1 Introduction

Upper respiratory tract infections are very prevalent and are the most common infectious diseases in adulthood every year. These infections are mainly caused by viruses, although bacterial infections are possible [1]. Respiratory viral infections are acute and transmitted rapidly, resulting in epidemics and pandemics. Recently, severe acute respiratory syndrome due to coronavirus-2 (SARS-Cov-2), also called coronavirus virus 2019 (COVID-19), has presented a major public health challenge. The COVID-19 outbreak has become a disastrous global pandemic with immeasurable impacts on communities. The main cause of mortality from the SARS-CoV family to date is atypical pneumonia. As the disease progresses, pro-inflammatory cytokines are secreted; in particular, interleukin (IL)-1β and IL-18 are secreted by type 1 T helper (Th1) immune cells and macrophages and exacerbate lung inflammation, finally causing fibrosis [2].

Observational studies have indicated that increases in serum vitamin D levels are associated with decreases in inflammatory markers, including hs-CRP [3]. In a systematic review and meta-analysis, Martineau et al. analyzed 25 randomized controlled trials and demonstrated that vitamin D could have a protective effect against acute respiratory tract infections [4]. Furthermore, evidence suggests that vitamin D has therapeutic effects or prevents adverse outcomes in COVID-19. Vitamin D inhibits the expression of inflammatory cytokines, such as tumor necrosis factor-α (TNF-α) and IL-1β [5]. Vitamin D insufficiency is correlated with overproduction of Th1-derived cytokines [6]. Diminished Th1 proliferation could reduce serum levels of pro-inflammatory cytokines and adaptive immune responses. Furthermore, vitamin D can impact T cell maturation and could shift the maturation of inflammatory Th17 to anti-inflammatory regulatory T cells, thereby reducing serum levels of pro-inflammatory cytokines, including IL-6, IL-1, TNF-α, IL-12, and IL-17, while elevating levels of the anti-inflammatory IL-10 [7]. A potential mechanism underlying the protective effect of vitamin D against acute lung injury is the modulation of Angiotensin-converting enzyme 2 (ACE2) receptors in lung tissue [8]. This mechanism may also explain the relationship between vitamin D deficiency and COVID-19. However, data on the relationship between serum vitamin D levels and COVID-19 are limited [9]. A negative association has been reported between serum 25(OH)D and mortality and morbidity rates due to SARS-CoV-2 infection in 20 European countries, especially in Spain, Switzerland, and Italy (countries with a high rate of COVID-19) [10]. The prevalence of vitamin D deficiency is high in Iran [11], and this region has a particularly high prevalence of infection and mortality rate due to SARS-CoV-2 [12]. However, little is known about the nutritional causes of COVID-19 susceptibility in the Iranian population. We assessed the association between serum 25(OH)D and clinical symptoms, mortality, and inflammatory markers among patients admitted for COVID-19 in Mashhad, Iran.

2 Materials and Methods

3 Results

A total of 452 patients were evaluated. The mean age of the patients was 63.87 ± 17.97 years. Demographic characteristics of the study patients are presented in Table 1. The median and IQR for vitamin D levels were 20.60 and 20.88 ng/mL, respectively. Only 30% of patients had serum vitamin D levels of > 30 ng/mL. Severe vitamin D deficiency was observed in 78 patients (17.3%), and vitamin D deficiency (29.4%) and insufficiency (23.2%) were more prevalent.

Age (p = 0.004), sex distribution (p = 0.025), and hypertension (p = 0.030) differed significantly with respect to vitamin D status (Table 1). Among the blood gas indices, O₂ saturation differed significantly among vitamin D categories (p = 0.013) (Table 2). Among the CBS indices, the platelet count (p = 0.018) and mixed cell percentage (p = 0.017) differed significantly among vitamin D categories (Table 2). There were also significant differences in terms of CRP (p < 0.001) and RPR (p = 0.014) among vitamin D categories (Table 2). After controlling for confounders, significant differences were observed in PR (p = 0.009), pH (p = 0.025), PaO₂ (p = 0.005), O₂ saturation (p = 0.006), admission duration (p = 0.004), CRP (p = 0.002), and INR (p = 0.027) among vitamin D categories (Table 2).

