Effect of Daily Vitamin D Supplementation on Risk of Upper Respiratory Infection in Older Adults: A Randomized Controlled Trial

Abstract

Background

Among individuals with vitamin D deficiency, daily vitamin D supplementation appears to lower risk of acute respiratory infection. However, recent trials, in different populations and using different regimens, have yielded null results. We investigated the effect of daily vitamin D supplementation (vs placebo) on risk of upper respiratory infection (URI) in older adults.

Methods

The VITamin D and OmegA-3 TriaL (VITAL) is a randomized, double-blind, placebo-controlled trial of supplemental vitamin D and/or omega-3 fatty acids in generally healthy men (age ≥50 years) and women (age ≥55 years). This prespecified analysis focuses on vitamin D₃ (2000 IU/day) versus placebo in the 15 804 (61%) participants with baseline serum total 25-hydroxyvitamin D level. The primary outcome was self-report of a recent URI at 1-year follow-up.

Results

Participants had a mean age of 68 years and 51% were women; 76% were non-Hispanic White, 16% Black, and 8% other race/ethnicity. The mean 25-hydroxyvitamin D level at baseline was 31 (standard deviation, 10) ng/mL, with <12 ng/mL in 2.4%. The overall effect of vitamin D supplementation on recent URI was nonsignificant (odds ratio [OR], 0.96 [95% confidence interval {CI}, .86–1.06]). In the prespecified subgroup of primary interest (<12 ng/mL and denied taking concurrent vitamin D), which had only 255 participants, vitamin D supplementation was nonsignificant (OR, 0.60 [95% CI, .28–1.30]). Statistical power to assess effect modification in other subgroups was limited.

Conclusions

In older adults not selected for vitamin D deficiency, supplemental vitamin D did not lower URI risk overall. Whether effects differ in subgroups requires further study.

Clinical Trials Registration. NCT01169259.

Graphical Abstract

Graphical Abstract.

This graphical abstract is also available at Tidbit: https://tidbitapp.io/tidbits/effect-of-daily-vitamin-d-supplementation-on-risk-of-upper-respiratory-infection-in-older-adults-a-randomized-controlled-trial

In a randomized trial of >15 000 older adults not selected for vitamin D deficiency, vitamin D supplementation did not prevent upper respiratory infections overall. Whether effects differ in subgroups requires further study.

Over the past 20 years, the health effects of vitamin D supplementation have garnered worldwide attention. While vitamin D is known to have major effects on bone health, many observational studies have demonstrated that lower vitamin D status—as measured by blood concentration of 25-hydroxyvitamin D, or 25(OH)D—is associated with higher risk of major chronic diseases, including cardiovascular disease and cancer [1]. Critics attribute these associations to reverse causality—that is, sicker individuals are less likely to be exposed to outdoor sunlight, which is the major driver of vitamin D status [2]. Indeed, randomized controlled trials (RCTs) testing the effect of vitamin D supplementation for the prevention of cardiovascular disease and cancer are largely null [3]. However, there is growing appreciation of the complexity of participant factors, vitamin D dosing regimen, and the chosen comparison group, which complicate RCT evaluation of the effect of vitamin D in the prevention (or treatment) of these diseases [4].

Concurrent with this research, many observational studies have shown an inverse association between vitamin D status and risk of acute respiratory outcomes, including acute respiratory infection (ARI) and exacerbations of asthma and chronic obstructive pulmonary disease (COPD) [5]. The observational evidence is particularly strong for an inverse nonlinear association between 25(OH)D and ARI [6]; the sharpest increase was among those with 25(OH)D <15 ng/mL. Moreover, a Mendelian randomization study supports an inverse association [7]. As with cardiovascular disease and cancer, initial observational studies were followed by RCTs and these trials, while yielding mixed results, provide overall evidence that vitamin D supplements may prevent 3 major respiratory outcomes: ARI [8], asthma exacerbation [9], and COPD exacerbation [10]. Focusing on ARI, an individual participant data meta-analysis (based on 25 trials with 10 933 participants) found an overall protective effect [8]. Benefit was greater among individuals with very low vitamin D status (blood 25(OH)D <10 ng/mL) and using nonbolus dosing (eg, daily supplements); benefits were similar for upper respiratory infection (URI) and lower respiratory infection (eg, pneumonia).

