One of the main reasons behind this occurrence is a concomitant impairment of the immune system against invading pathogens in patients with diabetes. This is caused by defects in both the innate immune response, such as neutrophil dysfunction, and the acquired immune response5. In both diabetes and COVID-19, there is an increase in interleukin-6 (IL-6) levels due to inflammation. Interestingly, several studies have shown that inhibiting IL-6 is associated with significant clinical improvement for patients hospitalized due to COVID-196.

In cases of viral infection, more specific mechanisms influencing the difference in course and mortality of patients with and without diabetes have been described. One such factor is the expression of angiotensin-converting enzyme 2 (ACE-2) in patients with diabetes. Both SARS CoV-1 and SARS CoV-2 use the ACE-2 receptor to enter cells. Diabetes and related characteristics may increase ACE2 expression, which can influence susceptibility to infection and its severity7. It has been noted that in patients with diabetes, the concentration of ACE2 is increased in some tissues—such as the lungs8—and decreased in others—such as the kidneys8. ACE2 may play a significant role in diabetes: in the pancreas, a relative deficiency of ACE2 as the disease progresses can contribute to reduced insulin secretion, while in the glomerulus it can promote proteinuria9. Another mechanism responsible for the greater virulence of SARS-CoV-2 in patients with diabetes is elevated serum furin activity. Furin is a type 1 membrane protease belonging to the proprotein convertase subtilisin/kexin (PCSK) family. It is involved in the entry of coronaviruses into the cell, and its significant elevation can facilitate viral replication10. Interestingly, higher activity of PCSK-9 (due to its overexpression) remains a main factor in developing familial hypercholesterolemia—a disease reported to modulate the odds of developing type 2 diabetes11.

SARS-CoV-2 infection can present as asymptomatic, with a range of manifestations from mild influenza-like illness (ILI) to life-threatening complications such as acute respiratory failure requiring non-invasive or invasive respiratory support12. Acute respiratory distress syndrome (ARDS) is the main cause of death in COVID-19 and results from the cytokine storm or excessive inflammatory response, causing the release of pro-inflammatory cytokines like interleukins and tumor necrosis factor-alpha (TNF-α).

Factors associated with the severity of the disease and risk of death in COVID-19 patients are being analyzed by considering various laboratory, physical, and subject examination parameters. RDW-SD is proposed as one such feature, as it was found to be significantly higher in fatal outcomes in patients receiving mechanical ventilation (MV) compared to survivors without oxygen support13,14. However, further research is needed to confirm this hypothesis. This study is based on data from the COLOS registry (COronavirus in the LOwer Silesia registry) and features a retrospective, exploratory data analysis, taking into account pre-hospitalization oxygen therapy, pre-diabetes or type 2 diabetes (T2DM) status, and variable saturation.

Materials and methods

Study population

The retrospective study presented was based on medical documentation from 2139 patients who were admitted and treated at the COVID-19 Hospital, organized by the University Medical Hospital in Wroclaw, Poland. The study period ranged from February 2020 to June 2021. The authors did not have access to patient identification information and had no direct contact with the participants. The Bioethics Committee of the Medical University of Wroclaw approved the use of this data for scientific research (No. KB-444/2021). All patients provided informed consent to participate in the study, understanding that the results may be used for research purposes. All patients displayed symptoms of COVID-19 and tested positive for SARS-CoV-2 RNA through the reverse-transcriptase polymerase chain reaction (RT-PCR) using nasopharyngeal swab material, following the protocol published by the World Health Organization (WHO)15. The observation period began on the day of admission to the hospital and ended on the day of discharge or death. Observations were recorded throughout the entire hospitalization period.

Adult patients who provided verbal and written consent to participate in the study and were pre-qualified by the surrounding emergency departments for hospitalization at the University Medical Hospital due to the need for intravenous medication (e.g. antibiotics) or oxygen therapy, were included. Patients who required treatment in the intensive care unit (ICU) from the beginning were not analyzed. Patients in the study typically exhibited severe COVID-19 with concomitant severe pneumonia and respiratory failure/pre-ARDS (with a Modified Early Warning Score (MEWS) of 3–4 points) and required close monitoring for intensive care. For patients with hypoxemia expressed by an SpO2 level below 90%, oxygen therapy was gradually increased, beginning with non-invasive methods such as nasal cannula, oxygen masks, and oxygen masks with a tank for high-flow oxygen therapy HFNOT (High Flow Nasal Oxygen Therapy) until saturation of 94% or higher was achieved. If these methods were ineffective, the patient was transferred to the intensive care team, who then qualified the patient for invasive therapy methods.

