Influence of Vitamin D Supplementation by Simulated Sunlight or Oral D3 on Respiratory Infection During Military Training

Sophie E. Harrison; Samuel J. Oliver; Daniel S. Kashi; Alexander T. Carswell; Jason P. Edwards; Laurel M. Wentz; Ross Roberts; Jonathan C. Y. Tang; Rachel M. Izard; Sarah Jackson; Donald Allan; Lesley E. Rhodes; William D. Fraser; Julie P. Greeves; Neil P. Walsh


Med Sci Sports Exerc. 2021;53(7):1505-1516. 

In This Article


British Army recruits voluntarily participated in study 1 and study 2 after providing fully informed written consent and passing a clinician-screened medical assessment, which excludes a number of medical conditions, including chronic lung diseases, and asthma symptoms or treatment in the last year. Men (studies 1 and 2) were located at Infantry Training Centre Catterick, UK (latitude 54°N), and women (study 1) were located at Army Training Centre Pirbright, UK (latitude 51°N). All volunteers were studied during 12 wk of basic military training that follows a syllabus of basic military skills, including physical training, weapon handling, map reading, and fieldcraft. The progressive, structured, physical training program included endurance training, circuit training, agility-based gymnasium work, assault course practice, and marching with a load. The studies received ethical approval from the UK Ministry of Defense Research Ethics Committee and were conducted in accordance with the Declaration of Helsinki (2013) (study registration references at [NCT02416895, NCT03132103]).

Study 1

Participants and Study Design. A total of 1644 men and women (n = 1220 men: 95% White ethnicity, age = 21 ± 3 yr; body mass = 75.3 ± 9.9 kg, height = 1.77 ± 0.06 m, body mass index [BMI] = 24.0 ± 2.7 kg·m−2, 38% smokers; n = 424 women: 95% White ethnicity, age = 22 ± 3 yr, body mass = 64.8 ± 8.2 kg, height = 1.65 ± 0.06 m, BMI = 23.7 ± 2.4 kg·m−2, 24% smokers) participated in this prospective cohort study between January 2014 and September 2015. Participants were included if they gave baseline blood samples, and URTI data were available during the entire 12 wk of military training.

Experimental Procedures. Baseline measures were collected from each participant during the initial medical assessment, including a venous blood sample for the determination of serum 25(OH)D, height, body mass, ethnicity, and smoking history by self-reported questionnaire (Figure 1). Medical records were accessed to obtain physician-diagnosed URTI and lost training days due to URTI. URTI was diagnosed by a single general practice-trained physician. A lost training day was recorded when a recruit was unavailable for normal military training.

Figure 1.

A schematic of the prospective cohort study (study 1) that investigated the association between vitamin D status (serum 25(OH)D), URTI and days lost from training, and the randomized controlled trial (study 2) that investigated the effects of vitamin D supplementation by solar simulated radiation (SSR), oral vitamin D3 (ORAL), or placebo (SSR-P or ORAL-P) on URTI and mucosal immunity. Blood samples were collected at baseline (studies 1 and 2), week 5, and the end of week 12 (study 2). Saliva samples were collected at baseline, week 5, and the end of week 12 (study 2). The syringe icon represents the blood sample; the head and the tube icon represent the saliva sample.

Study 2

Participants and Study Design. A total of 250 men (age = 22 ± 7 yr, body mass = 76.3 ± 10.8 kg, height = 1.77 ± 0.06 m, BMI = 24.2 ± 3.0 kg·m−2 participated in this double-blind, randomized placebo-controlled trial (Figure 1). Participants were recruited at the start of 12 wk of basic military training during January and February of 2016 and 2017, when ambient UVB is negligible at UK latitudes (50°N–60°N) and serum 25(OH)D is at its annual nadir. Participants were eligible to participate if they had sun-reactive skin type of I to IV on the Fitzpatrick Skin Type Scale,[26] were not consuming supplements containing vitamin D, and had not used a sunbed or traveled to a sunny climate in the 3 months before the study.

Experimental Procedures. Participants were randomized within their platoons to one of four intervention groups: 1) oral vitamin D3 supplementation (ORAL), 2) oral placebo supplementation (ORAL-P), 3) solar simulated radiation (SSR), or 4) solar simulated radiation placebo (SSR-P). Block randomization was used ( to achieve an equal distribution of intervention groups within each platoon so any differences in training conditions between platoons did not influence the outcomes of the study. The intervention strategy for the SSR and ORAL groups was to restore and then maintain IOM- and EFSA-recommended vitamin D sufficiency (serum 25(OH)D ≥50 nmol·L−1. Participants completed a 4-wk restoration phase, necessary because serum 25(OH)D was at its annual wintertime nadir, followed by an 8-wk maintenance phase.