Recently, 2 large ARI trials found no benefit from bolus dosing regimens of 100 000 IU monthly in New Zealand adults [11] and 14 000 IU weekly in Mongolian children [12]. An updated RCT meta-analysis including these 2 large trials confirmed a modest overall benefit, especially among children and with daily doses of 400–1000 IU for up to 12 months [13]. These publications have renewed interest in the role of vitamin D in ARI prevention, including identification of the optimal candidates for treatment and most effective dosing regimens. These questions heightened during the coronavirus disease 2019 (COVID-19) pandemic [4]. Moreover, because previous trials have lacked large numbers of Black participants, there is particular interest in the effects of vitamin D supplements in this population [14].

To address these important issues, we analyzed data from the recently completed Vitamin D and Omega-3 Trial (VITAL) [3]. VITAL enrolled >25 000 older and racially diverse adults into a multiyear RCT focused on cardiovascular disease and cancer. We hypothesized that daily vitamin D supplements (2000 IU/day), as compared to placebo, would lower risk of URI in participants with low vitamin D status.

METHODS

Trial Design, Participants, and Analytic Samples

VITAL was a randomized, double-blind, placebo-controlled trial, with a 2 × 2 factorial design, that examined the effect of vitamin D₃ (cholecalciferol, 2000 IU/day) and marine n-3 fatty acids (1 g/day) in the primary prevention of cardiovascular disease and cancer [3]. VITAL was approved by the local institutional review board and was monitored by an external data and safety monitoring board. All participants provided written informed consent before enrollment.

The 25 871 nationwide participants were generally healthy men (aged ≥50 years) and women (aged ≥55 years) and reported no history of cardiovascular disease nor cancer (except nonmelanoma skin cancer) at trial entry. Safety exclusions included renal failure or dialysis, cirrhosis, history of hypercalcemia, or other serious conditions. Because of the multiyear trial duration, and the risk of assigning placebo for 5 years to individuals with vitamin D deficiency, all participants were permitted to take up to 800 IU/day of supplemental vitamin D (the recommended dietary allowance [RDA] for older adults). Randomization was computer-generated within 5-year age groups, sex, and race in blocks of 8. The trial protocol has been described elsewhere [15] and the main results have been published [3].

This prespecified analysis focuses on the vitamin D versus placebo interventions among participants with baseline serum 25(OH)D level available (Figure 1). Baseline blood samples were provided by 16 956 individuals (66%) and baseline serum 25(OH)D level was available in 15 804 (61%). A second analytic sample involves 2013 participants who were enrolled during the first 2 winters of VITAL enrollment into a special URI subcohort that underwent serum 25(OH)D testing both at baseline and at 1 year; 1611 (80%) had both 25(OH)D values. Finally, we present results from comparable (albeit post hoc) analyses of all VITAL participants, regardless of their availability of baseline blood samples (Supplementary Figure 1, Supplementary Tables 1–3).

Figure 1.

Flowchart for trial participants with baseline vitamin D status (N = 15 804). Abbreviation: 25(OH)D, 25-hydroxyvitamin D.

Serum 25(OH)D Testing

Quest Diagnostics donated and performed the serum 25(OH)D assays with the use of liquid chromatography–tandem mass spectrometry [16]. For the 25(OH)D measurements, VITAL participated in the vitamin D standardization program of the Centers for Disease Control and Prevention [17]. Vitamin D deficiency was defined as serum total 25(OH)D <12 ng/mL [18]. Serum 25(OH)D levels were split at clinical cutpoints suggested by prior studies: <12, 12–19.9, 20–29.9, and ≥30 ng/mL. To convert to nmol/L, multiply by 2.496.

Interventions

The trial agents received Investigational New Drug Approval from the Food and Drug Administration. Pharmavite donated vitamin D and Pronova BioPharma/BASF donated fish oil (Omacor); the companies also donated matching placebos and packaging in the form of calendar packs. Calendar packs containing the trial capsules of vitamin D or corresponding placebo (and n-3 fatty acids or corresponding placebo) were mailed with questionnaires to the participants.

Baseline Characteristics and Follow-up

Baseline questionnaires collected data on demographic, clinical, and lifestyle factors, including age, sex, race/ethnicity (recategorized as non-Hispanic White, Black, other), body mass index (<25, 25–29.9, and ≥30 kg/m²), smoking status (never, past, current), asthma, COPD, current use of supplemental vitamin D (≤800 IU/day), and intake of foods related to vitamin D (milk, other vitamin D–fortified foods). Participants received follow-up questionnaires at 6 months and 1 year after randomization and annually thereafter to collect information on adherence to trial regimen, clinical outcomes, and potential adverse effects of the trial agents.