Patient characteristics were obtained from individual clinical records, including age, sex, length of hospitalization, respiratory support, smoking, comorbidities, symptoms of COVID-19, medications used before hospitalization, and results of laboratory tests at admission: complete blood count (5 DIFF), neutrophil–lymphocyte ratio, serum concentrations of sodium, potassium, calcium, C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 (IL-6), total protein, albumin, creatinine, urea, uric acid, D-dimers, fibrinogen, ferritin, glucose, total bilirubin, low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides (TG), troponin T, NT-pro-BNP, TSH (3rd generation), vitamin D, the activity of alanine aminotransferase (ALT), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), gamma-glutamyltransferase (GGT), amylase, and lipase, and values of APTT and INR.

Categorical parameters analyzed upon hospital admission included fever, hemorrhage, peripheral edema, hypertension, atrial fibrillation or flutter, heart failure, history of myocardial infarction, history of blood vessel revascularization (PCI, CEBG), peripheral arterial disease (intermittent claudication, arterial revascularization, necrosis or acute ischemia, aortic aneurysm of at least 6 cm, previous cerebral hemorrhage or transient ischemic attack [TIA]), thyroid dysfunction, current or past smoking history, chronic obstructive pulmonary disease (COPD), sleep apnea, presence of lung lesions, cough, shortness of breath, olfactory disorders, diarrhea, vomiting, abdominal pain, solid tumors, chronic renal failure, acute renal failure. Medications taken prior to hospitalization were also considered, including angiotensin-converting enzyme (ACE) inhibitors, β-blockers, amlodipine-type calcium channel blockers, acetylsalicylic acid and other platelet aggregation inhibitors, statins, diuretics, and inhaled anticholinergics.

During hospitalization, the parameters that were analyzed included the occurrence of pneumonia or obstruction, shock (hypovolemic, cardiac or septic), venous thromboembolism, PCI and/or CEBG, myocardial infarction, heart failure decompensation, stroke or TIA, new cognitive impairment, the presence of new neurological disorders, deterioration of the patient’s state, death, respiratory support, intubation, and the use of medications such as loop diuretics, low-molecular-weight heparin (LMWH), convalescent plasma, remdesivir, steroids, inhaled bronchodilators, β-agonists, inhaled anticholinergic bronchodilators, acetylsalicylic acid, and antibiotics.

The study group was divided according to a three-stage scale, taking into account T2DM and prediabetes, the use of pre-hospitalization oxygen therapy, and, in the absence of this therapy—the value of saturation (division criterion: saturation < 95%). The criteria for dividing the study group into subgroups are shown in Fig. 1. T2DM or prediabetes was selected as a criterion for the two main groups (Io level): group (A)—patients with T2DM or prediabetes; group (B)—patients without T2DM or prediabetes. Moreover, each of these groups was divided into two subgroups (IIo level), considering the use of pre-hospitalization oxygen therapy: group (C)—patients with T2DM/prediabetes with pre-hospitalization oxygen therapy; group (D)—patients with T2DM/prediabetes without pre-hospitalization oxygen therapy; group (E)—patients without T2DM or prediabetes with pre-hospitalization oxygen therapy; group (F)—patients without T2DM or prediabetes without pre-hospitalization oxygen therapy.

Figure 1
figure 1

The scheme featuring the division process carried out on the study group of patients. I0 level: group A – patients with type 2 diabetes or/and prediabetes; group B – patients without type 2 diabetes and prediabetes; II0 level: group C – patients with type 2 diabetes or/and prediabetes with prehospitalization oxygen therapy; group D – patients with type 2 diabetes or/and prediabetes without prehospitalization oxygen therapy; group E – patients without type 2 diabetes and prediabetes with prehospitalization oxygen therapy; group F – patients without type 2 diabetes and prediabetes without prehospitalization oxygen therapy; IIIo level: group G – patients with type 2 diabetes or/and prediabetes without prehospitalization oxygen therapy with SpO2 < 95% during the admission to the hospital; group H – patients with type 2 diabetes or/and prediabetes without prehospitalization oxygen therapy with SpO2 ≥ 95% during the admission to the hospital; group I – patients without type 2 diabetes and prediabetes without prehospitalization oxygen therapy with SpO2 < 95% during the admission to the hospital; group J – patients without type 2 diabetes and prediabetes without prehospitalization oxygen therapy with SpO2 ≥ 95% during the admission to the hospital.

Full size image

In addition, for cases where oxygen therapy was not used before admission to the hospital, a saturation level below 95% measured during admission (IIIo level) was included as an additional criteria for creating the following groups: group (G)—patients with T2DM/prediabetes without pre-hospitalization oxygen therapy and SpO2 < 95% during admission to the hospital; group (H)—patients with T2DM/prediabetes without pre-hospitalization oxygen therapy and SpO2 ≥ 95% during admission to the hospital; group (I)—patients without T2DM and prediabetes without pre-hospitalization oxygen therapy and SpO2 < 95% during admission to the hospital; group (J)—patients without T2DM and prediabetes without pre-hospitalization oxygen therapy and SpO2 ≥ 95% during admission to the hospital.

