Low Serum 25-Hydroxyvitamin D (25[OH]D) Levels in Patients Hospitalized With COVID-19 Are Associated With Greater Disease Severity

Grigorios Panagiotou; Su Ann Tee; Yasir Ihsan; Waseem Athar; Gabriella Marchitelli; Donna Kelly; Christopher S. Boot; Nadia Stock; James Macfarlane; Adrian R. Martineau; Graham Burns; Richard Quinton


Clin Endocrinol. 2020;93(4):508-511. 

Dear Editor,

The global pandemic of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is associated with higher fatality in respect of male sex, ageing, obesity, diabetes, hypertension, climatic factors (low ambient temperature and high geographic latitude) and, in the UK and North America, with darker-skinned ethnicities;[1] in all of which circumstances, vitamin D deficiency (VDD) is more common.[2,3]

Vitamin D3 is a preprohormone, whose biosynthetic pathway begins with solar UVB irradiation of 7-dehydrocholesterol in bare skin exposed to strong sunlight, and exhibits multifaceted effects beyond calcium and bone metabolism. Vitamin D receptors are highly expressed in B- and T-lymphocytes, suggesting a role in modulating innate and adaptive immune responses.[4] 25-hydroxyvitamin D [25(OH)D] levels reach their nadir at the end of winter, and low levels are associated with increased risk of acute respiratory tract infections during winter[5] and mitigated by vitamin D supplementation. Clinical trials involving vitamin D supplementation in COVID-19 are ongoing but may not report within the time-frame of this pandemic.

As North East England has a high prevalence of seasonal VDD,[6] physicians in Newcastle upon Tyne Hospitals (NuTH) decided to measure admission serum 25(OH)D levels in patients with COVID-19, so to inform a treatment protocol adjusted according to the severity of baseline deficiency and based on pharmacokinetic data from Romagnoli, et al[7] (Appendix 1). We audited this protocol as soon as practicable (Clinical Governance & Audit Registration No.10075), to determine whether data supported its continuation and whether there might also be lessons for a wider audience.

Serum 25(OH)D levels were measured in 134 (largely Caucasian) inpatients with positive SARS-CoV-2 swab or clinical/radiological diagnosis of COVID-19. A cut-off of >50 nmol/L was defined as normal. Patients with VDD were treated wherever possible. No adverse effects, such as hypercalcaemia, were reported after treatment. Clinical observations at presentation (NEWS-2 score, heart rate, respiratory rate, blood pressure and temperature), and markers of inflammatory response [C-reactive protein (CRP), procalcitonin] were retrieved from electronic records. Sicker patients were admitted to Intensive Therapy Unit (ITU) and milder cases, or those with ward-based ceilings of care managed on medical wards ('non-ITU group'). Final outcome was recorded as discharge or death. Statistical methods are described in Appendix 2.

Patient characteristics are summarized in Table 1. The majority of COVID-19 inpatients (ie 90/134 patients or 66.4%) had vitamin D insufficiency (25–50 nmol/L); 50/134 (37.3%) were deficient (<25 nmol/L), and 29/134 (21.6%) had severe deficiency (≤15 nmol/L).

ITU patients were younger (61.1 years ± 11.8 vs non-ITU: 76.4 years ± 14.9, P < .001), more frequently hypertensive, and had higher NEWS-2 score (P = .01), respiratory rate and CRP levels at presentation (Table 1). 25(OH)D levels were not associated with increased oxygen requirements, NEWS-2 score, COVID-19 radiological findings, CRP levels, or presence of co-morbidities (P > .05 for all).

ITU patients had lower 25(OH)D levels compared with non-ITU patients despite being younger, (33.5 nmol/L ± 16.8 vs non-ITU: 48.1 nmol/L ± 38.2; mean difference for logarithmically transformed-25(OH)D: 0.14; 95% Confidence Interval (CI): −0.15, 0.41), albeit not reaching statistical significance (P = .3) possibly due to limited sample size. Nevertheless, ITU patients exhibited a significantly higher prevalence of VDD, with only 19% being vitamin D replete compared with 39.1% of non-ITU patients (P = .02).

Overall, 63/113 (55.8%) of eligible patients received treatment. Of these, 33/63 patients (52.4%) were treated as per protocol and the rest given lower doses. Outcome data were available for 110/134 patients (82.1%) at the time of reporting. 94 (85.5%) patients were discharged, 16 (14.5%) died; and 24 are still receiving inpatient care. Serum 25(OH)D levels were not associated with mortality [95% CI 0.97 (0.42, 2.23), P = .94]. Further adjustments for potential covariates including age, gender, co-morbidities and CRP levels did not affect these results.

Mortality from COVID-19 is caused by severe acute respiratory syndrome, with cytokine storm and diffuse micro- and macrovascular thrombosis. Vitamin D may reduce severity of respiratory tract infections via three putative mechanisms: maintaining tight junctions, killing enveloped viruses through induction of cathelicidin and defensins and reducing pro-inflammatory cytokine production, thereby decreasing risk of cytokine storm.[8] Therefore, identifying and treating VDD may represent a promising modality for mitigating COVID-19-associated fatality.

Previous publications have highlighted potential associations between VDD and COVID-19 mortality.[9] We found no significant association between VDD and mortality, which was not unexpected given our proactive treatment protocol, small sample size and observational nature of our analysis.

In a small US study, 84.6% (11/13) ITU patients had VDD compared with 57.1% of patients on medical wards.[10] Only 19% of our ITU patients were vitamin D replete, despite being significantly younger and having fewer VDD-associated co-morbidities; challenging the dogma that VDD is a problem of the elderly. This may have implications for public health advice, especially given recent limitations on sun exposure resulting from lockdown measures.

A recent study from UK Biobank found no association between serum 25(OH)D and risk of COVID-19 infection, but likewise found no association with hypertension and diabetes—both well-established risk factors for fatality—and, moreover, sample collection was not standardized for late winter, when the UK's COVID-19 outbreak began.[11]

This is the first report exploring serum 25(OH)D levels in COVID-19 inpatients in Europe. VDD was more prevalent among patients requiring ITU admission, and thus VDD might be an under-recognized determinant of illness-severity. Strengths of our data include the acute assessment of serum 25(OH)D during COVID-19 admission. Limitations include small, nonethnically diverse sample and observational nature of this audit; cross-sectional analysis does not allow causality to be established, and therefore, our results should be interpreted with caution.

Nevertheless, these preliminary data provide impetus to the commissioning, design and interpretation of ongoing or future clinical trials to evaluate a potential therapeutic role of vitamin D in COVID-19.