Of the 452 patients, 326 (72.1%) had a partial improvement in symptoms (median and IQR for duration were 6 and 5 days, respectively), 102 (22.6%) died (median and IQR for duration were 9 and 9.5 days, respectively), 19 (4.2%) were discharged with consent, 3 (0.7%) had no improvement, and 1 (0.2%) was referred to another center. Serum vitamin D levels were similar between the groups (p = 0.310). Therefore, vitamin D was excluded from the model. Regression analysis was performed to assess the relationships between the study parameters and dichotomized outcome (deceased or not deceased, while excluding data for patients lost during follow up due to referral or discharge, n = 19) (Table 3). COVID-19 mortality was related to serum urea (p = 0.002, OR = 1.020, 95% CI: 1.008–1.033) among laboratory assessments, pulse rate (p = 0.015, OR = 1.047, 95% CI: 1.009–1.086) among clinical findings, and age (p = 0.002, OR = 1.076, 95% CI: 1.027–1.127) among clinical findings. A one unit increase in serum urea was related to 1.4% increase in the risk of death, while a 1-year increase in age and one beat/minute increase in pulse rate were related to 4.9% and 2.5% increases in the risk of death due to COVID-19, respectively.

The cut-offs and predictors of mortality in the study patients were determined using a classification tree. The accuracy, sensitivity, specificity, PPV, and NPV for the trained model (on 347 patients) were 86.2%, 93.2%, 63.4%, 89.2%, and 74.3%, respectively. The model was tested on data for 105 additional patients and the accuracy, sensitivity, specificity, and positive and negative predictive values were 80.2%, 87.9%, 55.0%, 86.6%, and 57.9%, respectively. The classification tree findings revealed that 10 variables were important determinants of mortality in patients with COVID-19 (Figure 1). Among the 10 variables, serum urea level below 83 mg/dL, admission duration of less than 22 days, serum vitamin D level higher than 6.9 ng/dL at admission, RDW less than 14, PLR less than 257, INR less than 1.3, and neutrophil count higher than 83 (× 1000/dL) were associated with survival, while pulse rate higher than 86/min, serum potassium less than 4.5 mg/dL, and PT higher than 15 were predictors for mortality.

4 Discussion

In this study of individuals who were hospitalized with confirmed SARS-Cov-2 infection, we detected significant differences in pulse rate, O₂ saturation, PaO₂, admission duration, pH, CRP, and RPR among vitamin D categories. Although the differences in pH and pulse rate were statistically significant, no significant changes in clinical findings were observed. Among the included patients, 102 (22.6%) died, 326 (72.1%) showed a partial improvement in symptoms, 19 (4.2%) were discharged with consent, 3 (0.7%) had no improvement, and 1 (0.2%) was referred to another center. Furthermore, we detected significant associations between mortality and urea, pulse rate, and age. A classification tree also revealed that serum urea, admission duration, serum vitamin D level, RDW, PLR, INR, and neutrophil count were related to survival, and serum potassium, pulse rate, and PT were related to mortality due to COVID-19.

The majority of clinical and laboratory parameters did not differ significantly among serum vitamin D categories. D'Avolio et al. observed lower serum vitamin D concentrations in patients with confirmed SARS-CoV-2 infection [16]. Radujkovic et al. showed a significant correlation between a vitamin D deficiency and the severity and mortality of COVID-19 [17]. It should be noted that the cut-off values for vitamin D deficiency or sufficiency are subject to debate. In our study, 17.3% of patients had serum vitamin D concentrations below 10 ng/mL and 29.4% had serum vitamin D concentrations between 10 and 20 ng/mL. The vitamin D status was defined based on a previous study in which the cut-offs were defined based on the impact on bone health [18].