All methods were performed in accordance with relevant guidelines and regulations.

Statistical analysis

Data preprocessing and visualization were performed using Python 3.10.7 and standard packages (numpy 1.21.4, pandas 1.4.4, matplotlib 3.5.3, seaborn 0.11.2). The statistical analysis was conducted using the Statistica 13.3 package, available under a Wrocław Medical University license. A frequentist approach with an alpha-value of 0.05 was used for statistical inference. The assumptions of normality and homoscedasticity were tested with the Shapiro–Wilk and Levene’s tests, respectively. Student’s t-test was used to compare two groups in terms of continuous variables if no assumptions were violated. Otherwise, the Mann–Whitney U test was used. Pairwise differences in distributions of discrete variables were checked with the χ2 test, after ensuring that the estimated counts were greater than or equal to 5. If any estimated count was lower than 5, Yates' correction for continuity was applied to counter false positive (p < 0.05) cases.

Results

The study group consisted of 2139 COVID-19 patients, with 1076 women (50.30%) and 1063 men (49.70%). The average age was 63.73 ± 15.69 years. Out of 2139 COVID-19 patients, 473 (22.11%) suffered from either T2DM or prediabetes (group A), while 1666 (77.89%) did not have these diabetic disorders (group B). Group C consisted of 241 (11.27%) patients with T2DM or prediabetes who received oxygen therapy before hospitalization, group D consisted of 232 patients with T2DM or prediabetes who did not receive oxygen therapy before hospitalization (10.85%), group E consisted of 1013 patients without T2DM or prediabetes who did not receive oxygen therapy before hospitalization (47.36%), group F consisted of 652 patients without T2DM or prediabetes who did not receive oxygen therapy before hospitalization (30.48%). Group G consisted of 35 (11.27%) patients with T2DM or prediabetes who did not receive oxygen therapy before hospitalization and had a saturation level of less than 95% during hospital admission (1.64%), group H consisted of 111 patients with T2DM or prediabetes who did not receive oxygen therapy before hospitalization and had a saturation level greater than or equal to 95% during hospital admission (5.19%). Group I consisted of 134 (6.26%) patients without diabetic disorders who did not require oxygen therapy before hospitalization and had a SpO2 level of less than 95% during hospital admission (1.64%), group J consisted of 472 patients without diabetic disorders who did not receive oxygen therapy before hospitalization and had a SpO2 level greater than or equal to 95% during hospital admission (22.10%). Detailed characteristics of the patient groups are presented in Table 2 and Fig. 1S.

Relationship between oxygen therapy, saturation and other clinical and laboratory parameters obtained on admission to the hospital and during hospitalization, in context of the occurrence of prediabetes or T2DM

Descriptive statistics of variables, in case of significant pairwise difference (p < 0.05), are shown in Tables 1 and 2.

Table 1 Comparison of the discretized quantitative clinical and laboratory parameters in tested groups of patients with COVID-19 at the time of admission to hospital and during hospitalization.
Full size table
Table 2 Baseline characteristics (quantitative data) of 2139 patients hospitalized due to COVID-19, divided into groups as shown in Fig. 1
Full size table

The variability associated with the occurrence of prediabetes/T2DM (group A vs. group B)

Prior to hospitalization, patients in group A were more likely to be under treatment with ACE1 (p < 0.001), β-blockers (p < 0.001), amlodipine-type calcium antagonists (p < 0.001), aspirin (p < 0.001), antiplatelet agents (p < 0.001), statins (p < 0.001), and diuretics (p < 0.001). These patients were also taking antidiabetic drugs such as insulin, metformin, SGLT2 inhibitors or oral antidiabetic drugs more frequently.

