|Home | About | Journals | Submit | Contact Us | Français|
Rationale: Restoration of vitamin D sufficiency may reduce asthma exacerbations, events that are often associated with respiratory tract infections and cold symptoms.
Objectives: To determine whether vitamin D supplementation reduces cold symptom occurrence and severity in adults with mild to moderate asthma and vitamin D insufficiency.
Methods: Colds were assessed in the AsthmaNet VIDA (Vitamin D Add-on Therapy Enhances Corticosteroid Responsiveness) trial, in which 408 adult patients were randomized to receive placebo or cholecalciferol (100,000 IU load plus 4,000 IU/d) for 28 weeks as add-on therapy. The primary outcome was cold symptom severity, which was assessed using daily scores on the 21-item Wisconsin Upper Respiratory Symptom Survey.
Measurements and Main Results: A total of 203 participants experienced at least one cold. Despite achieving 25-hydroxyvitamin D levels of 41.9 ng/ml (95% confidence interval [CI], 40.1–43.7 ng/ml) by 12 weeks, vitamin D supplementation had no effect on the primary outcome: the average peak WURSS-21 scores (62.0 [95% CI, 55.1–68.9; placebo] and 58.7 [95% CI, 52.4–65.0; vitamin D]; P=0.39). The rate of colds did not differ between groups (rate ratio [RR], 1.2; 95% CI, 0.9–1.5); however, among African Americans, those receiving vitamin D versus placebo had an increased rate of colds (RR, 1.7; 95% CI, 1.1–2.7; P=0.02). This was also observed in a responder analysis of all subjects achieving vitamin D sufficiency, regardless of treatment assignment (RR, 1.4; 95% CI, 1.1–1.7; P=0.009).
Conclusions: Our findings in patients with mild to moderate asthma undergoing an inhaled corticosteroid dose reduction do not support the use of vitamin D supplementation for the purpose of reducing cold severity or frequency.
Asthma exacerbations are frequently caused by viral respiratory tract infections (RTIs), but whether vitamin D supplementation can reduce the severity and rate of acute RTIs is unknown. Recent studies suggest that vitamin D supplementation may protect against RTIs, particularly when supplements are taken daily.
Our study suggests that replacement of vitamin D levels does not affect the severity of colds but may increase the frequency of cold symptoms in select groups of patients with asthma, particularly when the dose of inhaled corticosteroid is being reduced.
Vitamin D insufficiency is common in the general population and has been linked to susceptibility to respiratory infections (1–4). Physiologically, vitamin D alters the function of multiple cell types involved in both the innate and the adaptive immune systems, modulating pathways that can participate in the response to respiratory viral infections (5–9). Because of the results of epidemiologic and physiologic studies, trials have been conducted to determine whether vitamin D supplementation reduces respiratory infections, but the findings of these trials have been mixed. In children enrolled during the winter, vitamin D supplementation with 1,200 IU/d for 4 months reduced laboratory-confirmed influenza illnesses by 42% (10). Likewise, ingestion of vitamin D (300 IU/d)–fortified milk during the winter reduced the incidence of acute respiratory tract infections (RTIs) by 44% as assessed using questionnaires with children (11). However, monthly vitamin D doses of 100,000 IU for 18 months failed to reduce reported and laboratory-confirmed upper RTIs in adults (12). A metaanalysis of 11 randomized controlled trials indicated that vitamin D supplements may prevent RTIs when administered daily rather than in high doses intermittently, although heterogeneity among the trials prevented a definitive conclusion (13). Recently, researchers in a trial of 10,000 IU of vitamin D3 weekly for 2 months in the fall found that it did not affect the viral load or clinical rate of colds, but that it did reduce the incidence of laboratory-confirmed viral RTIs in healthy adults (14). Although very few of these studies selectively enrolled patients with vitamin D insufficiency, the results collectively suggest that vitamin D supplementation may protect against RTIs, particularly when the supplements are taken daily.
