Journal of Endocrinological Investigation https://doi.org/10.1007/s40618-021-01639-9 ORIGINAL ARTICLE Vitamin D supplementation and COVID‑19 risk: a population‑based, cohort study J. Oristrell1,2,3 · J. C. Oliva2 · E. Casado4 · I. Subirana5,6 · D. Domínguez7 · A. Toloba6 · A. Balado1 · M. Grau5,6,8 Received: 28 June 2021 / Accepted: 12 July 2021 © Italian Society of Endocrinology (SIE) 2021 Abstract Purpose To analyze the associations between cholecalciferol or calcifediol supplementation, serum 25-hydroxyvitamin D (25OHD) levels and COVID-19 outcomes in a large population. Methods All individuals ≥ 18 years old living in Barcelona-Central Catalonia (n = 4.6 million) supplemented with cholecal- ciferol or calcifediol from April 2019 to February 2020 were compared with propensity score-matched untreated controls. Outcome variables were SARS-CoV2 infection, severe COVID-19 and COVID-19 mortality occuring during the first wave of the pandemic. Demographical data, comorbidities, serum 25OHD levels and concomitant pharmacological treatments were collected as covariates. Associations between cholecalciferol or calcifediol use and outcome variables were analyzed using multivariate Cox proportional regression. Results Cholecalciferol supplementation (n = 108,343) was associated with slight protection from SARS-CoV2 infection (n = 4352 [4.0%] vs 9142/216,686 [4.2%] in controls; HR 0.95 [CI 95% 0.91–0.98], p = 0.004). Patients on cholecalciferol treatment achieving 25OHD levels ≥ 30 ng/ml had lower risk of SARS-CoV2 infection, lower risk of severe COVID-19 and lower COVID-19 mortality than unsupplemented 25OHD-deficient patients (56/9474 [0.6%] vs 96/7616 [1.3%]; HR 0.66 [CI 95% 0.46–0.93], p = 0.018). Calcifediol use (n = 134,703) was not associated with reduced risk of SARS-CoV2 infec- tion or mortality in the whole cohort. However, patients on calcifediol treatment achieving serum 25OHD levels ≥ 30 ng/ml also had lower risk of SARS-CoV2 infection, lower risk of severe COVID-19, and lower COVID-19 mortality compared to 25OHD-deficient patients not receiving vitamin D supplements (88/16276 [0.5%] vs 96/7616 [1.3%]; HR 0.56 [CI 95% 0.42–0.76], p < 0.001). Conclusions In this large, population-based study, we observed that patients supplemented with cholecalciferol or calcifediol achieving serum 25OHD levels ≥ 30 ng/ml were associated with better COVID-19 outcomes. Keywords Cholecalciferol · Calcifediol · Vitamin D · COVID-19 · SARS-CoV2 5 * J. Oristrell CIBER of Epidemiology and Public Health (CIBERESP), joristrell@tauli.cat Barcelona, Catalonia 6 * E. Casado Institut Hospital del Mar d’Investigacions Mèdiques, ecasado@tauli.cat Barcelona, Catalonia 7 1 Agència de Qualitat i Avaluació Sanitària, Generalitat de Internal Medicine Service, Corporació Sanitària Parc Taulí, Catalunya, Barcelona, Catalonia Parc Taulí s/n, 08208 Sabadell, Barcelona, Catalonia 8 2 Department of Medicine, University of Barcelona, Barcelona, Institut d’Investigació i Innovació I3PT, Sabadell, Catalonia Catalonia 3 Department of Medicine, Universitat Autònoma de Barcelona, Barcelona, Catalonia 4 Rheumatology Service, Corporació Sanitària Parc Taulí, Parc Taulí s/n, 08208 Sabadell, Barcelona, Catalonia 13 Vol.:(0123456789) Journal of Endocrinological Investigation Introduction Methods Infection with the new coronavirus SARS-CoV2 is character- Study design and population included ized by an important clinical variability, which suggests that there are important host-related factors that impact in COVID- A retrospective cohort was built using the databases of the 19 outcomes. One of these factors has been postulated to be public healthcare system in Catalonia. We analyzed all indi- vitamin D deficiency [1, 2], which is a prevalent condition viduals ≥ 18 years old living in Barcelona and Central Cata- worldwide [3]. Vitamin D is now being recognized as an lonia regions on 25 February 2020, date of the first positive hormonal system with many extra-skeletal actions, including PCR for SARS-CoV2 in our country (n = 4,643,139). important effects on the immunological system [4–6]. In this population, we performed three independent Several clinical trials and two meta-analysis have shown studies to investigate the association of cholecalciferol or that cholecalciferol or ergocalciferol supplementation may calcifediol supplementation with COVID-19 outcomes: help prevent acute respiratory infections [7, 8]. However, evi- dence that these drugs are helpful to treat or prevent SARS- Comparison of COVID outcomes between supplemented CoV2 infection is still controversial [9]. patients and propensity score-matched controls: We iden- Observational studies have described an association tified all patients receiving cholecalciferol (n = 201,445) between low serum levels of 25-hydroxyvitamin D (25OHD) or calcifediol (n = 207,136) supplementation from 1 April and higher risk of SARS-CoV2 infection [10–12], higher risk 2019 to 28 February 2020 and patients not receiving of severe COVID-19 or higher COVID-19 mortality [13–15]. any vitamin D supplement (4,267,430) during the same Three meta-analysis have also concluded that there is an asso- period. Since chronic kidney disease (CKD) is a strong ciation between low serum levels of 25OHD and higher risk of predictor of worse prognosis in COVID-19 [27], subjects COVID-19 mortality [16–18]. However, studies analyzing the without an available serum creatinine determination per- use of cholecalciferol or calcifediol supplementation to modify formed between 1 October 2018 and 28 February 2020 COVID-19 outcomes have offered inconclusive results. While were excluded from the study. After propensity score some observational studies in hospitalized patients have shown matching (see below), 108,343 patients on cholecalcif- reduced COVID-19 severity or mortality in patients supple- erol, 216,686 matched controls (cholecalciferol controls), mented with cholecalciferol or calcifediol [19–21], Cereda 134,703 patients on calcifediol and 269,406 matched con- et al. described a trend to an increased mortality in patients trols (calcifediol controls) were selected for the analysis. supplemented with calcifediol [22]. Association between mean daily cholecalciferol or cal- The effects of cholecalciferol or calcifediol as a treatment cifediol dose and COVID-19 outcomes: All patients for hospitalized COVID-19 patients have also been studied in receiving cholecalciferol or calcifediol supplementa- three low-powered clinical trials, without observing any sig- tion from 1 November 2019 to 28 February 2020, with nificant reduction in COVID-19 mortality [23–25]. an available