Our results suggest that patients with a vitamin D deficiency have lower O₂ saturation and PaO₂ and therefore need more oxygen therapy. Similarly, Faul et al. reported that low serum vitamin D concentrations in patients with COVID-19 are related to a significantly elevated risk of invasive mechanical ventilation [19]. Furthermore, in a study of 185 patients with COVID-19, a vitamin D deficiency increased the risk of invasive mechanical ventilation and the need for more oxygen therapy after adjusting for sex, comorbidities, and age [17]. These findings appeared to be consistent with our observations. The observed relationship might be related to the cytokine storm observed in SARS-CoV-2 infection and the anti-inflammatory effects of vitamin D in this regard [20]. Potential mechanisms by which vitamin D controls inflammation are presented in the following paragraphs.

In this study, RPR was significantly lower among patients with sufficient levels of vitamin D than in those with < 30 ng/mL serum vitamin D. RPR has been identified as a novel indicator of inflammation severity based on the ratio of RDW to platelets [21]. The anti-inflammatory function of vitamin D could explain the protective role of 1,25(OH)D against cytokine storm in patients with COVID-19 [22]. Our results were consistent with the anti-inflammatory effects attributed to vitamin D. Recently, RPR was recognized as a marker of inflammation in acute pancreatitis and myocardial infarction and as a powerful predictor of hepatitis and hepatic fibrosis [23]. Patients with major burn injuries may exhibit a high RPR [24]. An elevated RPR is also associated with levels of inflammatory markers in patients with systemic lupus erythematosus [25]. Accordingly, RPR could be considered an indicator of inflammation. To the best of our knowledge, the relationship between the RPR and vitamin D status has not been assessed in COVID-19. The low serum vitamin D levels among patients with elevated RPR could be explained by the inflammation-modifying role of vitamin D via suppressing the production of IL-2, interferon, and TNF-α as well as through the induction of anti-inflammatory cytokines, such as IL-10 [26]. Low serum vitamin D concentrations were correlated with a rise in inflammatory markers and inhibition of RBC maturation. Moreover, pro-inflammatory cytokines deform RBC membranes and decrease the RBC half-life, along with increases in RDW and O₂ capacity [27]. The proposed mechanism is summarized in Figure 2.

Many studies have evaluated the link between serum vitamin D and CRP. Li et al. analyzed laboratory data for 1873 individuals (NHANES, 2007–2008) and found a strong association between high CRP levels and severe vitamin D deficiency [28]. A study of COVID-19 showed that patients with severe vitamin D deficiency have a 1.4 times higher risk of high CRP levels [22]. Similarly, Daneshkhah et al. observed an inverse relationship between hs-CRP and serum vitamin D concentrations; they also reported that when compared with that for moderate COVID-19, severe COVID-19 could increase the risk of high CRP by 44.5% [29]. In this study, CRP levels were lower among patients with sufficient vitamin D than in those with less than 20 ng/mL serum vitamin D. This finding could be attributed to the anti-inflammatory effect of vitamin D, and this is supported by the low CRP levels observed in our study. Low serum vitamin D concentrations induce the production of IL-6, IL-1β, and TNF-α by means of intercellular calcium activity. As a result, cytokine storm and CRP elevation may occur in patients with COVID-19 [29, 30].

We detected a significant relationship between blood urea nitrogen (BUN) and COVID-19 mortality. BUN is one of the main indicators of renal function and neurohormonal activity [31]. Consistent with our findings, Cheng et al. reported that increased BUN was a good predictor of mortality in patients with COVID-19 [32]. Similarly, another study suggested that NLR and BUN/Cr could predict the severity of COVID-19 and COVID-19 -related mortality [33].

Coronavirus enters cells using ACE2 receptors, which are expressed in the kidney and lungs. SARS-CoV-2 interferes with kidney function in an ACE2-dependent manner. As a result, the renin-angiotensin-aldosterone system is activated and enhances water and sodium absorption in renal tubules, leading to the passive reabsorption of BUN [29]. Renin-angiotensin-aldosterone system activation also results in renal vasoconstriction and therefore glomerular filtration and BUN excretion [34]. While creatinine is an important marker for renal function, we did not observe a significant association between COVID-19 mortality and serum creatinine. Further studies are needed to explore the precise relationships between creatinine and BUN and COVID-19 mortality.