URI Outcome

The prespecified URI outcome was: “In the past few days, have you had a cough, cold, or other acute illness?” (no/yes), which was taken verbatim from the National Health and Nutrition Examination Survey [19] and was the primary outcome of an earlier investigation on vitamin D status and URI [20]. This question was asked on every annual questionnaire.

Statistical Analysis

The parent trial was designed for the primary end points of cardiovascular disease and cancer [3, 15]. The URI analysis was conducted on the available sample, with a prespecified focus on participants with baseline serum 25(OH)D level. Analyses of effect were based on the intention-to-treat principle, after omitting subjects with missing outcome data. Data were analyzed using SAS version 9.4 (SAS Institute, Cary, North Carolina). Additional analytic details are reported in the Supplementary Data.

Initial analyses compared baseline characteristics of participants according to their assigned vitamin D group, active or placebo (Table 1). In the subset of 1611 participants with serum 25(OH)D from both baseline and 1-year follow-up, we used Wilcoxon 2-sample test to compare serum 25(OH)D changes by treatment group.

Table 1.

Baseline Characteristics Among Vitamin D and Omega-3 Trial (VITAL) Participants With Baseline Serum 25-Hydroxyvitamin D, According to Randomized Vitamin D Assignment (N = 15 804)

Baseline Characteristic	Overall (N = 15 804)	Vitamin D
		Active (n = 7905)	Placebo (n = 7899)
Age, y, mean (SD)	68 (7)	68 (7)	68 (7)
Age group, y
50–54	462 (3)	235 (3)	227 (3)
55–64	4657 (29)	2315 (29)	2342 (30)
65–74	8369 (53)	4201 (53)	4168 (53)
≥75	2316 (15)	1154 (15)	1162 (15)
Sex, female	8033 (51)	4034 (51)	3999 (51)
Race/ethnicity
Non-Hispanic White	11 789 (76)	5907 (77)	5882 (76)
Black	2426 (16)	1211 (16)	1215 (16)
Other	1237 (8)	602 (8)	635 (8)
BMI, kg/m2, mean (SD)	27.9 (5.6)	28.0 (5.6)	27.8 (5.6)
BMI category, kg/m2
<25.0	4987 (32)	2439 (32)	2548 (33)
25.0–29.9	6240 (40)	3134 (41)	3106 (40)
≥30.0	4207 (27)	2152 (28)	2055 (27)
Smoking status
Never	8089 (52)	4065 (52)	4024 (52)
Past	578 (42)	3269 (42)	3309 (42)
Current	923 (6)	461 (6)	462 (6)
Asthma	1684 (11)	820 (10)	864 (11)
Chronic obstructive pulmonary disease	815 (5)	406 (5)	409 (5)
Baseline vitamin D supplement (≤800 IU/d)	7300 (46)	3651 (46)	3649 (46)
Intake of foods related to vitamin D, mean (SD) servings/d
Milk	0.7 (0.9)	0.7 (0.9)	0.7 (0.9)
Other vitamin D–fortified foods	0.6 (0.7)	0.6 (0.8)	0.6 (0.7)
Baseline 25(OH)D level, ng/mL, mean (SD)	31 (10)	31 (10)	31 (10)
Baseline 25(OH)D category
<12 ng/mL	377 (2.4)	193 (2.4)	184 (2.3)
12–19.9 ng/mL	1626 (10)	781 (10)	845 (11)
20–29.9 ng/mL	5092 (32)	2538 (32)	2554 (32)
≥30 ng/mL	8709 (55)	4393 (56)	4316 (55)
Concurrent randomization to n3FA
Placebo group	7891 (50)	3947 (50)	3944 (50)
Active agent group	7913 (50)	3958 (50)	3955 (50)

Data are presented as No. (%) unless otherwise indicated.

Abbreviations: 25(OH)D, 25-hydroxyvitamin D; BMI, body mass index; n3FA, n-3 fatty acids; SD, standard deviation.