serum creatinine level (n = 165,588 and To the best of our knowledge, there are not published results 132,590, respectively), were selected for this analysis. describing the effects, at the population level, of cholecalcif- This shorter period of time was chosen to minimize the erol or calcifediol supplementation on COVID-19 outcomes. effects of eventual changes in the dose of these drugs. To ascertain whether vitamin D supplementation protects Comparison of COVID-19 outcomes between chole- against SARS-CoV2 infection or from COVID-19 adverse calciferol- or calcifediol-supplemented patients with outcomes, we designed this large observational study. We have a sufficient vitamin D status (serum 25OHD > 30 ng/ already shown that calcitriol supplementation, the active form ml) and unsupplemented vitamin D-deficient (serum of vitamin D, mainly used in patients with advanced renal fail- 25OHD < 20 ng/ml) patients: In order to reduce the vari- ure, was associated with important reductions in COVID-19 ability in serum 25OHD levels due to seasonal sun expo- mortality [26]. Now, we present the results of cholecalciferol sure, we only analyzed serum levels determined between or calcifediol supplementation showing also beneficial effects 1 November 2019 and 28 February 2020. All patients of of these drugs in patients that reach normal serum 25OHD the cohort that had a serum 25OHD determination in this levels. period of time (n = 85,158) were included in this analysis. Data sources Given Catalonia’s universal health and medication cov- erage, we were able to utilize electronic databases to 13 Journal of Endocrinological Investigation examine the association of cholecalciferol and calcifediol tolerated difference between matched patients. To examine use with COVID-19 outcomes in a real world setting. We the balance of each covariate between the treatment and the used anonymized data provided by the Catalan Agency for control group, the standardized mean difference was cal- Health Quality and Evaluation (AQUAS) within the frame- culated before and after matching using Tableone package work of the Data Analytics Program for Health Research in R [30]. We considered the groups well balanced if the and Innovation (PADRIS). PADRIS databases include standardized mean difference was < 0.10 for each covariate. information on demographics (age and sex), diagnoses, laboratory data, drugs supplied by pharmacies, Primary Serum levels of 25‑hydroxyvitamin D Care physician diagnoses, laboratory results and diagno- ses, procedures and outcomes of medical admissions in the Serum levels of 25OHD determined in the laboratories of the public hospitals in Catalonia. This project was approved catalan public health system between 1 November 2019 and in a public call for grants for using PADRIS databases in 28 February 2020 in the whole cohort were obtained from research projects on COVID-19. PADRIS databases. A deficient vitamin D status was defined as a serum 25OHD level < 20 ng/mL and a sufficient vitamin Identification of patients on cholecalciferol D status was defined as a serum 25OHD level ≥ 30 ng/mL. or calcifediol supplementation Outcome variables Patients who had been supplied in pharmacies with drugs of the Anatomical Therapeutic Chemical Classification We analyzed the occurence of SARS-CoV2 infection, System groups A11CC05, A12AX, M05BB03, M05BB07, COVID-19 hospitalization, intensive care admission, the M05BB08, M05BB09, A11CC06 or A11CC55 from 1 procedures during hospitalization and mortality during the April 2019 to 28 February 2020 were analyzed. The sum first wave of the pandemic. Four main outcome variables of Defined Daily Doses (DDD) of cholecalciferol or the were defined, with different timings due to the natural course sum of calcifediol doses supplied from 1 November 2019 to of the disease: 28 February 2020 were identified, transformed into micro- grams, and the mean daily cholecalciferol or calcifediol dose SARS‑CoV2 infection received per patient, in micrograms, was calculated. Patients receiving formulations containing > 250 μg of cholecalcif- Positive PCR result for SARS-CoV2 or a clinical diagnosis erol (12.5 DDD) or > 250 μg of calcifediol per dose were made by a Primary Care physician, or a hospital discharge considered as receiving bolus doses. report stating a diagnosis of COVID-19 (ICD-10 codes used are displayed in Supplementary Table 1), from 25 February Identification of control subjects 2020 to 30 April 2020. Time (in days) from 24 February through propensity score matching 2020 until a positive PCR or a clinical diagnosis (the first event) was used for survival analysis. Censored time for We performed two independent propensity score matching those individuals without the event was the time from 24 to build the control groups for cholecalciferol and calcife- February to 30 April 2020. diol using the ’Matching’ package in R [28] as described [26]. First, we used multivariate logistic regression to model COVID‑19 mortality receiving or not each drug as a function of the following covariates: sex, age, fifteen comorbidities identified from Patients diagnosed with COVID-19 infection resulting in the International Classification of Diseases (ICD-10) diag- death between 25 February and 15 May 2020. Patients with nostic codes issued by family physicians (Supplementary COVID-19 admitted to hospital before 16 May 2020 and Table 1), estimated glomerular filtration rate (eGFR), his- dying before 7 June were also included. Time (in days) from tory of cigarette smoking, nursing home residence and use 24 February 2020 to COVID-19 death was used for survival of seven classes of drugs that could potentially affect the analysis. Censored time for those individuals without the prognosis (Supplementary Table 1). Estimated glomerular event was the time from 24 February to 7 June 2020. filtration rate was obtained from serum levels of creatinine, sex and age according to the Chronic Kidney Disease Epide- Severe COVID‑19 miology Collaboration (CKD-EPI) equation [29]. Propensity scores were matched using the nearest-neighbor matching Composite outcome of COVID-19 mortality, as already method without replacement at a 1:2 ratio of treated subjects defined, or COVID-19 hospital admission needing non-inva- and controls. A caliper of 0.2 of the standard deviation of sive mechanical ventilation, orotracheal intubation, mechanical the propensity score logit was established as the maximum ventilation or intensive care unit admission from 25 February 13 Journal of Endocrinological Investigation 2020 to 15 May 2020. Time (in days) from 24 February 2020 Results until