Low serum 25-hydroxyvitamin D (25D) levels in patients hospitalised with COVID-19 are associated with greater disease severity: results of a local audit of practice. (preprint)

Objectives: To audit implementation of a local protocol for the treatment of vitamin D deficiency (VDD) among patients hospitalized for Coronavirus Disease 2019 (COVID-19), including an assessment of the prevalence of VDD in these patients, and of potential associations with disease severity and fatality. Design: This was not a study or clinical trial, but rather a retrospective interim audit (Newcastle-upon-Tyne Hospitals Registration No. 10075) of a local clinical care pathway for hospitalized patients with COVID-19-related illness. The Information (Caldicott) Guardian permitted these data to be shared beyond the confines of our institution. Setting: A large tertiary academic NHS Foundation Trust in the North East of England, UK, providing care to COVID-19 patients. Participants: One hundred thirty-four hospitalized patients with documented COVID-19 infection. Main outcome measures: Adherence to local investigation and treatment protocol; prevalence of VDD, and relationship of baseline serum 25(OH)D with markers of COVID-19 severity and inpatient fatality versus recovery. Results: 55.8% of eligible patients received Colecalciferol replacement, albeit not always loaded as rapidly as our protocol suggested, and no cases of new hypercalcaemia occurred following treatment. Patients admitted to ITU were younger than those managed on medical wards (61.1 years +/- 11.8 vs. 76.4 years +/- 14.9, p<0.001), with greater prevalence of hypertension, and higher baseline respiratory rate, National Early Warning Score-2 and C-reactive protein level. While mean serum 25(OH)D levels were comparable [i.e. ITU: 33.5 nmol/L +/- 16.8 vs. Non-ITU: 48.1 nmol/L +/- 38.2, mean difference for Ln-transformed-25(OH)D: 0.14, 95% Confidence Interval (CI) (-0.15, 0.41), p=0.3], only 19% of ITU patients had 25(OH)D levels greater than 50 nmol/L vs. 60.9% of non-ITU patients (p=0.02). However, we found no association with fatality, potentially due to small sample size, limitations of no-trial data and, potentially, the prompt diagnosis and treatment of VDD. Conclusions: Subject to the inherent limitations of observational (non-trial) audit data, analysed retrospectively, we found that patients requiring ITU admission were more frequently vitamin D deficient than those managed on medical wards, despite being significantly younger. Larger prospective studies and/or clinical trials are needed to elucidate the role of vitamin D as a preventive and/or therapeutic strategy for mitigating the effects of COVID-19 infection in patients with VDD.

COVID-19 management in a UK NHS Foundation Trust with a High Consequence Infectious Diseases centre: a detailed descriptive analysis (preprint)

BackgroundRecent large national and international cohorts describe the baseline characteristics and outcome of hospitalised patients with COVID-19, however there is little granularity to these reports. We aimed to provide a detailed description of a UK COVID-19 cohort, focusing on clinical decisions and patient journeys. MethodsWe retrospectively analysed the management and 28-day outcomes of 316 consecutive adult patients with SARS-CoV-2 PCR-confirmed COVID-19 admitted to a large NHS Foundation Trust with a tertiary High Consequence Infectious Diseases centre in the North of England. FindingsMost patients were elderly (median age 75) with multiple comorbidities. One quarter were admitted from residential or nursing care. Symptoms were consistent with COVID-19, with cough, fever and/or breathlessness in 90.5% of patients. Two thirds of patients had severe disease on admission. Mortality was 81/291 (27.8%). Most deaths were anticipated; decisions to initiate respiratory support were individualised after consideration of patient wishes, premorbid frailty and comorbidities, with specialist palliative care input where appropriate. 22/291 (7.6%) patients were intubated and 11/22 (50%) survived beyond discharge. Multiple logistic regression identified age as the most significant risk factor for death (OR 1.09 [95% CI 1.06 - 1.12] per year increase, p < 0.001). InterpretationThese findings provide important clinical context to outcome data. Deaths were anticipated, occurring in patients with advance decisions on ceilings of treatment. Age was the most significant risk factor for death, confirming that demographic factors in the population are a major influence on hospital mortality rates. FundingFunding was not required.