Primary analyses compared the main effects of vitamin D on URI at 1 year using logistic regression models adjusted for age, sex, and race/ethnicity (Table 2), overall and by baseline 25(OH)D level. The models yielded odds ratios (ORs) and 95% confidence intervals (CIs). We specifically examined the prespecified subgroup of primary interest: participants with vitamin D deficiency (<12 ng/mL) who denied taking concurrent vitamin D; we hoped that this comparison would address the effect of the VITAL intervention (2000 IU daily) versus true placebo—that is, without influence of concurrent vitamin D supplementation (up to 800 IU daily). Given the known seasonality of both vitamin D status and URI [20], and the focus on recent URI, we also examined the effect of vitamin D on recent URI by season, using the month of the Year 1 follow-up questionnaire response to classify each participant into summer (June–August), fall (September–November), winter (December–February), or spring (March–May).

Table 2.

Odds Ratios and 95% Confidence Intervals for Recent Upper Respiratory Infection at 1-Year Follow-up, According to Randomized Vitamin D (Versus Placebo) Assignment and Baseline Vitamin D Status, Overall and by Season (N = 15 804)

Outcome	Overall		Fall		Winter		Spring		Summera
	OR (95% CI)	P Value	OR (95% CI)	P Value	OR (95% CI)	P Value	OR (95% CI)	P Value	OR (95% CI)	P Value
Overall	n = 15 013 1674 events		n = 4533 406 events		n = 4378 636 events		n = 2664 348 events		n = 3438 284 events
	0.96 (.86–1.06)	.38	1.06 (.86–1.30)	.61	1.02 (.86–1.21)	.79	0.94 (.75–1.18)	.59	0.72 (.56–.93)	.01
By 25(OH)D level
<12 ng/mLb	n = 315 36 events		n = 45 3 events		n = 87 8 events		n = 112 14 events		n = 71 11 events
	0.64 (.31–1.29)	.21	0.48 (.04–6.47)	.58	0.42 (.09–1.90)	.26	1.25 (.39–3.96)	.71	0.40 (.10–1.63)	.20
12–19.9 ng/mL	n = 1502 217 events		n = 346 43 events		n = 424 70 events		n = 396 70 events		n = 336 34 events
	0.92 (.69–1.24)	.60	1.06 (.55–2.06)	.86	1.01 (.60–1.70)	.96	0.76 (.45–1.29)	.31	1.02 (.49–2.11)	.97
20–29.9 ng/mL	n = 4818 569 events		n = 1356 132 events		n = 1420 213 events		n = 919 130 events		n = 1123 94 events
	1.00 (.83–1.19)	.97	1.04 (.72–1.49)	.85	1.02 (.76–1.38)	.88	1.14 (.78–1.68)	.50	0.75 (.49–1.16)	.20
≥30 ng/mL	n = 8378 852 events		n = 2786 228 events		n = 2447 345 events		n = 1237 134 events		n = 1908 145 events
	0.96 (.83–1.11)	.55	1.08 (.82–1.42)	.59	1.06 (.84–1.33)	.63	0.88 (.61–1.27)	.49	0.67 (.47–.95)	.03

Sample sizes (n) and numbers of events are raw counts after omitting subjects with missing outcome data. All models include age, sex, and race/ethnicity.

Abbreviations: 25(OH)D, 25-hydroxyvitamin D; CI, confidence interval; OR, odds ratio.

^a P for interaction for randomized assignment × recent upper respiratory infection (URI) × summer (yes/no) = .02.

^b P for interaction for randomized assignment × recent URI × vitamin D deficiency (yes/no) = .30.

We further investigated possible variations in the treatment effect according to prespecified demographic factors (including age, sex, race/ethnicity), clinical and lifestyle factors, and concurrent randomization to the n-3 fatty acid group (Table 3). Last, for completeness, we used repeated-measures logistic regression to address the overall effect of vitamin D on URI across all 4 years of follow-up.

Table 3.