hospital admission (if severe COVID-19 developed dur- ing hospitalization) or time (in days) from 24 February 2020 COVID‑19 outcomes until COVID-19 death was used for survival analysis. Cen- in cholecalciferol‑supplemented patients sored time for those individuals without the event was the time versus propensity score‑matched controls from 24 February to 7 June 2020. A total of 201,445 patients ≥ 18 years old were identified Statistical analysis as being on cholecalciferol treatment between 1 April 2019 and 28 February 2020 in Barcelona and Central Catalonia Continous variables are reported as mean and standard devia- regions. After propensity-score matching, 108,343 patients tion and qualitative variables are summarized by frequen- on cholecalciferol and 216,686 matched control patients cies and percentages. Basal differences between treated and were included in the analysis. untreated groups were assessed using Student’s t test or chi- Clinical characteristics of the patients treated with square test and standardized mean differences. cholecalciferol and their matched controls are shown in Once the control groups were established, associations Table 1. Mean age was 70 years and more than 83% were between cholecalciferol or calcifediol supplementation and women. SARS-CoV2 infection was diagnosed in 4352 outcome variables were further analyzed using unadjusted and patients supplemented with cholecalciferol (2113 of them multivariate Cox proportional hazards regression models. All [48.6%] confirmed by PCR) and in 9142 untreated con- the variables that approached statistical significance (p < 0.2) trols (4300 of them [47.0%] confirmed by PCR). Chole- were initially selected for inclusion in the adjusted analyses. calciferol use was associated with a mild, but significant, Multivariable models were constructed by means of a stepwise lower risk of SARS-CoV2 infection, both in the univariate forward inclusion procedure and only the significant variables and in the multivariate Cox regression analysis (n = 4352 were retained in the final model. Unadjusted and adjusted haz- [4.0%] versus n = 9142 [4.2%] in controls; HR 0.95 [CI ard ratios and their 95% confidence intervals are reported. 95% 0.91–0.98], p = 0.004). This mild reduction in the Taking into account that vitamin D supplementation may infection rate was not observed in the subgroup of patients be prescribed to treat a low vitamin D status, we also com- diagnosed by PCR (HR 0.98 [0.93–1.04; p = 0.51]. We did pared the outcome variables between patients with sufficient not observe any significant association between cholecal- vitamin D status, while being vitamin D-supplemented, with ciferol supplementation and COVID-19 severity (n = 798 patients deficient in vitamin D and not supplemented, also [0.7%] versus n = 1650 [0.8%] in controls) or COVID-19 using multivariate Cox regression analysis. Finally, the asso- mortality (n = 716 [0.7%] versus n = 1492 [0.7%] in con- ciations between the mean daily cholecalciferol or mean daily trols), when comparing with matched untreated controls calcifediol dose and COVID-19 outcomes were also analyzed (Table 2). using multivariate Cox regression analysis. Patients receiving bolus cholecalciferol had lower For all statistical tests, a p value < 0.05 was used for statisti- COVID-19 mortality (100/20,715; [0.5%]) than patients on cal significance. daily cholecalciferol treatment (616/87628; [0.7%]). How- Descriptive statistics and survival analysis were carried out ever, in the multivariate Cox regression analysis, receiving using SPSS version 25.0 for Windows (SPSS, Chicago, IL, bolus cholecalciferol was not significantly associated with USA), and Survival and Survminer packages in R [31, 32]. lower COVID-19 severity nor mortality. We analyzed the cholecalciferol DDD received in the Ethical issues and confidenciality 165,588 patients that had been supplied between 1 Novem- ber 2019 and 28 February 2020. Mean daily cholecalciferol All data were treated anonymously in order for this study to doses were similar in patients who became infected or died comply with the provisions of Spanish and European laws due to COVID-19 than in patients without these outcomes on Protection of Personal Data. The study was approved by (Table 3). However, in the multivariate Cox regression the ethics committee of the Corporació Sanitària Parc Taulí- analysis, the mean daily dose of cholecalciferol, measured Universitat Autònoma de Barcelona. in 10 μg intervals (equivalent to 400 IU), was associated with small but significant reductions in the risk of SARS- Cov2 infection (HR 0.97 [CI 95% 0.95–0.99], p = 0.007), in SARS-CoV2 infection confirmed by PCR (HR 0.94 [CI 95% 0.91–0.97], p < 0.001), in severe COVID-19 (HR 0.92 [CI 95% 0.87–0.97], p = 0.002) and in COVID-19 mortal- ity (HR 0.90 [CI 95% 0.85–0.96], p = 0.001) (Table 3). 13 Journal of Endocrinological Investigation Table 1  Clinical characteristics of patients on cholecalciferol or calcifediol treatment and their matched controls Cholecalciferol Cholecalcif- SMD Calcifediol treated Calcife- SMD treated (n = 108,343) erol matched (n = 134,703) diol matched (n = 216,686) (n = 269,406) Variables used for matching Female gender, n (%) 90,417 (83.5) 181,414 (83.7) 0.007 105,229 (78.1) 209,824 (77.9) 0.006 Age, mean (SD) 70.0 (14.0) 70.0 (14.6) 0.001 68.8 (14.9) 68.8 (15.1) 0.004 Nursing home residence, n (%) 2884 (2.7) 5649 (2.6) 0.003 3548 (2.6) 6989 (2.6) 0.002 Cigarette smoking, n (%) 23,329 (21.5) 46,003 (21.2) 0.007 32,507 (24.1) 64,571 (24.0) 0.004 Comorbidities Hypertension, n (%) 59,157 (54.6) 116,565 (53.8) 0.016 75,058 (55.7) 148,784 (55.2) 0.010 Obesity, n (%) 39,941 (36.9) 78,711 (36.3) 0.011 51,547 (38.3) 102,367 (38.0) 0.006 Diabetes, n (%) 23,899 (22.1) 46,978 (21.7) 0.009 33,284 (24.7) 66,443 (24.7) 0.001 Heart failure, n (%) 11,572 (10.7) 22,939 (10.6) 0.003 14,692 (10.9) 29,260 (10.9) 0.001 COPD, n (%) 14,838 (13.7) 29,376 (13.6) 0.004 18,431 (13.7) 36,947 (13.7) 0.001 Asthma, n (%) 11,791 (10.9) 23,332 (10.8) 0.004 13,983 (10.4) 27,894 (10.4) 0.001 eGFR, mean (SD) 78.35 (20.65) 78.46 (20.96) 0.005 78.01 (22.91) 78.01 (21.62) < 0.001 Cerebrovascular disease, n (%) 7949 (7.3) 15,674 (7.2) 0.004 10,234 (7.6) 20,346 (7.6) 0.002 Dementia, n (%) 6517 (6.0) 12,783 (5.9) 0.005 7841 (5.8) 15,590 (5.8) 0.001 Malignant neoplasia, n (%) 28,731 (26.5) 57,158 (26.4) 0.003 32,604 (24.2) 65,240 (24.2) < 0.001 Liver cirrhosis, n (%) 1426 (1.3) 2846 (1.3) < 0.001 1641 (1.2) 3346 (1.2) 0.002 Osteoporosis, n (%) 19,332 (17.8) 37,668 (17.4) 0.012 19,052 (14.1) 36,908 (13.7) 0.013 Past femur fracture, n (%) 1746 (1.6) 