At baseline, during the routine initial medical assessment, height and body mass were measured, a venous blood sample was collected for the determination of serum 25(OH)D, and a lifestyle questionnaire was completed to determine smoking and alcohol use. Additional blood samples were obtained at week 5 and week 12. At baseline, week 5, and week, 12 saliva samples were collected in the evening, between 1800 and 2130 h, at least 15 min postprandial. Participants were excluded from analysis if they did not achieve ≥80% compliance with the intervention. Compliance with the interventions was calculated from researcher weekly counts of oral capsules remaining in recruit pill boxes and SSR cabinet visit records. Vitamin D from the diet was estimated in week 12 using a food frequency questionnaire, and solar UVR exposure was measured in weeks 4 and 11 using polysulfone badges, worn on the upper chest/anterior shoulder region on the outer clothes, as described.[10,27] The change in absorbance of the badges due to exposure was measured using a spectrophotometer and related to the erythemal effective UVR (sunburning) through a standard polynomial relationship; data are expressed as standard erythemal dose (SED) per day.[27] Participant dietary vitamin D intake was calculated excluding the oral D3 supplement participants received in the ORAL group. On completion of the study, to confirm participant blinding, participants were asked to guess the intervention they had received.

Simulated Sunlight Intervention. Simulated sunlight was provided following guidelines on safe, low-level sunlight exposure for vitamin D synthesis,[6] described previously to achieve serum 25(OH)D ≥50 nmol·L−1 in the majority of individuals with sun-reactive skin type of I to IV.[28] Those assigned to the SSR intervention were exposed three times a week during the restoration phase and once per week during the maintenance phase to an experimenter-controlled constant UVR dose using a whole body irradiation cabinet (Hapro Jade, Kapelle, The Netherlands) fitted with Arimed B fluorescent tubes (Cosmedico, Stuttgart, Germany). The fluorescent tubes emitted a UVR spectrum similar to sunlight (λ = 290–400 nm, 95% UVA = 320–400 nm, 5% UVB = 290–320 nm) that was characterized by a spectroradiometer (USB2000+; Ocean Optics BV, Duiven, The Netherlands) radiometrically calibrated with traceability to UK national standards.

During each exposure, participants received a 1.3× SED while wearing shorts and a T-shirt to expose ~40% skin surface area. This dose is equivalent to ~15 min of midday summer sun exposure six times per week for a casually dressed individual in northern England (latitude 53.5°N).[28] A constant SSR dose was maintained during the study by monitoring irradiance using a spectroradiometer (USB2000+, Ocean Optics BV) and adjusting for any decrease in measured irradiance emitted by increasing exposure time, as described[28] (mean duration of SSR exposures was 222 ± 23 s). The exposure time was controlled by using an electronic timer on the irradiation cabinet. For the SSR-P participants, the number and the duration of intervention exposures were the same as SSR, except the irradiation cabinet fluorescent tubes were covered with transparent UVR blocking film (DermaGard UV film; SunGard, Woburn, MA). A spectroradiometer confirmed that the UVR blocking film was effective at preventing transmission of 99.9% of UVR.

Oral Vitamin D3. Participants receiving the ORAL intervention consumed a vitamin D3 capsule daily, containing 1000 and 400 IU during the restoration and maintenance phases, respectively (Pure Encapsulations, Sudbury, MA). The restoration dose was based on previous predictive modeling to achieve serum 25(OH)D ≥50 nmol·L−1[29] and pilot investigations that showed it achieved similar serum 25(OH)D concentrations to SSR, and it was less than the tolerable upper intake recommended by the IOM and EFSA.[7,8] The ORAL maintenance dose was shown in a pilot investigation to maintain serum 25(OH)D ≥50 nmol·L−1 and when accounting for typical habitual dietary intake[5–7] was similar to IOM- and EFSA-recommended dietary allowances.[7,8] For 12 wk, ORAL-P participants consumed an identical-looking cellulose placebo capsule daily (Almac Group, County Armagh, UK). Independent analysis found the vitamin D3 content of the 1000- and 400-IU capsules to be 1090 and 460 IU, respectively, and confirmed the placebo did not contain vitamin D (NSF International Laboratories, Ann Arbor, MI).

URTI Diagnosis (Study 2). As in study 1, medical records were accessed to obtain data on physician-diagnosed URTI and lost training days due to URTI. However, URTI was principally monitored by self-reported daily symptoms recorded using the Jackson common cold questionnaire.[30] A strength of the Jackson common cold questionnaire compared with physician-diagnosed URTI is that URTI duration and severity, as well as prevalence, can be assessed. Participants were asked to rate eight symptoms (sneezing, headache, feeling generally unwell, runny nose, blocked nose, sore throat, cough, and chilliness) on a 4-point Likert scale (not at all = 0, mild = 1, moderate = 2, severe = 3). Data were included when participants completed ≥80% of their daily Jackson questionnaires. A URTI was defined by a daily total symptom score of ≥6 for two or more consecutive days.[31] Further, the average URTI duration (average duration of all URTI episodes), the peak URTI symptom severity (maximum URTI severity score on a single day of any URTI episode; maximum possible peak severity is 24 arbitrary units [AU]), and the total number of days with a URTI during basic military training for each participant (total days with URTI; military training is 84 d in total) were also determined. Self-reported URTI data were not reported back to the military and therefore did not influence physician diagnosis of URTI or lost training days due to URTI.