Patients with asthma are prone to more severe chest cold symptoms despite having rates of RTIs similar to those of healthy cohabitants (15). Because vitamin D insufficiency is also common in this population and may have independent effects on lung function and corticosteroid responsiveness (16), we recently completed a trial of vitamin D supplementation in patients with asthma and baseline 25-hydroxyvitamin D [25(OH)D] insufficiency (17). Vitamin D supplements taken daily for 28 weeks did not reduce the rates of first asthma treatment failure or exacerbations, although this treatment modestly spared the daily dose of inhaled corticosteroid (ICS) when the corticosteroid was tapered (17). There was a trend for a reduction of the overall rate of exacerbations, including multiple events per subject associated with assignment to vitamin D supplementation (hazard ratio, 0.63; 95% confidence interval [CI], 0.39–1.01; P=0.05). In a responder analysis, patients achieving vitamin D sufficiency at 12 weeks, regardless of treatment assignment, had a 40% reduction in overall exacerbations compared with placebo, with a 20% decrease noted for every 10–ng/ml improvement in 25(OH)D levels (17). Because asthma exacerbations are frequently caused by viral RTIs (15, 18–20), we sought to determine whether vitamin D supplementation could reduce the severity and rate of acute RTIs (common colds) and the associated changes in asthma control in adults with mild to moderate asthma and vitamin D insufficiency. A portion of this work has been presented previously as a poster (21).
We carried out a prospectively planned secondary analysis of a multicenter, randomized, double-blind, placebo-controlled clinical trial (the VIDA [Vitamin D Add-on Therapy Enhances Corticosteroid Responsiveness] trial) (17). Briefly, 408 adult patients with mild to moderate asthma, baseline serum levels of 25(OH)D3 less than 30 ng/ml, and asthma symptoms despite low-dose ICS therapy were randomized to receive placebo or cholecalciferol for 28 weeks (100,000 IU once, then 4,000 IU/d for 28 wk) as add-on therapy in the background of a tapering ICS protocol.
The visit structure and assessment of treatment failure, exacerbation, lung function, airway hyperresponsiveness, asthma symptoms, asthma control (measured using the Asthma Control Test [ACT] score ), asthma-specific quality of life, 25(OH)D levels, and ICS exposure have been described previously (17). Cold symptoms were assessed using the 21-item Wisconsin Upper Respiratory Symptom Survey (WURSS-21), a validated instrument with a range of 0–140 points and a minimal important difference of 18.5 (23, 24). Instructions on how to complete these surveys and their distribution occurred at the randomization visit, with reinforcement of its use at all subsequent visits, which occurred at 4- to 6-week intervals. Electronic diaries were used to assess asthma symptoms. In the diaries, participants answered the question, “Did you have a cold today”? If they answered “yes,” they were instructed to start completing WURSS-21 surveys daily. Survey completion continued until the first question on the survey (“How sick do you feel today?”) was answered “not sick” for 2 days in a row. All completed surveys were returned to the study coordinator at the participants’ next study visit. Each participant was given 21 copies of the WURSS-21 to keep at home in the event that he or she experienced a cold symptom between visits.
Melanin-dependent skin pigmentation is a measure of sun exposure, which can correlate with endogenous production of 25(OH)D (25). Melanin levels were estimated with a SmartProbe 400 spectrophotometer (IMS, Portland, ME). This device is used to measure degrees of pigmentation on a continuous scale from 0 to 100, with 0 being absolute black and 100 being absolute white. Pigmentation measurements were made from the forehead, outer forearm, inner upper arm, and abdomen, with two readings averaged and recorded at each location. We proposed that the difference between the forearm and abdominal readings would be a surrogate measure of degree of sun exposure. Skin pigmentation was measured before and after the 28 weeks of vitamin D supplementation.