2939 (1.4) 0.021 1676 (1.2) 2877 (1.1) 0.016 Dyslipidemia, n (%) 57,688 (53.2) 115,410 (53.3) < 0.001 70,817 (52.6) 142,020 (52.7) 0.003 Ischemic heart disease, n (%) 7204 (6.6) 14,097 (6.5) 0.006 10,331 (7.7) 20,784 (7.7) 0.002 Peripheral arteriopathy,n (%) 3085 (2.8) 6064 (2.8) 0.003 4679 ( 3.5) 9372 ( 3.5) < 0.001 Use of drugs Proton pump inhibitors 49,488 (45.7) 98,961 (45.7) < 0.001 58,522 (43.4) 118,032 (43.8) 0.007 Oral corticosteroids, n (%) 11,986 (11.1) 24,149 (11.1) 0.003 11,098 (8.2) 22,562 (8.4) 0.005 DPP4-inhibitors, n (%) 2659 (2.5) 5194 (2.4) 0.004 4220 (3.1) 8440 (3.1) < 0.001 Statins, n (%) 33,471 (30.9) 66,992 (30.9) < 0.001 43,130 (32.0) 86,714 (32.2) 0.004 ACE inhibitors, n (%) 26,353 (24.3) 52,413 (24.2) 0.003 32,101 (23.8) 64,087 (23.8) 0.001 ARB, n (%) 18,988 (17.5) 37,532 (17.3) 0.005 25,486 (18.9) 50,685 (18.8) 0.003 Immunosuppressants, n (%) 3095 (2.9) 6082 (2.8) 0.003 3004 (2.2) 5941 ( 2.2) 0.002 SMD standardized mean difference, SD standard deviation, COPD chronic obstructive pulmonary disease, eGFR estimated glomerular filtration rate, ACE angiotensin convertint enzyme, ARB angiotensin-II receptor blockers, DPP4 dipeptidyl peptidase-4 COVID‑19 outcomes in calcifediol‑supplemented SARS-CoV2 infection was diagnosed in 5662 patients patients versus propensity score‑matched controls supplemented with calcifediol (2607 of them [46.0%] con- firmed by PCR) and in 11,401 untreated controls (5413 of A total of 207,136 patients ≥ 18 years old were identified as them [47.5%] confirmed by PCR). We did not observe any being on calcifediol treatment between 1 April 2019 and 28 significant association between calcifediol supplementation February 2020 in Barcelona-Central Catalonia region. After and the risk of SARS-CoV2 infection (n = 5662 [4.2%] ver- propensity-score matching, 134,703 patients on calcifediol sus 11,401 [4.2%] in matched controls), nor between cal- and 269,406 matched control patients were included in the cifediol use and COVID-19 severity (n = 1037 [0.8%] versus study. Calcifediol was supplied as bolus doses in 99.4% of n = 2073 [0.8%] in controls) or mortality (n = 934 [0.7%] patients or as a daily drop formulation in 0.6% of patients. versus n = 1859 [0.7%] in controls) (Table 4). Clinical variables in patients treated with calcifediol Among calcifediol treated patients, 132,590 had received and their respective control group are shown in Table 1. the drug between 1 November 2019 and 28 February 2020. Mean age of patients was 69 years, with a high proportion In the multivariate Cox regression analysis, the mean daily of women and similar comorbidities to the patients treated calcifediol dose received in this period, measured in 10 μg with cholecalciferol. intervals, was not associated with the risk of SARS-CoV2 13 Table 2  Variables associated with COVID-19 outcomes in patients on cholecalciferol treatment (n = 108,343) and matched controls (n = 216,686) SARS-CoV2 ­infection1 Severe COVID-192 COVID-19 mortality 3 4 3 4 13 Univariate HR p Multivariate HR p Univariate HR p Multivariate HR p Univariate3 HR p Multivariated HR p (CI 95%) (CI 95%) (CI 95%) (CI 95%) (CI 95%) (CI 95%) Cholecalciferol 0.95 (0.92–0.99) 0.007 0.95 (0.91–0.98) 0.004 0.97 (0.89–1.05) 0.439 0.96 (0.88–1.05) 0.366 treatment Female sex 0.88 (0.84–0.92) < 0.001 0.42 (0.39–0.46) < 0.001 0.52 (0.48–0.57) < 0.001 0.42 (0.38–0.46) < 0.001 0.56 (0.50–0.62) < 0.001 Age5 1.14 (1.13–1.16) < 0.001 0.94 (0.92–0.95) < 0.001 2.38 (2.28–2.48) < 0.001 1.48 (1.41–1.55) < 0.001 2.90 (2.77–3.03) < 0.001 1.79 (1.69–1.89) < 0.001 Cigarette smok- 0.93 (0.89–0.97) < 0.001 0.97 (0.88–1.07) 0.554 0.93 (0.83–1.03) 0.147 1.28 (1.13–1.45) < 0.001 ing Nursing home 8.89 (8.51–9.29) < 0.001 6.25 (5.91–6.60) < 0.001 17.49 (16.06– < 0.001 6.20 (5.59–6.87) < 0.001 20.25 (18.55– < 0.001 6.10 (5.49–6.78) < 0.001 residence 19.05) 22.11) Hypertension 1.12 (1.08–1.16) < 0.001 2.37 (2.17–2.60) < 0.001 2.61 (2.37–2.88) < 0.001 Obesity 1.06 (1.03–1.10) 0.001 1.10 (1.06–1.14) < 0.001 1.23 (1.13–1.33) < 0.001 1.25 (1.15–1.36) < 0.001 1.19 (1.09–1.30) < 0.001 1.26 (1.15–1.38) < 0.001 Diabetes 1.22 (1.17–1.27) < 0.001 1.06 (1.02–1.11) 0.004 2.27 (2.09–2.46) < 0.001 1.33 (1.22–1.45) < 0.001 2.35 (2.15–2.55) < 0.001 1.36 (1.25–1.49) < 0.001 Heart failure 1.97 (1.89–2.06) < 0.001 1.33 (1.27–1.40) < 0.001 4.11 (3.77–4.47) < 0.001 1.35 (1.23–1.48) < 0.001 4.55 (4.17–4.96) < 0.001 1.36 (1.23–1.50) < 0.001 COPD 1.42 (1.36–1.49) < 0.001 1.11 (1.06–1.16) < 0.001 2.46 (2.25–2.69) < 0.001 1.17 (1.07–1.29) 0.001 2.54 (2.31–2.78) < 0.001 1.13 (1.02–1.25) 0.021 Asthma 1.25 (1.19–1.31) < 0.001 1.11 (1.06–1.17) < 0.001 1.11 (0.98–1.26) 0.094 1.03 (0.90–1.17) 0.705 eGFR 0.92 (0.92–0.93) < 0.001 0.97 (0.96–0.98) < 0.001 0.70 (0.69–0.71) < 0.001 0.88 (0.86–0.90) < 0.001 0.68 (0.67–0.69) < 0.001 0.88 (0.86–0.91) < 0.001 Cerebrovascular 1.69 (1.61–1.78) < 0.001 1.23 (1.17–1.30) < 0.001 2.85 (2.57–3.16) < 0.001 1.20 (1.08–1.33) 0.001 3.10 (2.79–3.44) < 0.001 1.22 (1.09–1.36) < 0.001 disease Dementia 3.56 (3.41–3.72) < 0.001 1.78 (1.68–1.87) < 0.001 7.63 (7.01–8.31) < 0.001 2.18 (1.97–2.41) < 0.001 8.78 (8.05–9.58) < 0.001 2.22 (2.00–2.45) < 0.001 Malignant neo- 1.11 (1.07–1.15) < 0.001 1.10 (1.06–1.14) < 0.001 1.52 (1.40–1.65) < 0.001 1.13 (1.04–1.23) 0.006 1.58 (1.45–1.72) < 0.001 1.16 (1.06–1.27) 0.002 plasia Liver cirrhosis 1.06 (0.91–1.22) 0.460 1.09 (0.78–1.52) 0.616 0.96 (0.66–1.40) 0.848 Osteoporosis 0.91 (0.87–0.95) < 0.001 0.89 (0.84–0.93) < 0.001 0.97 (0.87–1.07) 0.511 0.99 (0.88–1.10) 0.814 Past femur 0.95 (0.92–0.99) 0.007 1.23 (1.11–1.36) < 0.001 0.97 (0.89–1.05) 0.439 4.32 (3.62–5.16) < 0.001 1.24 (1.03–1.48) 0.022 fracture Dyslipidemia 0.89 (0.86–0.92) < 0.001 0.92 (0.88–0.95) < 0.001 1.11 (1.03–1.21) 0.008 1.11 (1.02–1.20) 0.018 0.91 (0.84–0.99) 0.038 Ischemic heart 1.33 (1.25–1.42) < 0.001 1.09 (1.02–1.16) 0.012 2.10 (1.87–2.36) < 0.001 2.18 (1.93–2.47) < 0.001 disease Peripheral arte- 1.37 (1.25–1.49) < 0.001 2.68 (2.30–3.12) < 0.001 2.81 (2.40–3.29) < 0.001 riopathy Use of PPI 1.41 (1.37–1.46) < 0.001 1.17 (1.12–1.21) < 0.001 2.40 (2.21–2.61) < 0.001 1.14 (1.04–1.24) 0.005 2.53 (2.31–2.77) < 0.001 1.14 (1.03–1.25) 0.008 Use oral corti- 1.63 (1.55–1.70) < 0.001 1.39 (1.33–1.46) < 0.001 2.61 (2.38–2.86) < 0.001 1.79 (1.63–1.98) < 0.001 2.62 (2.38–2.89) < 0.001 1.79 (1.61–1.98) < 0.001 costeroids Use of DPP4- 1.15 (1.04–1.27) 0.008 1.89 (1.56–2.29) < 0.001 1.84 (1.50–2.26) < 0.001 inhibitors Use of statins 0.84 (0.81–0.87) < 0.001 0.87 (0.83–0.91) < 0.001 1.04 (0.96–1.14) 0.340 1.02 (0.93–1.12) 0.683 Journal of Endocrinological Investigation Journal of Endocrinological Investigation