Low 25-Hydroxyvitamin D Levels on Admission to the Intensive Care Unit May Predispose COVID-19 Pneumonia Patients to a Higher 28-Day Mortality Risk: A Pilot Study on a Greek ICU Cohort

We aimed to examine whether low intensive care unit (ICU) admission 25-hydroxyvitamin D (25(OH)D) levels are associated with worse outcomes of COVID-19 pneumonia. This was a prospective observational study of SARS-CoV2 positive critically ill patients treated in a multidisciplinary ICU. Thirty (30) Greek patients were included, in whom 25(OH)D was measured on ICU admission. Eighty (80%) percent of patients had vitamin D deficiency, and the remaining insufficiency. Based on 25(OH)D levels, patients were stratified in two groups: higher and lower than the median value of the cohort (15.2 ng/mL). The two groups did not differ in their demographic or clinical characteristics. All patients who died within 28 days belonged to the low vitamin D group. Survival analysis showed that the low vitamin D group had a higher 28-day survival absence probability (log-rank test, p = 0.01). Critically ill COVID-19 patients who died in the ICU within 28 days appeared to have lower ICU admission 25(OH)D levels compared to survivors. When the cohort was divided at the median 25(OH)D value, the low vitamin D group had an increased risk of 28-day mortality. It seems plausible, therefore, that low 25(OH)D levels may predispose COVID-19 patients to an increased 28-day mortality risk.


Introduction
Vitamin D deficiency is known to aggravate the incidence and outcome of infectious complications, especially in patients admitted to the intensive care unit (ICU) [1]. Host defense against intracellular pathogens depends upon innate and adaptive antimicrobial effector pathways. One such important pathway is the Toll-like receptor (TLR) pathway, which is activated by 25-hydroxyvitamin D (25(OH)D) [2,3]. More specifically, 25(OH)D regulates the expression of the antimicrobial peptides cathelicidin and β-defensin, which may help improve endothelial barrier function [4].
In our cohort of critically ill COVID-19 pneumonia patients, six (20%) patients were vitamin D insufficient (20-29.9 ng/mL), and the remaining 80% were deficient (<19.9 ng/mL); no patient exhibited vitamin D sufficiency (>30 ng/mL). Therefore, given the high rate of vitamin D deficiency in our patient cohort and the limited sample size, we classified the patients according to the median of the whole cohort (low, <15.2 ng/mL, N = 15) and patients with 25(OH)D levels ≥15.2 ng/mL (high, N = 15). Most experts agree that anyone with a 25(OH)D level of less than 15 ng/mL needs more vitamin D; indeed, the Institute of Medicine has suggested that levels of 16 ng/mL meet the needs of approximately half the population [8]. Demographics, clinical and biochemical characteristics on ICU admission and important outcomes of the two patient groups are listed in Table 2. This table serves for descriptive purposes only. We note that potential confounders of differences observed are not accounted for. The two groups differed only with respect to 28-day mortality. Overall ICU mortality in the two groups did not differ; however, vitamin D-low patients died within 20 ± 7 days, whereas vitamin D-high patients within 44 ± 7 days, from ICU admission (p = 0.001).   Median 25(OH)D levels of the survivors were significantly higher than those of the non-survivors. For the survivors, the median was 16.7 ng/mL (interquartile range (IQR) of 10.6-20.7), whereas at the time of ICU admission, median 25(OH)D levels for the non-survivors were 9.4 ng/mL (IQR of 6.2-13.2; p = 0.03; Figure 1A). Hence, both continuous and categorical 25(OH)D levels on ICU admission differed between survivors (N = 25) and non-survivors (N = 5). Figure 1 depicts the difference observed in 25(OH)D levels on ICU admission (continuous; p = 0.03; Figure 1A and categorical; p = 0.02; Figure 1B). As seen in Figure 1B, all patients who died belonged to the low vitamin D group. Finally, the Kaplan-Meier method was performed for 28-day survival absence probability estimation. The ICU cohort was independently dichotomized above and below the vitamin D median: a high vitamin D group (≥ 15.2 ng/mL) and a low vitamin D group (< 15.2 ng/mL). The probability of survival absence with time was significantly elevated in the low vitamin D group (logrank test, p = 0.01; Figure 2). All patients belonging to the high group survived, while all nonsurvivors belonged to the low group. A Cox survival analysis was also performed; no parameter was identified as an independent predictor of survival.  Finally, the Kaplan-Meier method was performed for 28-day survival absence probability estimation. The ICU cohort was independently dichotomized above and below the vitamin D median: a high vitamin D group (≥15.2 ng/mL) and a low vitamin D group (<15.2 ng/mL). The probability of survival absence with time was significantly elevated in the low vitamin D group (log-rank test, p = 0.01; Figure 2). All patients belonging to the high group survived, while all non-survivors belonged to the low group. A Cox survival analysis was also performed; no parameter was identified as an independent predictor of survival. Finally, the Kaplan-Meier method was performed for 28-day survival absence probability estimation. The ICU cohort was independently dichotomized above and below the vitamin D median: a high vitamin D group (≥ 15.2 ng/mL) and a low vitamin D group (< 15.2 ng/mL). The probability of survival absence with time was significantly elevated in the low vitamin D group (logrank test, p = 0.01; Figure 2). All patients belonging to the high group survived, while all nonsurvivors belonged to the low group. A Cox survival analysis was also performed; no parameter was identified as an independent predictor of survival.