Upon hospital admission the following diagnoses were more frequent in group A: obesity (p = 0.001), chronic kidney disease (p < 0.001), acute kidney injury (p < 0.001) hemorrhage (p = 0.031), lung lesions (p < 0.001), peripheral edema (p < 0.001), hypertension (p < 0.001), atrial fibrillation or flutter (p < 0.001), heart failure (p < 0.001), COPD (p = 0.001), sleep apnea (p < 0.001), peripheral arterial disease (e.g., intermittent claudication, arterial revascularization, necrotic changes or acute ischemia, aortic aneurysm at least 6 cm, p = 0.007), and thyroid function disorders (p < 0.001). Additionally, patients in group A were more likely to have a history of myocardial infarction (p = 0.028), blood vessels revascularization (p < 0.001), cerebral hemorrhage or TIA (p < 0.001), and smoking (p = 0.007). However, olfactory disturbances (p = 0.028), fever (p = 0.000) and cough (p = 0.025) occurred less frequently in this group. Group A also had higher values of BMI (p < 0.001), age (p < 0.001), leukocyte count (p < 0.001), RDW-SD (p < 0.001), PDW (p = 0.001), MPV (p < 0.001), potassium concentration (p < 0.001), hsCRP (p = 0.013), procalcitonin (p < 0.001), INR (p < 0.001), glucose (p < 0.001), urea (p < 0.001), and creatinine (p < 0.001). In group A, upon admission to the hospital, an increase in the following laboratory parameters was observed to be above the assumed cut-off points more frequently: troponin concentration > 15.6 pg/ml (p < 0.001), as well as threefold (p < 0.001) and fivefold (p < 0.001) increases in troponin T concentrations above the maximum reference threshold of 15.6 pg/ml, bilirubin concentration > 2 mg/dl (p < 0.001), RDW-SD > 47 (p < 0.001), procalcitonin concentration > 0.1 ng/ml, glucose concentration > 100 mg/mL (p < 0.001), creatinine > 1.1 mg/mL (p < 0.001), creatinine > 2 mg/mL (p < 0.001), eGFR < 60 (p < 0.001), NT-ProBNP > 100 p/dL (p < 0.001) and triglycerides > 150 mg/mL (p = 0.015). Moreover, outside the reference range, the following parameters were more common: WBC (p < 0.001), MCHC (p = 0.002), MCV (p < 0.001), sodium concentration (p = 0.039), potassium concentration (p = 0.004), INR (p < 0.001), APTT (p < 0.001) and uric acid level (p < 0.001). Conversely, group B showed higher values of body temperature (p < 0.001), MCHC (p < 0.001), sodium concentration (p = 0.008) and eGFR (p < 0.001).

Group A showed a higher frequency of individuals who required respiratory support during hospitalization (p < 0.001), intubation (p < 0.001), were subject to PCI or CEBG (p < 0.001), showed signs of lung obstruction (p < 0.001), and experienced shock (cardiac, hypovolemic, septic) during hospitalization (p < 0.001). They also suffered from myocardial infarction (p = 0.001), decompensation of heart failure (p < 0.001), cerebral hemorrhage (p = 0.005), and developed cognitive (p < 0.001) and neurological disorders (p = 0.004). Additionally, these patients more often required treatment with loop diuretics (p < 0.001), inhaled anticholinergics (p < 0.001), LMWH (p < 0.001), aspirin (p < 0.001), and antibiotics (p < 0.001) during their hospitalization. Furthermore, they were hospitalized for more than 49 days (p = 0.001) and experienced more frequent health deterioration (p < 0.001), ultimately leading to in-hospital death more often (p < 0.001).

The variability associated with being subject to prehospital oxygen therapy among patients suffering from prediabetes/T2DM (group C vs. group D)

Upon admission to hospital, acute kidney injury (p = 0.011) and/or lung lesions (p < 0.001) were observed more frequently in group C. However, this stratum was less likely to be diagnosed with chronic kidney disease (p = 0.044) and had less frequent findings of elevated bilirubin levels above 2 mg/dL (p = 0.013), solid tumors (p = 0.031), and smoking history (p = 0.014). Additionally, this group showed more frequent deviations of the following values (Table 1) outside of their reference ranges: CRP ≥ 5 mg/dL (p < 0.001), procalcitonin ≥ 0.1 ng/dL (p = 0.049), IL-6 ≥ 0.1 mg/dL (p < 0.001), D-dimers > 0.5 µg/mL (p = 0.018), Hb concentration < 11 g/L (p < 0.001), and both WBC (p = 0.013) and LDH (p < 0.001). Higher values were found in group C for the following quantitative parameters: heart rate (p = 0.036), hsCRP (p < 0.001), ALT (p = 0.047), and total bilirubin (p = 0.011). However, INR levels were higher in group D (p = 0.003). Detailed quantitative measures for both groups can be found in Table 2.

During hospitalization, group C frequently exhibited shallow breathing (p < 0.001) and/or cough (p < 0.001). Additionally, these patients often required respiratory support (p < 0.001), intubation (p < 0.001), and showed signs of lung obstruction or pneumonia (p < 0.001). They also experienced hypovolemic, cardiac, or septic shock (p < 0.001) and venous thromboembolism (p < 0.001). Furthermore, they required treatment with loop diuretics (p = 0.001), steroids (p < 0.001), LMWH (p < 0.001), convalescent plasma (p < 0.001), remdesivir (p < 0.001), and antibiotics (p < 0.001). The hospital stay for group C patients was typically less than 49 days (p = 0.036), however, their overall clinical status often deteriorated (p < 0.001) and death (p < 0.001) occurred more frequently in this group.