infection, nor with COVID-19 severity or mortality eGFR estimated glomerular filtration rate (ratios are calculated for every 10 ml increase of creatinine clearance), HR hazard ratio, CI 95% confidence interval 95%, COPD chronic obstructive < 0.001 0.007 (Table 3). Multivariated HR p Since renal proximal tubular reabsortion of calcifediol may be impaired in CKD, and low serum calcifediol lev- < 0.001 1.54 (1.24–1.91) < 0.001 0.88 (0.80–0.97) els are a frequent finding in this disease [33], we also ana- lyzed the association between calcifediol supplementation (CI 95%) and COVID-19 outcomes in the subgroup of patients with advanced CKD (stages 4 and 5). In these patients, calcifediol supplementation was associated with a significant reduc- < 0.001 tion in SARS-CoV2 infection (311/4380 [7.1%] vs 568/5533 [10.3%] in untreated controls; HR 0.77 [0.67–0.88]; COVID-19 mortality p Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admission or death p < 0.001 in multivariate Cox proportional hazards regres- < 0.001 1.50 (1.22–1.85) 1.29 (1.16–1.43) 0.003 1.21 (1.10–1.33) pulmonary disease, PPI proton pump inhibitors, DPP4 dipeptidyl peptidase-4, ACE angiotensin-converting enzyme, ARB angiotensin-II receptor blockers Univariate3 HR sion analysis) and a reduced risk of severe COVID-19 (144/4380; [3.3%] vs 260/5533 [4.7%] in untreated controls; (CI 95%) HR 0.78 [0.63–0.96]; p < 0.001). A reduction in COVID-19 mortality was also observed in patients with advanced CKD supplemented with calcifediol, but without reaching statis- tical significance (138/4380; [3.2%] vs 245/5533; [4.4%]; HR 0.83 [0.67–1.02]; p = 0.077). These associations were Multivariate HR p not observed in patients with better renal function (Table 5). < 0.001 1.49 (1.22–1.82) < 0.001 0.87 (0.80–0.95) 4 Serum 25‑hydroxyvitamin D levels and COVID‑19 (CI 95%) risk. Serum 25OHD levels were determined in 85,158 patients < 0.001 in the whole cohort (71,972 patients on cholecalciferol or calcifediol supplementation and 13,186 patients of the p Severe COVID-192 < 0.001 1.27 (1.15–1.40) 1.61 (1.33–1.95) < 0.001 1.18 (1.08–1.29) untreated control groups) between 1 November 2019 and Univariate HR 28 February 2020. 3 Mean serum 25OHD levels were significantly lower in (CI 95%) patients that developed SARS-CoV2 infection (22.7 [SD 14.1] ng/ml, n = 3091) than in non-infected patients (24.0 [14.1] ng/ml, p < 0.001). Mean serum 25OHD levels were also significantly lower in patients that developed severe COVID-19 (22.0 [SD 15.7] ng/ml [n = 538] versus 24.0 Multivariate HR p [14.3] ng/ml, p = 0.004) and in patients that died due to < 0.001 0.88 (0.84–0.93) < 0.001 0.85 (0.81–0.89) COVID (21.9 [15.7] ng/ml [n = 475] versus 24.0 [14.3] ng/ 4 Positive PCR or clinical diagnosis of SARS-CoV2 infection ml, p = 0.004). (CI 95%) In the multivariate analysis, lower serum 25OHD levels Cox regression analysis controlling for all covariates were also associated with increased risk of SARS-CoV2 0.134 infection (HR 0.97 [CI 95% 0.94–0.99]; p = 0.017), but were Ratios are calculated for every 10 years of age not significantly associated with severe COVID-19 or with SARS-CoV2 ­infection1 COVID-19 mortality. p Unadjusted Cox regression analysis 1.08 (0.98–1.19) 0.92 (0.88–0.96) 0.91 (0.88–0.95) Univariate HR COVID‑19 outcomes 3 in cholecalciferol‑supplemented, vitamin (CI 95%) D‑sufficient patients, versus unsupplemented, vitamin D‑deficient patients. Table 2  (continued) Use immunosup- The rate of SARS-CoV2 infection was significantly lower in Use of ARB Use of ACE inhibitors pressants vitamin D-sufficient patients supplemented with cholecalcif- erol (309/9474 [3.3%]) than in patients vitamin D deficient not receiving vitamin D supplements (430/7616 [5.6%]; 1 2 3 4 5 13 13 Table 3  Association between daily cholecalciferol (n = 165,588) or calcifediol dose (n = 132,590) and COVID-19 outcomes Outcome Cholecalciferol dose, Univariate ­analysis3 Multivariate ­analysis4 Calcifediol dose, daily Univariate ­analysis3 Multivariate daily mean (SD) μg mean (SD) μg ­analysis4 HR (CI 95%)5 p HR (CI 95%)5 p HR (CI 95%)5 p HR (CI p 95%)5 SARS-CoV2 ­infection1, n (%) Yes 15.4 (11.9) 1.00 (0.98–1.03) 0.729 0.97 (0.95–0.99) 0.005 19.8 (10.0) 1.06 (1.03–1.09) < 0.001 – NS No 15.3 (12.3) 19.3 (8.5) Severe COVID-192, n (%) Yes 15.1 (12.3) 0.99 (0.94–1.04) 0.986 0.92 (0.87–0.97) 0.002 20.1 (9.1) 1.06 (1.02–1.09) 0.002 – NS No 15.3 (12.3) 19.3 (8.6) COVID-19 mortality, n (%) Yes 15.0 (11.8) 0.98 (0.93–1.03) 0.442 0.90 (0.85–0.96) 0.001 20.2 (9.4) 1.06 (1.02–1.11) 0.002 – NS No 15.4 (12.3) 19.3 (8.6) 1 Positive PCR or clinical diagnosis of SARS-CoV2 infection 2 Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admission or death 3 Unadjusted Cox regression analysis 4 Cox regression analysis controlling for all covariates 5 Ratios are calculated for every 10 µg increase in the mean daily cholecalciferol or calcifediol dose Journal of Endocrinological Investigation Journal of Endocrinological Investigation Table 4  SARS-CoV2 infection, severe infection and death in patients treated with calcifediol (n = 134,703) and their matched controls (269,406) Outcome Calcifediol treatment Matched controls Univariate ­analysis3 Multivariate ­analysis4 (n = 134,703) (n = 269,406) HR (CI 95%) p HR (CI 95%) p SARS-CoV2 ­infection1, n (%) 5662 (4.2) 11,401 (4.2) 0.99 (0.96–1.03) 0.646 – – Severe COVID-192, n (%) 1037 (0.8) 2073 (0.8) 1.00 (0.93–1.08) 0.995 – – COVID-19 mortality, n (%) 934 (0.7) 1859 (0.7) 1.01 (0.93–1.09) 0.908 – – 1 Positive PCR or clinical diagnosis of SARS-CoV2 infection 2 Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admis- sion or death 3 Unadjusted Cox regression analysis 4 Cox regression analysis controlling for all covariates Table 5  SARS-CoV2 infection, severe infection and death in patients treated with calcifediol and controls, according to CKD stages Patients on stages 1–3 CKD Calcifediol Controls (n = 263,873) Univariate ­analysis3 Multivariate ­analysis4 (n = 130,323) HR (CI 95%) p HR (CI 95%) p 1 SARS-CoV2 ­infection , n (%) 5351 (4.1) 10,833 (4.1) 1.00 (0.97–1.03) 0.96 – NS Severe COVID-192, n (%) 893 (0.7) 1813 (0.7) 1.00 (0.92–1.08) 0.94 – NS COVID-19 mortality, n (%) 796 (0.6) 1614 (0.6) 1.00 (0.92–1.09) 0.97 – NS Patients on stages 4–5 CKD Calcifediol Controls (n = 5533) HR (CI 95%) p HR (CI 95%) p (n = 4380) SARS-CoV2 ­infection1, n (%) 311 (7.1) 568 (10.3) 0.68 (0.59–0.78) < 0.001 0.77 (0.67–0.88) < 0.001 Severe COVID-192, n (%) 144 (3.3) 260 (4.7) 0.69 (0.57–0.85) < 0.001 0.78 (0.63–0.96) 0.018 COVID-19 mortality, n (%) 138 (3.2) 245 (4.4) 0.71 (0.57–0.87) 0.001 0.83 (0.67–1.02) 0.077 CKD chronic