Blood Analysis (Studies 1 and 2). Whole blood samples were collected by venipuncture from an antecubital vein into plain vacutainer tubes (Becton Dickinson, Oxford, UK) and left to clot for 1 h. Subsequently, samples were centrifuged at 1500g for 10 min at 4°C, and the serum was aliquoted into universal tubes before being immediately frozen at −80°C for later analysis. Total serum 25(OH)D was measured with high-pressure liquid chromatography–mass spectrometry. Analyses were performed in a Vitamin D External Quality Assurance Scheme–certified laboratory (Bioanalytical Facility, University of East Anglia, Norwich, UK). The mean intra-assay coefficient of variation for 25(OH)D3 and 25(OH)D2 was <10%, and the lower limit of quantification was 0.1 nmol·L−1.[32]

Saliva Collection and Analysis (Study 2). Saliva was collected for 5 min in a preweighed 30-mL tube using the passive dribble method.[33] Samples were weighed immediately after collection, centrifuged at 1500g and 4°C for 10 min, aliquoted, and then stored at −80°C. Samples were analyzed in duplicate by enzyme-linked immunosorbent assay for secretory immunoglobulin A (SIgA) and cathelicidin concentration (Salimetrics, State Colelge, PA, and Hycult Biotech, Wayne, PA). The mean intra-assay coefficient of variation was 2.3% for saliva SIgA concentrations ranging from 0.02 to 0.51 mg·mL−1 and 10.2% for saliva cathelicidin concentrations ranging from 0.30 to 65.90 μg·L−1. Assuming the density to be 1.00 g·mL−1 for saliva, the secretion rate was calculated by multiplying the saliva flow rate by concentration.[33]

Statistical Analysis. Statistical analyses were performed using SPSS Version 25 (IBM Corp, New York, NY). Data points that were more than three times the interquartile range were deemed as outliers and removed. Where data were not normally distributed, they were transformed using square-root calculation. Significance was set at P < 0.05. For study 1, an estimated minimum required sample size of 1286 was calculated, using a type 1 error (one-tailed) of 5%, a power of 80%, and an anticipated odds ratio of 1.5 (equivalent to a small effect size) and including a binomial variable at 20%. This was based on previous literature describing the difference in URTI prevalence between individuals with low and high vitamin D status, whereby 20% of individuals with high vitamin D status reported a URTI,[4] while also anticipating that 20% of individuals would have low vitamin D status across the whole year.[34] Logistic regression was used to compare vitamin D status (25(OH)D ≥50 vs <50 nmol·L−1 and ≥75 vs <30, ≥50–<75, and <75 nmol·L−1 with URTI prevalence during the 12-wk military training and the first 3 wk of military training; circulating 25(OH)D has an estimated 3-wk half-life.[35,36] Sex and smoking were included as covariates as they have previously been shown to influence URTI susceptibility.[37,38] Chi-square tests were used to compare URTI prevalence between vitamin D–sufficient participants and those with serum 25(OH)D <50 nmol·L−1 and the proportion of vitamin D–sufficient participants between seasons. We used a one-way ANOVA to compare 25(OH)D between seasons. For study 2, an estimated minimum required sample size of 74 (37 in each comparison group) was calculated, using the anticipated odds ratio of 0.3 for URTI prevalence between vitamin D and placebo supplemented individuals with low vitamin D status,[15] and that 60% would self-report URTI during basic military training,[18,31,39] with a type 1 error (one-tailed) of 5% and a power of 80%. URTI prevalence between vitamin D (SSR and ORAL) and placebo (SSR-P and ORAL-P) supplementation groups was compared by logistic regression. Independent samples t-tests (two groups [SSR and ORAL combined, SSR-P and ORAL-P combined]) were used to compare vitamin D and placebo supplementation effects on average URTI duration, total days with URTI, peak URTI severity, saliva flow rate, SIgA, and cathelicidin. Serum 25(OH)D, total days with URTI, URTI duration, URTI severity, saliva flow rate, SIgA, and cathelicidin were compared between vitamin D strategies and placebo groups by mixed-model ANOVA (4 groups [SSR, ORAL, SSR-P, and ORAL-P] × 3 time points [baseline, week 5, and week 12]). Sunlight exposure and dietary vitamin D intake between SSR, ORAL, SSR-P, and ORAL-P groups were compared by one-way ANOVA. Cohen's d effect sizes (d) are presented to indicate the meaningfulness of group differences for total days with URTI, URTI duration, and URTI severity, whereby values greater than 0.2, 0.5, and 0.8 represent small, medium, and large effects, respectively.[40]