Testing the primary hypothesis regarding the effect of vitamin D supplementation on the severity of colds was addressed by fitting a repeated-measures analysis of covariance (RM-ANCOVA) model to the peak WURSS-21 cold scores during each cold, allowing for multiple colds for each participant. Assuming that the incidence of colds in our study was similar to that seen in the Asthma Clinical Research Network’s Post-cold Asthma Control and Exacerbation study (24), we estimated that 65 participants per group would experience at least one cold (32.4%). With this sample size, we expected to have over 90% power to detect an effect size of 18, assuming that the power in our analyses was increased by the inclusion of multiple colds for each person (two-sided α, 0.05; SD, 32).
In addition to evaluating the average peak cold score from each cold, the average scores on Day 1 and Day 2 were also compared between the intervention and control groups. The rates of colds were compared between the treatment groups using Poisson regression. Each model included adjustment for center, African American race, and body mass index (BMI) greater than 25 kg/m2 (similar to the parent trial ). Exploratory analyses to investigate the effect of treatment on cold severity and rate at different levels of baseline factors [African American race, BMI >25 kg/m2, baseline 25(OH)D level, cold season] were conducted via the inclusion of nested effects in the RM-ANCOVA and Poisson regression models.
A responder analysis was performed to compare the cold severity and rate measures between those participants who became vitamin D sufficient [achieving a 25(OH)D level ≥30 ng/ml by 12 wk postrandomization] with those of participants who did not become vitamin D sufficient, regardless of treatment. This analysis was also performed using RM-ANCOVA and Poisson regression models as described above. In addition, 25(OH)D level at 12 weeks and change in the difference between the forearm and abdominal readings for melanin pigmentation were evaluated as potential factors in each of these models. The impact of an RTI on the subsequent development of a treatment failure or exacerbation, regardless of treatment assignment, was assessed using Kaplan-Meier methods for the first event and Poisson regression for the analysis of multiple events. The change in the ACT score during a cold was compared with the change in the ACT score throughout the duration of the study in those who did not experience a cold. This comparison was made using an RM-ANCOVA model adjusted for center, African American race, and BMI greater than 25 kg/m2. All tests were two-sided, with a P value less than 0.05 denoting significance. Without formal adjustment for the number of secondary analyses that were performed, the secondary results should be considered exploratory. All analyses were performed using SAS version 9.3 software (SAS Institute, Cary, NC).
Overall, 203 (49.8%) of the 408 participants experienced at least one self-reported cold during the course of the study. Tables 1 and and22 show the baseline (precold) characteristics of subjects who developed a cold during the study, stratified by treatment assignment. In the original trial, there were no baseline characteristic differences noted between treatment assignments (17). This randomization process was also successful when we evaluated participants who had a cold during the course of the trial; the exceptions noted were small differences in baseline melanin (Tables 1 and and22).
Figure 1 demonstrates the month of occurrence of the colds. The average peak WURSS-21 score for the group of subjects with colds was 61.3 (95% CI, 57.6–65.0). We defined time until RTI improvement as the time until the WURSS-21 score was less than 50% of the peak value for 2 consecutive days (24). The median time to improvement was 4 days with an interquartile range of 3–6 days. Participants with a cold had a decline in their ACT scores (average, −1.3; 95% CI, −1.8 to −0.8) at the visit after the cold that was significantly different from the change in ACT scores over the course of the entire treatment period for subjects who did not experience a cold (average, 0.2; 95% CI, −0.4 to 0.8; P<0.001) (Figure 2, left). The minimally important differences for the ACT score range between 2 and 3, depending on the methods and population studied (26). In our study, a cold increased the odds of having an ACT score decline of at least 2 (OR, 1.59; 95% CI, 1.01–2.49; P=0.043) (Figure 2, right).
There were no significant differences between treatment groups in the adjusted models for the WURSS-21 scores on Day 1, Day 2, the sum of Days 1 and 2, the sum of Days 1–4, or the sum of Days 1–7 (Figure 3A). The time course of the WURSS-21 scores is shown in Figure 3B, reflecting symptoms experienced by the participants during the first 14 days after the start of their colds, stratified by treatment assignment. The peak WURSS-21 cold score for participants assigned to placebo was 62.0 (95% CI, 55.1–68.9), and the peak for subjects receiving vitamin D was 58.7 (95% CI, 52.4–65.0). This difference was not significant before or after adjustment for center, African American race, and BMI greater than 25 kg/m2 (P=0.39 by RM-ANCOVA). There were no differences between treatment groups in terms of the time to improvement, even when we allowed for multiple colds per person.