Recent Upper Respiratory Infection at 1-Year Follow-up, Comparing Vitamin D and Placebo Groups, According to Baseline Characteristics (N = 15 804), Overall and During Summer

Characteristic	Overall				During Summer
	No. of Events		OR (95% CI)	P interaction	No. of Events		OR (95% CI)	P interaction
	D	P			D	P
Age
< median of 67.6 y	400	454	0.88 (.76–1.02)	.12	78	102	0.75 (.54–1.03)	.71
≥ median of 67.6 y	416	404	1.04 (.89–1.20)		42	62	0.67 (.45–1.02)
Sex
Male	403	421	0.98 (.85–1.14)	.62	47	56	0.88 (.58–1.33)	.25
Female	413	437	0.93 (.80–1.08)		73	108	0.64 (.47–.88)
Race/ethnicity
Non-Hispanic White	568	573	0.98 (.87–1.11)		78	97	0.78 (.57–1.06)
Black	151	185	0.79 (.63–.99)	.11	33	52	0.59 (.37–.95)	.35
Other	80	75	1.15 (.82–1.61)	.39	6	9	0.75 (.26–2.21)	.91
Body mass index, kg/m2
<25.0	232	250	0.98 (.81–1.19)		31	34	0.92 (.55–1.54)
25.0–29.9	314	341	0.91 (.77–1.07)	.55	44	72	0.55 (.37–.83)	.12
≥30.0	254	243	1.00 (.83–1.21)	.89	42	53	0.79 (.51–1.22)	.67
Smoking
Never	407	435	0.92 (.80–1.07)		62	83	0.73 (.52–1.04)
Past	321	347	0.95 (.80–1.11)	.82	44	56	0.76 (.50–1.15)	.92
Current	73	67	1.11 (.77–1.61)	.38	13	20	0.67 (.30–1.47)	.78
Asthma
Yes	127	137	0.97 (.74–1.26)	.78	23	23	0.92 (.49–1.73)	.33
No	682	717	0.95 (.85–1.06)		94	140	0.67 (.50–.88)
COPD
Yes	76	66	1.19 (.82–1.73)	.17	17	22	0.80 (.38–1.70)	.65
No	721	783	0.92 (.82–1.02)		101	139	0.70 (.53–.92)
Baseline vitamin D supplement
No	455	475	0.95 (.83–1.09)	.90	77	94	0.78 (.57–1.08)	.36
Yes	361	383	0.96 (.82–1.12)		43	70	0.61 (.41–.92)
Concurrent randomization to n3FA
Placebo group	434	430	0.99 (.86–1.15)	.44	55	77	0.67 (.46–.97)	.59
Active agent group	382	428	0.92 (.79-1.06)		65	87	0.77 (.55–1.09)

Numbers of events are raw counts after omitting subjects with missing outcome data. All models include age, sex, and race/ethnicity.

Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; D, vitamin D group; n3FA, n-3 fatty acids; OR, odds ratio; P, placebo group.

A 2-sided P < .05 was considered statistically significant. P values and 95% CIs were not corrected to account for the multiple hypothesis tests because, given the known heterogeneity in the effectiveness of vitamin D supplementation in ARI prevention [5], we did not want to miss any potentially important findings [21]. Thus, subgroup findings should be interpreted with caution.

RESULTS

Figure 1 shows the Consolidated Standards for Reporting Trials (CONSORT) diagram for this URI analysis. In brief, among the overall VITAL sample of 25 871 adults, baseline serum 25(OH)D level was available in 15 804 (61% of randomized subjects). The mean age of the VITAL participants in this ancillary study was 68 years and 51% were women; 76% were non-Hispanic White, 16% Black, and 8% other race/ethnicity. The mean body mass index was 27.9 kg/m². Many participants (46%) were taking a vitamin D supplement at baseline, which they continued during the trial. At baseline, the mean serum 25(OH)D level was 31 ng/mL (standard deviation [SD], 10). Vitamin D deficiency (<12 ng/mL) was present in 377 participants (2.4%). Table 1 shows balance between the 2 randomly allocated groups across many baseline characteristics.

Adherence-related data support fidelity to the protocol. For example, in the subset of 1611 participants who underwent repeat serum 25(OH)D testing, the serum 25(OH)D level of the vitamin D group went from 30 ng/mL to 42 ng/mL, whereas the serum 25(OH)D in the placebo group did not change (30 to 29 ng/mL); mean individual change was +12 (SD, 9) versus −1 (SD, 7) ng/mL, respectively (P < .0001). As reported in the primary report [3], participant retention also was high (>93%) and the vitamin D intervention did not affect risk of participant-reported adverse events.