kidney disease 1 Positive PCR or clinical diagnosis of SARS-CoV2 infection 2 Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admis- sion or death 3 Unadjusted Cox regression analysis 4 Cox regression analysis controlling for all covariates HR 0.66 [CI 95% 0.57–0.77], p < 0.001 in multivariate Cox [CI 95% 0.61–0.79], p < 0.001 in multivariate Cox propor- proportional hazards regression analysis). Similarly, vitamin tional hazards regression analysis). Similarly, vitamin D-suf- D-sufficient patients on cholecalciferol supplementation had ficient patients on calcifediol supplementation had lower risk lower risk of severe COVID-19 (65/9474 [0.7%] vs 99/7616 of severe COVID-19 (n = 100/16276 [0.6%] vs n = 99/7616 [1.3%]; HR 0.72 [0.52–1.00]; p = 0.050), and lower COVID- [1.3%]; HR 0.61 [0.46–0.81]; p = 0.001), and lower COVID- 19 mortality (56/9474 [0.6%] vs 96/7616 [1.3%]; HR 0.66 19 mortality (n = 88/16276 [0.5%] vs n = 96/7616 [1.3%]; [CI 95% 0.46–0.93], p = 0.018) compared to vitamin D-defi- HR 0.56 [0.42–0.76]; p < 0.001) compared to vitamin cient unsupplemented patients (Table 6). D-deficient unsupplemented patients (Table 7). COVID‑19 outcomes in calcifediol‑supplemented, vitamin D‑sufficient patients, Discussion versus unsupplemented, vitamin D‑deficient patients. In this large population-based cohort, we have compared COVID-19 outcomes in patients supplemented with chole- The rate of SARS-CoV2 infection was significantly lower in calciferol or calcifediol versus untreated matched controls, vitamin D-sufficient patients supplemented with calcifediol finding only a mild reduction in the risk of SARS-CoV2 (535/16276 [3.3%]) than in patients vitamin D deficient not infection (diagnosed clinically or by PCR) in patients sup- receiving vitamin D supplements (430/7616 [5.6%]; HR 0.69 plemented with cholecalciferol, and a small reduction in 13 Journal of Endocrinological Investigation Table 6  COVID-19 outcomes in 25OHD-sufficient, cholecalciferol-supplemented patients compared to unsupplemented, 25OHD-deficient patients Cholecalciferol treated and Untreated controls with Univariate ­analysis3 Multivariate ­analysis4 serum 25OHD ≥ 30 ng/ml serum 25OHD < 20 ng/ml (n = 9474) (n = 7616) HR (CI 95%) p HR (CI 95%) p SARS-CoV2 ­infection1, 309 (3.3%) 430 (5.6%) 0.57 (0.50–0.66) < 0.001 0.66 (0.57–0.77) < 0.001 n (%) Severe COVID-192, n (%) 65 (0.7%) 99 (1.3%) 0.53 (0.39–0.72) < 0.001 0.72 (0.52–1.00) 0.050 COVID-19 mortality, n (%) 56 (0.6%) 96 (1.3%) 0.47 (0.34–0.65) < 0.001 0.66 (0.46–0.93) 0.018 25OHD 25-hydroxyvitamin D 1 Positive PCR or clinical diagnosis of SARS-CoV2 infection 2 Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admis- sion or death 3 Unadjusted Cox regression analysis 4 Cox regression analysis controlling for all covariates Table 7  COVID-19 outcomes in 25OHD-suficient, calcifediol-supplemented patients compared to unsupplemented, 25OHD-suficient patients Calcifediol treated and Untreated controls with Univariate ­analysis3 Multivariate ­analysis4 serum 25OHD ≥ 30 ng/ml serum 25OHD < 20 ng/ml (n = 16,276) (n = 7616) HR (CI 95%) p HR (CI 95%) p SARS-CoV2 ­infection1, 535 (3.3%) 430 (5.6%) 0.58 (0.51–0.66) < 0.001 0.69 (0.61–0.79) < 0.001 n (%) Severe COVID-192, n (%) 100 (0.6%) 99 (1.3%) 0.47 (0.36–0.62) < 0.001 0.61 (0.46–0.81) 0.001 COVID-19 mortality, n (%) 88 (0.5%) 96 (1.3%) 0.43 (0.32–0.57) < 0.001 0.56 (0.42–0.76) < 0.001 25OHD 25-hydroxyvitamin D 1 Positive PCR or clinical diagnosis of SARS-CoV2 infection 2 Composite outcome of need for non-invasive mechanical ventilation, orotracheal intubation, mechanical ventilation, intensive care unit admis- sion or death 3 Unadjusted Cox regression analysis 4 Cox regression analysis controlling for all covariates the rates of infection and COVID-19 mortality associated respiratory distress syndrome in experimental models of with the use of higher cholecalciferol doses. Since vitamin lipopolysaccharide-induced acute lung injury [35–37]. D supplementation may be prescribed as a result of a low From the methodological point of view, our study differs vitamin D status, we also compared COVID-19 outcomes from previously published studies, where the effects of vita- between supplemented, vitamin D-sufficient patients with min D supplementation were mostly analyzed in hospitalized unsupplemented, vitamin D-deficient patients showing a COVID-19 patients [20–25]. To the best of our knowledge, significant reduction in the risk of infection, hospitaliza- this is the largest study to analyze the association of chole- tion and COVID-19 mortality in patients with a normal calciferol or calcifediol supplementation with COVID-19 vitamin D status supplemented either with cholecalciferol outcomes at the population level. This has allowed to detect or calcifediol. Overall, our results suggest that reaching a small differences in outcomes after adjusting for multiple sufficient vitamin D status in patients supplemented with covariates. Our study also differs from that of Loucera et al., these vitamin D metabolites is associated with a reduced a study that compares COVID-19 inhospital mortality in risk of SARS-CoV2 infection and lower COVID-19 patients under vitamin D supplementation with unsupple- mortality. mented, propensity score-matched, hospitalized COVID-19 There are several pathophysiological mechanisms that patients, since it only analyzed hospitalized patients and not could explain the benefits of vitamin D against COVID- all the population at risk [38]. 19. Calcitriol, the hormonal form of vitamin D, can protect Our study has some similarities with that of Meltzer et al. against infections by increasing the production of LL-37, [11], who also combined the results of serum 25OHD levels β-defensin2 and nitric oxide in respiratory epithelia [34] and the vitamin D supplied to categorize 489 patients from and it has been shown to reduce the incidence of adult an urban academic center as likely vitamin D sufficient or 13 Journal of Endocrinological Investigation likely deficient; however, in that study, the authors only ana- analysis. IS methodological support in the matching process. DD data lyzed the risk of SARS-CoV2 infection, but not the risk of curation. EC design, writing review. AT methodological support in the matching process. AB conceptualization and design. MG design and severe COVID-19 or COVID-19 mortality. methodological support, writing review. We also found a lower risk of SARS-CoV2 infection and reduced COVID-19 mortality in patients in stages 4–5 CKD Funding This research received no external funding. The Catalan supplemented with calcifediol. These results are