The variability associated with being subject to prehospital oxygen therapy among non-diabetic individuals (group E vs. group F)

Prior to hospitalization, the following diagnoses were more frequent among individuals in group E (Table 1): obesity (p = 0.008), peripheral edema (p = 0.005), hypertension (p < 0.001), heart failure (p = 0.291), myocardial infarction (p = 0.004), chronic obstructive pulmonary disease (p < 0.001), lung lesions (p < 0.001), shortness of breath (p < 0.001), cough (p < 0.001), and acute renal failure (p < 0.001). During hospitalization, patients in group E more often showed deterioration of their health status (p < 0.001) and/or death (p < 0.001). The higher mortality rate in this group may have been influenced by a higher frequency of pulmonary obstruction or pneumonia (p < 0.001), shock (p < 0.001), decompensated heart failure (p = 0.006), venous thromboembolism (p = 0.001), and a greater likelihood of further cognitive impairment (p = 0.001). Additionally, group E required respiratory support (p < 0.001) and intubation (p < 0.001) more frequently.

Upon admission to the hospital, an increase above the assumed cut-off point was observed more often in cases of troponin T concentration (> 15.6 pg/ml, p < 0.001; > threefold, p < 0.001; > fivefold, p = 0.004), BMI ≥ 30 (p = 0.008), CRP ≥ 5 mg/dl (p < 0.001), procalcitonin ≥ 0.1 ng/dl (p < 0.001), IL-6 ≥ 0.1 mg/dl (p < 0.001), D-dimer > 0.5 μg/ml (p = 0.006), ALT > 35 U/L (p < 0.001), bilirubin > 1.2 ng/mL (p = 0.004), glucose ≥ 100 mg/dl (p < 0.001), triglycerides > 150 mg/dL (p = 0.002), and hemoglobin < 11 g/ml (p = 0.008). Additionally, the following parameters were also found to be out-of-range: WBC (p = 0.004), lymphocyte count (p < 0.001), sodium (p < 0.001), potassium (p = 0.014), ferritin (p < 0.001), albumin (p < 0.001), urea (p < 0.001), and uric acid (p = 0.012). The activity of GGT (p < 0.001), LDH (p < 0.001), and INR values (p < 0.001) were also higher in this group (E). Interestingly, this group also exhibited a lower frequency of LDL > 115 mg/dL (p = 0.028), while values of HDL (for both females and males) were more frequently outside of the reference range (p = 0.003) compared to group F.

The values of the following quantitative parameters were increased in group E compared to those of group F: systolic blood pressure (p < 0.001), heart rate (p = 0.045), RDW-SD (p = 0.012), MPV (p = 0.009), sodium (p = 0.043), hsCRP (p < 0.001), procalcitonin (p < 0.001), INR (p < 0.001), glucose (p < 0.001), urea (p < 0.001), serum creatinine (p < 0.001), ALT (p < 0.001), and GGT (p < 0.001). Conversely, patients in group F showed higher values for neutrophil–lymphocyte ratio (p = 0.005), APTT (p < 0.001), and LDH (p < 0.001) than patients in group E. More detailed data is provided in Table 2. In addition, patients in group E, after admission to the hospital, were more frequently treated with loop diuretics (p = 0.002), LMWH (p < 0.001), convalescent plasma (p < 0.001), remdesivir (p < 0.001), steroids (p = 0.050), antibiotics (p < 0.001), and bronchodilators such as β2-agonists (p = 0.002) or anticholinergic agents (p = 0.030) (see Table 1).

Treatment-wise, patients in group E were more commonly prescribed ACEIs (p < 0.001), β-blockers (p = 0.013), amlodipine-type calcium agonists (p = 0.001), and statins (p = 0.001) prior to hospitalization.

The saturation-associated variability observed among patients suffering from prediabetes/T2DM who did not receive prehospital oxygen therapy (group G vs. group H)

Upon admission to the hospital, the following diagnoses were more common in group G: cerebral hemorrhages or TIAs (p = 0.035), lung lesions (p = 0.004), and shallow breathing (p < 0.001). The same group also showed higher rates of abnormal potassium levels (p = 0.015), urea levels (p = 0.014), albumin levels (p = 0.042), procalcitonin levels (p < 0.001), creatinine levels above 1.1 mg/mL (p = 0.047), and higher lymphocyte counts (p = 0.042). Additionally, the following quantitative parameters had higher values in group G: BMI (p = 0.002), age (p < 0.001), MCV (p = 0.004), PDW (p = 0.040), hsCRP (p < 0.001), INR (p < 0.001), glucose levels (p < 0.001), urea levels (p < 0.001), ALT levels (p < 0.001), GGT levels (p < 0.001), and LDH levels (p < 0.001). In contrast, patients in group H had higher levels of procalcitonin (p < 0.001) and APTT (p < 0.001) compared to group G (Table 2).

During hospitalization, participants in group G were more likely to experience a deterioration in their health status (p < 0.001). Additionally, signs of obstruction or pneumonia were more frequently observed (p < 0.001).

There were no significant differences in treatment between the two groups before or during hospitalization.