Of those participants with at least one cold, the median number of colds per subject was 1, with an interquartile range of 1–2. Seventeen subjects had three colds, two subjects had four colds, and one subject had five colds. With respect to overall rate in the two treatment groups, there were 139 colds among the 207 subjects in the placebo group, for a rate of 1.24 colds per person-year, and 161 colds experienced by the 201 subjects receiving vitamin D, with a rate of 1.48 colds per person-year. The rate ratio for the effect of vitamin D on the rate of colds was 1.2 with a 95% confidence interval of 0.9–1.5 (P=0.15 in adjusted Poisson regression model). This effect was not different by center or BMI class; however, among African American participants, those receiving vitamin D had a higher rate of colds than those who received placebo (rate ratio, 1.7; 95% CI, 1.1–2.7; P=0.02).
Despite enrollment of 217 (53.2%) of 408 subjects with baseline 25(OH)D levels less than 20 ng/ml, 82% of the participants assigned to receive vitamin D achieved sufficiency, reaching a level greater than 30 ng/ml after 12 weeks of treatment, with a mean level of 41.9 ng/ml (95% CI, 40.1–43.7 ng/ml). Additionally, 9% of the subjects who received placebo treatment achieved 25(OH)D sufficiency (17). Figure 4 (light gray bars) shows similar overall rates of colds in patients stratified by their baseline 25(OH)D levels. In a responder analysis in which we compared only those subjects who achieved sufficiency with those who did not, regardless of treatment assignment, achieving vitamin D sufficiency had no effect on any of the cold severity measures (P>0.1 in all cases). Also, as shown in Figure 4 (dark gray bars), vitamin D responders had a higher rate of colds relative to nonresponders (1.4; 95% CI, 1.1–1.7; P=0.009). Analysis of the 25(OH)D level achieved at 12 weeks postrandomization as a predictor of cold severity did not reveal a difference when this was evaluated as a main effect (P>0.4 for all cold severity measures) or nested within cold season (P>0.1 for all seasons). Additionally, adjustment by the change in the difference between the forearm and abdominal readings for melanin pigmentation did not show significance (P>0.05 in all cases).
In symptomatic patients with mild to moderate asthma and baseline insufficiency of vitamin D in the setting of ICS dose reduction (17), those assigned to an effective vitamin D supplementation regimen did not have less severe or less frequent colds than subjects who received placebo treatment (Figure 3). Although this was a negative result, it is worth noting that the frequency of colds in this prespecified analysis conferred over 90% power to observe an effect size that was smaller than the minimal clinically important difference for the survey instrument used. The observation period was slightly longer than 6 months, ensuring that each subject was followed during at least one cold season. Adjustment for season of enrollment or sun exposure did not change these results. We conclude that the lack of effect on cold severity in the overall study population is valid and unlikely to change without significant modification of the study design.
A few caveats prevent broad generalization of this study’s results. These include the mild to moderate severity of baseline asthma and acute cold symptoms in the patient population studied. Patients with more severe asthma who had a history of frequent RTI-induced exacerbations might be an ideal population in which to evaluate respiratory samples during the course of the cold to confirm virus-associated events. Additionally, the ICS tapering protocol may have had unanticipated effects. For example, we hypothesized but did not observe (17) that achieving vitamin D sufficiency would allow for enhanced responsiveness to ICS with respect to lung function owing to the ability of vitamin D to influence steroid metabolism. Conversely, it is also possible that the change in ICS doses during the protocol influenced vitamin D metabolism and/or the expression of the vitamin D receptor and binding protein. With these considerations in mind, it is possible that we did not give enough vitamin D and that 25(OH)D levels in the serum do not reflect the changes relevant to airway epithelium innate immunity. This is unlikely in that the supplementation strategy used in this trial achieved a 25(OH)D level of 41.9 ng/ml (95% CI, 40.1–43.7 ng/ml) by 12 weeks (17) and epithelial cell conversion of 25(OH)D to the 1,25-dihydroxy active metabolite starts to plateau at around 40 ng/ml. These levels are sufficient for subsequent expression of cathelicidin and CD14 (27). Researchers who conduct future studies in this area should use a fixed dose of ICS (or none) and verify end-target effects of vitamin D supplementation, perhaps by including evaluation of changes in epithelial cell vitamin D–responsive gene expression.