Discussion
In this observational, single center study, we demonstrated that COVID-19 non-survivors had lower ICU admission 25(OH)D levels compared to survivors, implying a possible association of low 25(OH)D levels with poor prognosis of COVID-19 pneumonia patients.
To the best of our knowledge, this is the first, albeit, pilot study in ICU patients showing an association between 25(OH)D levels and 28-day mortality. In another pilot study with a similar number of patients, low serum levels of vitamin C and 25(OH)D were found in most of the critically ill COVID-19 ICU patients. However, older age and low vitamin C levels appeared to be co-dependent risk factors for mortality [9]. A study performed in a respiratory intermediate care unit (RICU), showed similar results to ours, i.e., survival analysis pointed out that, after 10 days of hospitalization, severe vitamin D deficiency patients (<10 ng/mL) had a 50% mortality probability [10]. A very recent study demonstrated that 25(OH)D levels are very low in critically ill COVID-19 patients, and that the corresponding inflammatory response and fatality rates are higher compared to asymptomatic carriers [11].
The remaining studies focused on patients hospitalized in wards. More specifically, vitamin D deficiency has been associated with the progression and severity of COVID-19, such as higher risk of invasive mechanical ventilation and/or death [12][13][14][15][16][17]. Thus, it has been stipulated that diagnosis of vitamin D deficiency could be helpful in assessing patients' potential of developing severe COVID-19, defined as presence of ARDS and/or mechanical ventilation, ICU admission versus ward admission and lower probability of survival. In the elderly, patients with low 25(OH)D levels exhibited elevated cytokine storm markers, resulting in hypoxia that required non-invasive ventilatory support. The study, however, was underpowered to detect a significant difference in mortality [18]. Bolus 25(OH)D supplementation during or just before COVID-19 was associated in frail elderly with less severe COVID-19 and better survival rate [19,20]. In the general population, it has been suggested that low plasma 25(OH)D levels constitute an independent risk factor for COVID-19 infection and hospitalization [21], while significantly lower 25(OH)D levels have been observed in SARS-CoV2-positive patients compared with negative patients [22].
Complex associations and interactions, along with a variety of risk factors for low 25(OH)D levels, make it difficult to prove cause and effect of 25(OH)D on the outcomes of COVID-19 ICU patients. Exploratory studies with carefully chosen matched control groups are of outmost importance.
The limitations of our study should be stated. This pilot study concerned a Greek-only ethnicity population of a limited size in a single ICU. Potential confounders on low 25(OH)D levels were not accounted for. We also had one single 25(OH)D value on ICU admission; we did not have 25(OH)D measurements prior to ICU admission or serial measurements during ICU stay. We were able, however, to demonstrate that 28-day ICU non-survivors had lower continuous and categorical 25(OH)D admission levels than survivors.