Overall, vitamin D supplementation did not prevent URI at the 1-year follow-up (Table 2). Recent URI was reported by 816 (10.8%) in the vitamin D group and 858 (11.5%) in the placebo group (OR, 0.96 [95% CI, .86–1.06]). In the seasonal analysis, we observed a significantly lower occurrence of URI in the vitamin D group, compared to placebo, during summer only (OR, 0.72 [95% CI .56–.93]; P for interaction = .02). Table 2 also shows the effect of the vitamin D supplement on URI, stratified by baseline vitamin D status. The overall OR was lower among those with <12 ng/mL (OR, 0.64) than those with higher vitamin D status (ORs, 0.92–1.00) but the stratified results were nonsignificant (P for interaction = .30). The prespecified subgroup of primary interest—baseline 25(OH)D <12 ng/mL and denied taking concurrent vitamin D—had only 255 participants and the OR was lower but nonsignificant (OR, 0.60 [95% CI, .28–1.30]).

Table 3 explores potential effect modification by baseline characteristics other than serum 25(OH)D status. There was a suggestion of benefit among Black participants (OR, 0.79 [95% CI, .63–.99]; P for interaction = .11). An exploratory analysis of the summer subgroup revealed that Black participants were the most likely to benefit (OR, 0.59 [95% CI, .37–.95]). However, when we explored this finding in a post hoc analysis of all VITAL participants (Supplementary Table 3), we observed no such association (OR, 0.98 [95% CI, .83–1.15]).

To address potential differences in short-term (1-year) versus longer-term (multiyear) effects, we performed a repeated-measures analysis to examine all 4 years of follow-up. The overall, multiyear OR was null (OR, 1.02 [95% CI, .97–1.07]). Restricting this analysis to Black participants, the multiyear OR was 0.96 (95% CI, .87–1.07).

To provide context for these largely null RCT findings, we also examined the observational (noninterventional) association between baseline 25(OH)D levels and risk of URI in the placebo group only (Supplementary Table 4). Curiously, we did not observe the strong inverse association seen in many prior cohort studies [5, 6], even for those with baseline 25(OH)D <12 ng/mL. However, we did observe elevated odds of URI among Black participants, particularly in summer (OR, 3.17 [95% CI, 1.99–5.04]).

DISCUSSION

In this large RCT of generally healthy older US adults, we found that a vitamin D supplement of 2000 IU daily did not lower risk of URI, as measured by the prespecified primary outcome of recent URI [19] after 1 year of treatment. We examined potential effect-modifiers, including baseline 25(OH)D level, and these analyses were mostly nonsignificant. Although Black participants may have derived benefit, this also may have been due to chance. Unexpectedly, an analysis by season of self-reported URI showed significant benefit during summer. Taken together, the vitamin D intervention appeared safe and did not prevent URI in the overall study population.

To better understand the potential effect of vitamin D on ARI, which encompasses both upper and lower respiratory infections, it is important to consider factors related to the individual participant, the vitamin D regimen, and the comparison group [4]. Starting with the participant, the most important factor in observational research has been baseline vitamin D status, with lower baseline 25(OH)D levels consistently associated with higher risk of ARI [5–7]. Indeed, prior RCTs have shown that daily vitamin D supplementation is more beneficial among individuals with very low vitamin D status [8]. Regardless, individuals with clear-cut vitamin D deficiency (<12 ng/mL) should be treated with vitamin D because of its known benefits for bone health [8]. The role of vitamin D supplementation in Black individuals remains unclear; to date, no RCTs (except VITAL) have included many Black participants.

In the current trial, only 377 (2.4%) participants had baseline levels of <12 ng/mL. The overall percentage of VITAL participants with 25(OH)D <12 ng/mL (2.4%) reflects the proportion seen among non-Hispanic White Americans (2.3%) [22]. However, vitamin D deficiency is more common in non-White populations. For example, 24% of African Americans have blood 25(OH)D levels <12 ng/mL [22]. If low baseline vitamin D status is essential for ARI benefit, this suggests that a daily dose of vitamin D would be more (or less) effective in populations with differing levels of baseline 25(OH)D. However, the relative role of the baseline 25(OH)D level versus race/ethnicity in identifying patients who benefit remains uncertain. We found no clear evidence of a modifying effect of baseline serum 25(OH)D or baseline personal use of vitamin D supplements up to RDA doses (a design feature given the long duration of the main VITAL trial and ethical concerns about a vitamin D–deficient participant being assigned to placebo for several years); VITAL did not target vitamin D–deficient patients for enrollment.