We observed that achieving vitamin D sufficiency increased cold frequency. The direction of this effect is opposite to our hypothesis and appears to be driven by the subjects of African American race, who also had the lowest mean baseline vitamin D level (15.6 ng/ml [95% CI, 14.4–16.8] vs. 20.4 ng/ml [95% CI, 19.6–21.2] for all other races ). This is also contrary in some ways to the observation that patients achieving vitamin D sufficiency have a lower overall rate of exacerbations (17). It is unlikely that vitamin D levels have any influence on exposure to respiratory viruses. One speculation is that patients with 25(OH)D levels less than 20 ng/ml may be more likely than patients with higher levels to have asymptomatic upper RTI when exposed to common cold viruses, and in this case replacement of vitamin D could reconstitute the inflammatory responses related to symptom development. This reconstitution could lead to a greater likelihood of upper airway symptoms while also reducing the likelihood of lower airway infection and related risk of exacerbation. Of note, the median severity of the cold symptoms in a qualitative assessment was mild. Reconciling these possibilities would likely be best addressed by using an inoculation study design, before and after vitamin D supplementation.
In conclusion, the statistical power, frequency of study visits, and lack of a signal in the present trial suggest that longer studies or those in which higher vitamin D levels are achieved are unlikely to change the implication that restoration of vitamin D sufficiency does not impact cold severity in patients with mild to moderate asthma undergoing an ICS dose reduction and may increase the rate of symptomatic colds in patients with the lowest baseline levels of vitamin D.
The authors are grateful to the study participants, the AsthmaNet clinical research coordinators, and the data coordinating center. The authors thank Bruce Barrett, M.D. (University of Wisconsin, Madison), for the use of the AsthmaNet version of the short-form Wisconsin Upper Respiratory Symptom Survey.
Supported by NHLBI grants HL098102, U10HL098096, UL1TR000150, UL1TR000430, UL1TR000050, HL098075, UL1TR001082, HL098090, HL098177, UL1TR000439, HL098098, UL1TR000448, HL098107, HL098112, HL098103, UL1TR000454, and HL098115. In addition, L.C.D. has an R01 grant (HL115118) focused on the mechanisms of recovery from virus-induced asthma exacerbations.
Author Contributions: T.S.K. had full access to all the data in the study and had final responsibility for the decision to submit the manuscript for publication. Study concept and design: L.C.D., T.S.K., F.H., H.A.B., N.N.J., M.K., S.P.P., K.S., M.E.W., S.E.W., M.C., and P.C.A.; acquisition, analysis, or interpretation of data: L.C.D., T.S.K., J.C.C., F.H., H.A.B., N.N.J., S.P.P., K.R., M.E.W., M.C., and P.C.A.; drafting of the manuscript: L.C.D., T.S.K., T.C., N.N.J., K.S., S.P.P., M.E.W., S.E.W., M.C., and P.C.A.; critical revision of the manuscript for important intellectual content: L.C.D., T.S.K., J.C.C., T.C., F.H., D.J.J., S.P.P., K.R., K.S., H.A.B., M.E.W., S.E.W., M.C., and P.C.A.; statistical analysis: T.S.K.; study supervision: L.C.D., F.H., S.P.P., K.R., H.A.B., M.E.W., M.C., and P.C.A.
Originally Published in Press as DOI: 10.1164/rccm.201506-1169OC on November 4, 2015