similar, but Agency for Health Quality and Evaluation (AQUAS) offered a data of lower magnitude, than those observed in patients treated analyst for extracting all the data necessary for the development of this project. with calcitriol [26]. The better results obtained with calci- triol may be a consequence of its greater potency, being the Data availability statement All data can be offered to researchers upon active metabolite of vitamin D, not requiring any hydrox- request. ylation process in cytochromes and having a high affin- ity for the vitamin D receptor [39]. However, since renal Declarations proximal tubular cells express both ACE2 and mitochondrial 1-hydroxylase [40, 41], and they are commonly infected and Conflicts of interest The authors declare no conflict of interest. The damaged in severe COVID-19 [42, 43], it is also possible Catalan Agency for Health Quality and Evaluation (AQUAS), a gov- ernmental office, offered a data analyst for extracting all the data nec- that during SARS-CoV2 infection, an acute decrease in essary for the development of this project. They had no role in the renal calcitriol synthesis takes place that cannot be restored design of the study, analyses, interpretation of data; in the writing of through cholecalciferol or calcifediol supplementation. the manuscript, or in the decision to publish these results. We think our study has some strengths, including the Institutional review board statement This study was approved by assessment of COVID-19 outcomes in a large population the Institutional Ethics Committee of the Corporació Sanitària Parc under supplementation with cholecalciferol or calcifediol Taulí—Universitat Autònoma Barcelona (Project number 2020588). and the use of a matched cohort of controls. This study also has some limitations. First, there are the inherent limitations Informed consent Not applicable. of an observational cohort. Although we were comprehen- Patient and public involvement It was not appropriate or possible to sive in analyzing many covariables, it is possible that there involve patients or the public in the design, or conduct, or reporting, are still important variables not considered in the matching or dissemination plans of our research. process that may disbalance the treated and control groups. Transparency statement The lead author (J. Oristrell) affirms that the Second, our data were obtained from the registries of the manuscript is an honest, accurate, and transparent account of the study health administration of the government of Catalonia, which being reported. No important aspects of the study have been omitted, are fed by the diagnoses issued by family physicians, hos- and there were no significant discrepancies from the study as it was pital discharge reports, laboratory data of public hospitals, planned. or medicines supplied by pharmacies, with the inherent Dissemination declaration Dissemination to the patients supplemented limitations of administrative data. Finally, we decided to with vitamin D compounts is not applicable. focus our analysis on the first wave of the pandemic, with higher number of severe cases and mortality. However, the diagnosis of SARS-CoV2 in that phase could not be ascer- tained with PCR in all the cases, and some patients received References a clinical diagnosis without a confirmatory microbiological confirmation. 1. Grant WB, Lahore H, McDonnell SL, Baggerly CA, French CB, In conclusion, in this large, population-based study, we Aliano JL et al (2020) Evidence that vitamin D supplementation observed that cholecalciferol or calcifediol supplementa- could reduce risk of influenza and COVID-19 infections and tion seem to be beneficial against SARS-CoV2 infection, deaths. Nutrients 12:988. https://​doi.​org/​10.​3390/​nu120​40988 2. Teymoori-Rad M, Marashi SM (2020) Vitamin D and COVID- COVID-19 severity and COVID-19 mortality in patients 19: from potential therapeutic effects to unanswered qüestions. achieving serum 25OHD levels ≥ 30 ng/ml. Rev Med Virol. https://​doi.​org/​10.​1002/​r mv.​2159 3. Hilger J, Friedel A, Herr R, Rausch T, Roos F, Wahl DA et al Supplementary Information The online version contains supplemen- (2014) A systematic review of vitamin D status in populations tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 40618-0​ 21-0​ 1639-9. worldwide. Br J Nutr 111:23–45. https://​doi.​org/​10.​1017/​S0007​ 11451​30018​40 Acknowledgements Agència de Qualitat i Avaluació Sanitàries 4. Peelen E, Knippenberg S, Muris AH, Thewissen M, Smolders (AQUAS). Programa d’Analítica de Dades per a la Recerca i la Inno- J, Cohen Tervaert JW et al (2011) Effects of vitamin D on the vació en Salut (PADRIS). Generalitat de Catalunya. peripheral adaptive immune system: a review. Autoimmun Rev 10:733–743 5. Kongsbak M, Levring TB, Geisler C, von Essen MR (2013) The Author contributions JO conceptualization, data analysis, writing vitamin D receptor an T cell function. Front Immunol 4:1–10 original draft. JCO design and methodological support. Formal data 13 Journal of Endocrinological Investigation 6. Bscheider M, Butcher EC (2016) Vitamin D immunoregulation 22. Cereda E, Bogliolo L, Lobascio F et al (2019) Vitamin D supple- through dendritic cells. Immunology 148:227–236. https://​doi.​ mentation and outcomes in coronavirus disease 2019 (COVID- org/​10.​1111/​imm.​12610 19) patients from the outbreak area of Lombardy, Italy. Nutrition 7. Bergman P, Lindh AU, Björkhem-Bergman L et al (2013) 2020:111055. https://​doi.​org/​10.​1016/j.​nut.​2020.​111055 Vitamin D and respiratory tract infections: a systematic review 23. Entrenas Castillo M, Entrenas Costa LM, Vaquero Barrios JM and meta-analysis of randomized controlled trials. PLoS ONE et al (2020) Effect of calcifediol treatment and best available 8:e65835 tehrapy on intensive care unit admission and mortality among 8. Martineau AR, Jolliffe DA, Hooper RL et al (2017) Vitamin D patients hospitalized for COVID-19: a pilot randomized clinical supplementation to prevent acute respiratory tract infections: sys- study. J Steroid Biochem Mol Biol 203:105751 tematic review and meta-analysis of individual participant data. 24. Rastogi A, Bhansali A, Khare N et al (2020) Short term, high- BMJ 356:i6583 dose vitamin D supplementation for COVID-19 disease: a ran- 9. Pérez Castrillón JL, Casado E, Corral Gudino L et al (2020) domised, placebo-controlled, study (SHADE study). Postgrad COVID-19 and vitamin D. Position paper of the Spanish Soci- Med J. https://​doi.​org/​10.