The saturation-associated variability observed among non-diabetic patients who did not receive prehospital oxygen therapy (group I vs. group J)

Prior to hospitalization, patients in group I showed a higher frequency of fever (p < 0.001), COPD (p = 0.002), lung lesions (p < 0.001), cough (p = 0.025), shallow breathing (p < 0.001), olfactory disorders (p < 0.030), and diarrhea, vomiting, and/or abdominal pain (p = 0.019). Furthermore, these patients were more often treated with diuretics (p = 0.040).

Upon hospital admission, the following quantitative parameters were higher in group I: body temperature (p = 0.012), hsCRP (p < 0.001), procalcitonin (p < 0.001), glucose (p = 0.007), total bilirubin (p = 0.002), GGT (p = 0.006), and LDH (p < 0.001). Additionally, patients in group I more frequently showed potassium concentration outside the reference range (p = 0.008), CRP ≥ 5 mg/dl (p < 0.001), procalcitonin concentration > 0.1 ng/mL (p = 0.014), glucose concentration > 100 mg/mL (p < 0.001), creatinine > 1.1 mg/mL (p = 0.045), and concentrations of urea (p = 0.037), albumin (p = 0.035), and LDH (p < 0.001) outside the reference range. Conversely, patients in group J showed higher values of eGFR (p = 0.020) and ALT (p < 0.001).

During hospitalization, participants in group I were more likely to be treated with LMWH (p = 0.050), steroids (p < 0.001), and antibiotics (p = 0.027) (see Table 1).

Discussion

Although the COVID-19 pandemic has ended, due to its high mortality rate and late consequences, science continues to strive towards better understanding its mechanism and further implications. Attempts are being made to develop novel risk assessment models aimed at reducing treatment costs while maintaining or improving clinical sensitivity. However, as of now, no universal model accounting for comorbidity has been successfully developed16.

This study shows that among patients with T2DM or prediabetes, pre-hospitalization oxygen therapy was associated with a decreased count of out-of-range values for selected diagnostic parameters. Due to the inclusion of 10 different patient groups in the comparative analysis and the presence of statistically significant differences in various parameters, three profiles were created: morphological, inflammatory, and biochemical (Fig. 1). This arrangement of results (Figs. 2, 3, 4) allows for easier comparison with findings from other scientific publications. Furthermore, the impact of antidiabetic therapy used in the study group of patients on the severity of COVID-19 was thoroughly discussed in a previous publication17.

Figure 2
figure 2

Morphological panel: Frequency of RDW-SD > 47 and values: Hb, MCH, MCHC LYMPH, WBC outside the reference range in patient groups A–J.

Full size image
Figure 3
figure 3

Panel of the inflammatory parameters: Frequency of values of D-dimers, CRP, albumin, PCT and ferritin outside the reference range in patient groups A–J.

Full size image
Figure 4
figure 4

Panel of the biochemical parameters: Frequency of values of TG, HDL-Chol, LDL-Chol, ALAT, Troponin T, bilirubin, GGTP, creatinine, Na+, K+, glucose outside the reference range in patient groups A–J.

Full size image

The parameters that were often found outside the reference range include the concentrations of lymphocytes, hemoglobin, WBC, and red blood cell parameters—specifically RDW-SD, MCH, MCHC, and MCV (Fig. 2). In terms of inflammatory markers, there were elevated levels of D-dimer, IL-6, CRP, albumin, procalcitonin, and ferritin (Fig. 3). Additionally, in the biochemical profile, there were increased levels of creatinine, Na + , K + , glucose, triglycerides, HDL, LDL, troponin T, bilirubin, and activities of ALT and GGT (Fig. 4).

Red blood cell parameters are often used as prognostic indicators in COVID-19 exacerbation due to their accessibility and frequency of analysis18,19,20. In severe COVID-19 cases, there is a decrease in lymphocyte concentration below the reference range21 and hemoglobin concentration22. Some publications have also reported an increase in RDW values and a decrease in platelet concentration23. Furthermore, studies have shown that an RDW-SD value of ≥ 47 indicates severe disease and a high risk of death24.

This study showed that approximately 60% of patients who used pre-hospitalization oxygen therapy (groups C and E) had a lymphocyte concentration that exceeded the reference range (a decrease in lymphocyte concentration is observed, as shown in Fig. 1S), while 35% and 40% of patients in groups D and E, respectively, had the same result. This is most likely due to the more severe form of COVID-19 upon admission to the hospital. None of the comparisons between the groups showed statistically significant differences. However, according to Table 1, it was found that patients in group A had a much higher prevalence of RDW-SD ≥ 47 compared to patients in group B. This relationship has not yet been demonstrated in the literature. It is known that diabetics infected with SARS-CoV-2 are more likely to be hospitalized, develop severe pneumonia, and have a higher mortality rate compared to people without T2DM. This is most likely because chronic hyperglycemia can impair innate and humoral immunity and is associated with chronic inflammation, which can lead to an excessive inflammatory response and, consequently, the occurrence of acute respiratory distress syndrome25.