Overall, the study findings indicated that vitamin D has predictive value for the severity of COVID-19; however, its performance is negligible compared with those of more powerful predictors, such as BUN, age, and vital signs at admission. Our study had some limitations that should be noted. First, it was a single-center study although the hospital was one of the three centers for COVID-19 referrals in the province. Second, owing to the cross-sectional design, we could not establish causation. Finally, several unmeasured confounding factors were not recorded or could not be controlled. Therefore, larger, multi-center studies are needed to assess the relationship between severity and mortality due to COVID-19 and renal function tests, including possible confounders. Furthermore, our results indicate the need for cohort studies to assess the cause of the observed clinical and paraclinical findings in patients with COVID-19. Owing to the high admission load in the first two peaks of COVID-19 and the limitations set for external researchers in accessing patients and records, some variables could not be collected for all patients. Therefore, statistical analyses were performed on available variables, which might result in the exclusion of some patients in regression models. However, considering the unknown behavior of the disease, data imputation was not performed. Nevertheless, considering the number of patients in the models, the findings could still be considered valid. The authors also acknowledge that the amount of vitamin D can be affected by sun exposure; however, considering the high prevalence of vitamin D deficiency among the 20- to 50-year-old Iranian population (72%) and the reduced outdoor activities due to the COVID-19 quarantine and movement restrictions, the effect of sun exposure was considered minimal in this study [35].

A strength of the study is that it addressed the limited information on clinical, laboratory, and demographic characteristics of patients with COVID-19 in Iran and their associations with outcomes. Furthermore, owing to the large number of patients who were referred to this center, the findings could be generalized to the COVID-19 patient population of the city and even province. Another strength of our study is the inclusion of the newly proposed indices in the assessment of the outcome of COVID-19.

5 Conclusion

Serum urea was significantly associated with COVID-19 in-hospital mortality. Additionally, levels of inflammatory markers, including CRP and RPR, were higher in patients with vitamin D deficiency than in patients with normal serum vitamin D concentrations. Larger cross-sectional or cohort studies are necessary to further investigate the association between serum vitamin D levels and clinical outcomes in COVID-19 before proceeding to clinical trials. Furthermore, serum urea at admission may serve as a potential risk marker for in-hospital mortality among patients with COVID-19 in conjunction with general risk factors for disease severity; however, further studies are needed to confirm its predictive value. It may be useful to include serum urea and vitamin D levels in the assessment of all admitted patients with COVID-19.

Author Contributions

Zahra Khorasanchi: conceptualization (equal), project administration (equal), validation (equal), writing – review and editing (equal). Ali Jafazadeh Esfehani: data curation (equal), formal analysis (equal), methodology (equal), software (equal), visualization (equal). Mohammad Reza Shadmand Foumani Moghadam: project administration (equal), visualization (equal), writing – review and editing (equal). Sara Shojaei Saadatqoli: project administration (equal). Payam Sharifan: data curation (equal), writing – review and editing (equal). Majid Ghayour Mobarhan: conceptualization (equal), supervision (equal), validation (equal). Sajjad Arefinia: project administration (equal). Afshin Roghani: writing – original draft (equal), writing – review and editing (equal). Naghme Mirhossini: writing – original draft (equal), writing – review and editing (equal). Masoud Pezeshki Rad: investigation (equal), validation (equal). Saeid Eslami: data curation (equal), formal analysis (equal). Hamidreza Naderi: writing – original draft (equal), writing – review and editing (equal). Hassan Vatanparast: supervision (equal), validation (equal), writing – original draft (equal), writing – review and editing (equal). Reza Rezvani: conceptualization (equal), supervision (equal), writing – review and editing (equal).

Ethics Statement

The study was approved by the Ethical Committee of the Mashhad University of Medical Sciences (No. IR.MUMS.REC.1398.314).

Consent

Informed consent was obtained from all participants.

Conflicts of Interest

The authors declare no conflicts of interest.