​1136/​postg​radme​dj-​2020-​139065 ety for Bone Research and Mineral Metabolism (SEIOMM). Rev 25. Murai IH, Fernandes AL, Sales LP et al (2021) Effect of a single Osteoporos Metab Miner 12:155–159 high dose of vitamin D3 on hospital length of stay in patients 10. Kaufman HW, Niles JK, Kroll MH, Bi C, Holick MF (2020) with moderate to severe COVID-19. A randomized clinical trial. SARS-CoV-2 positivity rates associated with circulating JAMA. https://​doi.​org/​10.​1001/​jama.​2020.​26848 25-hydroxyvitamin D levels. PLoS ONE 15:e0239252. https://​ 26. Oristrell J, Oliva JC, Subirana I, Casado E, Domínguez D, Toloba doi.​org/​10.​1371/​journ​al.​pone.​02392​52 A et al (2021) Association of calcitriol supplementation with 11. Meltzer DO, Best TJ, Zhang H, Vokes T, Arora V, Solway J (2020) reduced COVID-19 mortality in patients with chronic kidney Association of vitamin D status and other clinical characteristics disease:a population-based study. Biomedicines 9:509. https://​ with COVID-19 test results. JAMA Netw Open 3(9):e2019722. doi.​org/​10.​3390/​biome​dicin​es905​0509 https://​doi.​org/​10.​1001/​jaman​etwor​kopen.​2020.​19722 27. Williamson EJ, Walker AJ, Bhaskaran K et al (2020) Factors asso- 12. Merzon E, Tworowski D, Gorohovski A, Vinker S, Golan Cohen ciated with COVID-19-related death using OpenSAFELY. Nature A et al (2020) Low plasma 25(OH) vitamin D level is associated 584:430–443 with increased risk of COVID-19 infection: an Israeli population- 28. Sekhon JS (2011) Multivariate and propensity score matching based study. FEBS J. https://​doi.​org/​10.​1111/​febs.​15495 software with automated balance Optimization: the matching 13. Radujkovic A, Hippchen T, Tiwari-Heckler S, Dreher S, Box- package for R. J Stat Soft 42:1–52 berger M, Merle U (2020) Vitamin D deficiency and outcome of 29. Levey AS, Stevens LA, Schmid CH et al (2009) CKD-EPI COVID-19 patients. Nutrients 12:2757. https://​doi.​org/​10.​3390/​ (Chronic Kidney Disease Epidemiology Collaboration). A new nu120​92757 equation to estimate glomerular filtration rate. Ann Intern Med 14. Abrishami A, Dalili N, Torbati PM, Asgari R, Arab-Ahmadi M, 150:604–612 Behnam B et al (2020) Possible association of vitamin D status 30. Yoshida K, Bohn J. Package ‘tableone’ for R. Accessible online with lung involvement and outcome in patients with COVID-19: at https://​cran.r-​proje​ct.​org/​web/​packa​ges/​table​one/​table​one.​pdf a retrospective study. Eur J Nutr 30:1–9. https://​doi.​org/​10.​1007/​ 31. Therneau T (2020) A package for survival analysis in R. R pack- s00394-​020-​02411-0 age version 3.2-7, https://​CRAN.R-​proje​ct.​org/​packa​ge=​survi​val 15. Luo X, Liao Q, Shen Y, Li H, Cheng L (2021) Vitamin D defi- 32. Kassambara A, Kosinski M, Biecek P et al. Survminer: drawing ciency is associated with COVID-19 incidence and disease sever- survival curves using ’ggplot2’, https://C ​ RAN.R-p​ rojec​ t.o​ rg/p​ acka​ ity in Chinese people. J Nutr 151:98–103 ge=​survm​iner 16. Bassatne A, Basbous M, Chakhtoura M, El Zein O, Rahme M, 33. Dusso AS, Tokumoto M (2011) Defective renal maintenance of El-Hajj FG (2021) The link between COVID-19 and vitamin D the vitamin D endocrine system impairs vitamin D renoprotection: (VIVID): a systematic review and meta-analysis. Metab Clin Exp a downward spiral in kidney disease. Kidney Int 79:715–729 119:154753 34. Greiller CL, Martineau AR (2015) Modulation of the immune 17. Kazemi A, Mohammadi V, Aghababaee SK, Golzarand M, Clark response to respiratory viruses by vitamin D. Nutrients 7:4240– CCT, Babajafari S (2021) Association of vitamin D status with 4270. https://​doi.​org/​10.​3390/​nu706​4240 SARS-CoV2 infection or COVID-19 severity: a systematic review 35. Kong J, Zhu X, Shi Y et al (2013) VDR attenuates acute lung and meta-analysis. Adv Nutr. https://​doi.​org/​10.​1093/​advan​ces/​ injury by blocking Ang-2-Tie-2 pathway and renin-angiotensin nmab0​12 system. Mol Endocrinol 27:2116–2125. https://​doi.​org/​10.​1210/​ 18. Pereira M, Dantas Damascena A, Galvão Azevedo LM, de me.​2013-​1146 Almeida OT, da Mota SJ (2020) Vitamin D deficiency aggravates 36. Dancer RCA, Parekh D, Lax S et al (2015) Vitamin D deficiency COVID-19: systematic review and meta-analysis. Crit Rev Food contributes directly to the acute respiratory distress syndrome Sci Nutr. https://​doi.​org/​10.​1080/​10408​398.​2020.​18410​90 (ARDS). Thorax 70:617–624. https://​d oi.​o rg/​1 0.​1 136/​t hora​ 19. Hernández JL, Nan D, Fernández-Ayala M, García-Unzueta M, xjnl-​2014-​206680 Hernández-Hernández MA, López-Hoyos M et al (2020) Vitamin 37. Xu J, Yang J, Chen J et al (2017) Vitamin D alleviates lipopoly- D status in hospitalized patients with SARS-CoV2 infection. J saccharide-induced acute lung injury via regulation of the renin- Clin Endocrinol Metab 20:1–11. https://​doi.​org/​10.​1210/​clinem/​ angiotensin system. Mol Med Rep 16:7432–7438. https://​doi.o​ rg/​ dgaa7​33 10.​3892/​mmr.​2017.​7546 20. Annweiler G, Corvaisier M, Gautier J et al (2020) Vitamin D 38. Loucera C, Peña-Chilet M, Esteban-Medina M, Muñoyerro-Muñiz supplementation associated to better survival in hospitalized frail D, Villegas R, López-Miranda J et al (2021) Real-world evidence elderly COVID-19 patients: the GERIA-COVID quasi-experimen- of calcifediol use and mortality rate of COVID-19 hospitalized in tal study. Nutrients 12:3377 a large cohort of 16,401 Andalusian patients. MedRxiv Preprint. 21. Ling SF, Broad E, Murphy R, Pappachan JM, Pardesi-Newton S, https://​doi.​org/​10.​1101/​2021.​04.​27.​21255​937 Kong MF et al (2020) High-dose cholecalciferol booster therapy 39. Haussler MR, Professor R, Jurutka PW et al (2011) Vitamin D is associated with a reduced risk of mortality in patients with receptor (VDR)-mediated actions of 1α,25(OH)2vitamin ­D3: COVID-19: a cross-sectional multi-centre observational study. genomic and non-genomic mechanisms. Best Pract Res Clin Nutrients 12:3799. https://​doi.​org/​10.​3390/​nu121​23799 Endocrinol Metab 25:543–559 13 Journal of Endocrinological Investigation 40. Brunette MG, Chan M, Ferriere C, Roberts KD (1978) Site Publisher’s Note Springer Nature remains neutral with regard to of 1,25(OH)2 vitamin D3 synthesis in the kidney. Nature jurisdictional claims in published maps and institutional affiliations. 276:287–289 41. Lely AT, Hamming I, van Goor H, Navis GJ (2004) Renal ACE2 expression in human kidney disease. J Pathol 204:587–593 42. Werion A, Belkhir L, Perrot M, Schmit G, Aydin S, Chen Z et al (2020) SARS-CoV-2 causes a specific dysfunction of the kidney proximal tubule. Kidney Int 98:1296–1307 43. Armaly Z, Kinaneh S, Skorecki K (2021) Renal manifestations of COVID-19: physiology and pathophysiology. J Clin Med 10:1216. https://​doi.​org/​10.​3390/​jcm10​061216 13