Coagulation disorders, along with inflammatory mechanisms, are implicated in the onset of neurological manifestations in COVID-19 (such as demyelinating and neurodegenerative lesions, taste, smell, and visual dysfunctions, and consciousness disorders)26. The concentration of D-dimers is useful in the diagnosis and prediction of deep vein thrombosis recurrence and DIC. In COVID-19, the average value of D-dimer concentration positively correlates with the severity of the disease and is inversely proportional to survival27. Moreover, a systematic review and meta-analysis have demonstrated that interleukin-6 (IL-6) levels are elevated in cases of complicated COVID-19, and increased IL-6 levels are significantly associated with adverse clinical outcomes. This suggests that the progression of initial SARS-CoV-2 infection may be the consequence of an excessive host immune response and autoimmune injury28. A previous study showed that a high level of CRP is strongly associated with venous thromboembolism, acute kidney injury, critical illness, and mortality during COVID-19 infection29. Furthermore, a decrease in serum albumin concentration is associated with a higher risk of intensification of COVID-1930.

High levels of procalcitonin are observed in inflammatory conditions primarily caused by bacteria31. Although COVID-19 is caused by a virus, it often leads to bacterial superinfection, which affects the course of the disease. It has been proven that increased PCT levels are associated with a nearly fivefold higher risk of severe SARS-CoV-2 infection32. Additionally, high serum ferritin levels are associated with mortality and the development of severe outcomes in COVID-19. Cytokine storm syndrome can cause multiorgan failure and hyperferritinemia, as observed in all severe patients upon admission in one of the studies33.

High D-dimer levels were more common in patients of group G compared to patients of group C (100% vs. 85%). This suggests that early oxygen therapy may reduce the number of patients who develop deep vein thrombosis or DIC. Additionally, group G showed a high frequency of records with increased procalcitonin levels and decreased albumin levels, indicating that bacterial superinfections often coexisted with poor endothelial nutrition in this group. These observations lead to two conclusions: firstly, patients with T2DM or prediabetes with SpO2 levels below 95% are at a higher risk of severe COVID-19 compared to other groups studied, and secondly, pre-hospitalization oxygen therapy contributed to a less severe disease course in patients with T2DM or prediabetes as levels of D-dimers, procalcitonin, and albumin were less frequently outside of the reference values in this group. Additionally, the lack of difference in pro-inflammatory parameters between groups C and D may be due to oxygen therapy being used in individuals with a more severe disease course.

Due to the fact that many different types of cells contain receptors with which SARS-CoV-2 interacts, the impact of oxygen therapy on various biochemical parameters was also analyzed in the literature. The analysis of the usefulness of serum total, LDL, non-HDL, and HDL cholesterol to predict the COVID-19 course showed that a severe outcome was associated with lower HDL cholesterol levels and higher triglycerides. These parameters were found to be correlated with ferritin and D-dimer levels, but not with CRP levels34. However, the analyses of the results of measuring LDL, HDL, and triglycerides in the course of COVID-19 in previous studies for different patient groups are inconsistent35. In some infectious diseases, LDL cholesterol levels are decreased, which has been confirmed in some studies on COVID-1936. However, some studies have not shown an association of low LDL levels with a more severe course of COVID-1934. Furthermore, previous clinical studies have found correlations between the severity of COVID-19 and its unfavorable evolution, and the degree of liver damage37,38. In some cases, SARS-CoV-2 can cause viral myocarditis and cardiac injury, which can progress into multi-organ failure39. This is characterized by a significant elevation in troponins, ALT and AST, which is associated with critical changes in renal function parameters and coagulation markers40. Additionally, AST/ALT values in COVID-19 patients were slightly elevated (1–2 times higher than normal values) in approximately 60% of cases, moderate (2–5 times higher than normal) in about 30% of cases and increased by more than 5 times the normal value in approximately 10% of cases41,42. Moreover, patients with COVID-19 showed higher levels of GGT43 and total bilirubin44. Interestingly, SARS-CoV-2 may affect the development of hyperglycemia through the Na + /H + exchanger and lactate pathways45, leading to hypoxia and extracellular acidification, which can cause the accumulation of calcium and sodium ions in the cells. This can promote oxidative stress and collateral damage to the pancreatic cells46. Additionally, SARS-CoV-2 may also cause potassium metabolism disorders, with hypokalemia being more common than hyperkalemia in COVID-19 patients47.