Looking beyond participant factors, the vitamin D regimen also plays a role [8]. Prescribing decisions include the frequency of dosing (eg, daily or nondaily), dose (eg, 600–800 IU or higher), and duration of supplementation (eg, <12 months or longer). The vitamin D regimen in VITAL was 2000 IU daily, which has been found effective for ARI prevention in some earlier trials [8]. The trial duration of 4 years allowed us to extend prior work by investigating potential differences in early versus late effects, but overall results were nonsignificant both at 1 year and longer.

Background use of low-dose vitamin D supplements did not appear to modify the findings (Table 3). Moreover, we examined the 255 participants who had both vitamin D deficiency (<12 ng/mL) and who denied taking concurrent vitamin D at baseline. While statistical power was limited, the odds ratio (OR, 0.60) resembled the earlier observational results with this URI outcome (OR, 0.74) [20] but the RCT result was nonsignificant (95% CI, .28–1.30).

While we accept the likelihood that vitamin D supplementation does not reduce risk of ARI in the general population [5], we remain mindful of the myriad challenges of vitamin D supplement trials [4] and the many laboratory studies linking activated vitamin D with anti-infective immunologic effects [23, 24]. The challenge for vitamin D researchers is in translating these impressive laboratory findings to the bedside—and beyond. If vitamin D supplementation has any role in preventing ARI, who is the optimal recipient and what is the optimal regimen? Unfortunately, there remain no easy answers [25].

The VITAL study has many strengths, including its large size, nationwide participation, large number of Black participants, and rigorous multiyear RCT design. As with all studies, it also has limitations. Given the importance of low 25(OH)D levels to ARI risk, the generally replete baseline vitamin D status in VITAL and other contemporary trials is a major challenge. Analyses in the subgroup of interest (n = 255) required exclusion of >98% of the study population. Moreover, an observational analysis of the placebo group did not confirm the expected association between baseline 25(OH)D level and risk of URI—but did show an association between Black race and URI. Another limitation was the relatively simple self-report URI outcome. Although recall bias is minimized when asking about the “past few days,” it would have been better to have viral testing to confirm the URI diagnosis. Less precise outcomes (eg, URI in the past year) are even more likely to drive results toward the null.

In summary, in >15 000 generally healthy older adults not selected for vitamin D deficiency, supplemental vitamin D did not lower URI risk overall. The intervention appeared safe, without evidence of harm over 5 years. We encourage further research on the potential role of vitamin D supplementation in preventing URIs and other infectious diseases, and, more specifically, on the established heterogeneity in the RCT literature [8, 13]. Vitamin D also warrants study in the prevention (and treatment) of COVID-19 since there are observational studies suggesting benefit [26] and RCTs have mixed results [27, 28]. Consistent with the goals of precision medicine, we need to focus interventions (including vitamin D supplements) on those who are most likely to obtain true health benefits.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors are indebted to the VITAL participants and to the entire VITAL research group and staff for their dedicated collaboration. Voting members of the VITAL Data and Safety Monitoring Board include Lawrence S. Cohen, MD (Yale University School of Medicine); Theodore Colton, ScD (Boston University); Mark A. Espeland, PhD (Wake Forest University); Craig Henderson, MD (University of California, San Francisco); Alice H. Lichtenstein, ScD (Tufts University); Rebecca A. Silliman, MD, PhD (Boston University); and Nanette K. Wenger, MD (chair; Emory University); these members received a stipend for their service. Ex officio members at the National Institutes of Health (NIH) include Josephine Boyington, PhD, MPH; Rebecca B. Costello, PhD; Cindy D. Davis, PhD; Peter Greenwald, MD; Gabriela Riscuta, MD; and Harold Seifried, PhD. They also thank the Centers for Disease Control and Prevention (Drs Hubert Vesper and Julianne Cook Botelho) for their collaboration on the standardization and calibration of the serum 25(OH)D measurements throughout the study.

Data availability. Data are not publicly available.

Disclaimer. The views expressed in this article are those of the authors and do not necessarily represent the views of the National Institutes of Health. The funding/supporting organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Financial support. This work was supported by the National Institutes of Health (NIH) (grant numbers U01 CA138962, R01 CA138962, R01 AT011729, R01 AI093723, K24 HL136852). Pharmavite, Pronova BioPharma/BASF, and Quest Diagnostics provided study medications and diagnostic services.