Similarly to the findings featured in the literature, this analysis showed a decrease in LDL concentration (as in36) and an increase in triglyceride levels in the course of COVID-19 (as in34). The highest reductions of LDL-Chol values were observed in groups G and I. Such an observation was not demonstrated in group C. In the presented studies, an increase in both AST and GTP values was most often observed in group F. This is most likely due to the fact that patients in this group had more advanced COVID-19. Similarly, troponin levels were significantly more frequently elevated in group F. However, it was observed that in group G, the levels of troponin and creatinine were increased much more often than in group C. Perhaps, the use of pre-hospitalization oxygen therapy in patients with diabetic disorders affected the rate of increase of troponin and creatinine during hospitalization. It is possible that the use of pre-hospitalization oxygen therapy, especially in people with T2DM or pre-diabetes, contributed to reducing the number of cases of viral myocarditis or renal failure in multi-organ failure.

Previous researchers have hypothesized that furin has a key role in the occurrence of hyperkalemia in COVID-19. This enzyme is necessary for cleaving both the SARS-CoV-2 spike protein and epithelial sodium channels, resulting in the retention of potassium ions and hyperkalemia47. However, in our study, we noticed that hyperkalemia was more common in patients who did not receive pre-hospitalization oxygen therapy with an SpO2 of less than 95%, compared to patients with an SpO2 of 95% or higher (in both the group with T2DM or prediabetes and in patients without diabetes). Additionally, we observed that in group I patients, elevated glucose levels were much more likely to occur compared to patients with neither any diabetic disorders nor pre-hospitalization oxygen therapy, with an SpO2 of 95% or higher. Therefore, we hypothesize that elevated glucose levels and hyperkalemia are caused by an activating mutation of the SUR1/Kir6.2 pancreatic potassium channel (KATP). Presumably, SARS-CoV-2 causes a mutation in the gene encoding this channel. The lack of closing of potassium channels in the pancreas prevents depolarization of the membrane, which means that intracellular potassium ions are not secreted, and the pancreas is not stimulated to secrete insulin. This mechanism has been visualized in Fig. 5.

Figure 5
figure 5

A scheme which pictures the function of the SUR1/KIR.2.6 potassium channel in pancreatic β-cells—in a physiological state (4a and 4b), and in the case of the gain-of-function mutation, hypothesized to be caused by SARS-CoV-2. Upon polarization of the cell membrane, K+ are released through KATP to the extracellular fluid (Fig. 3a). An increase in glycemia induces the shutting of the KATP and increasing the intracellular concentration of ATP and Ca2+, inducing membrane depolarization and insulin secretion (Fig. 4b). The mutations in genes coding for SUR1 or KIR.2.6 activate the KATP channel, preventing the closure of the KATP channel, regardless of glycemia. These mutations prevent insulin secretion, consequently promoting hyperglycemia (similarly to the pathomechanism of MODY diabetes).

Full size image

The SUR1/Kir6.2 pancreatic potassium channel (KATP) plays a vital role in regulating insulin secretion by β-cells (Fig. 5). At rest, the channel remains open, allowing for the free flow of K+ ions out of the cell. However, an increase in the cytosolic concentration ratio of [ATP]/[ADP] (caused by, for example, an increase in glucose concentration) leads to depolarization of the cell membrane, causing the KATP channel to close. This results in hypercalcemia and stimulates the release of insulin. Based on the findings presented in this paper, it is hypothesized that SARS-CoV-2 may cause a mutation in the KATP channel similar to what is seen in maturity onset diabetes of the young (MODY). This mutation prevents the channel from closing, resulting in a constant release of K+ into the extracellular fluid (plasma/serum). In this state, insulinemia cannot be maintained48, leading to hyperglycemia.

Conclusions

The current publication presents the impact of pre-hospitalization oxygen therapy in patients with type 2 diabetes, prediabetes, or without diabetic disorders. It considers saturation levels and how they affect the course of COVID-19, as determined by diagnostic parameters known to potentially indicate the severity of the disease. Based on statistical analysis, it was observed that:

  • Among pre-/diabetic individuals, pre-hospitalization oxygen therapy was associated with fewer cases in which the values of D-dimer, albumin, and procalcitonin were outside of the reference range. Additionally, despite a more severe course of COVID-19, pre-hospitalization oxygen therapy was linked to lower levels of inflammation markers. This suggests that the use of pre-hospitalization oxygen therapy may have contributed to a less severe course of the disease.

  • In patients without pre-hospitalization oxygen therapy, regardless of diabetic status, lower saturation (SpO2 < 95%) was associated with a higher frequency of hyperkalemia. Furthermore, patients without diabetic disorders who did not receive pre-hospitalization oxygen therapy and had SpO2 < 95% showed a higher prevalence of hyperglycemia compared to those who had neither diabetic disorders nor pre-hospitalization oxygen therapy but had a SpO2 ≥ 95%. Based on this observation, we hypothesize that elevated glucose levels and hyperkalemia may be caused by an activating mutation of the SUR1/Kir6.2 pancreatic potassium channel (KATP).