Vitamin D3 for Life

Live Enlightened

Physiology

Vitamin D Status, Inflammation and Cardiovascular Risk Markers: Full Study

Previous related pageVitamin D Status, Inflammation and Cardiovascular Risk Markers: Study SummaryNext related pageVitamin D and Cardiovascular Health: Mechanisms and Clinical Outcomes
Study at a glance

Do vitamin D supplements change blood signs of inflammation?

This page compares studies linking vitamin D blood levels with later heart and circulation problems and trials testing changes in blood signs of inflammation. The main analysis followed 1,054 older adults assigned vitamin D3 or placebo within the VITAL trial. Most did not have severe vitamin D deficiency. A separate VITAL analysis examined blood samples after one year.

Which inflammation measurements changed with vitamin D3?

One inflammation-related blood marker, C-reactive protein measured with a high-sensitivity test, was lower with vitamin D3 after two years. The other three markers did not show significant reductions. After four years, none of the four markers showed a significant benefit. The exact size of the two-year change in this marker is uncertain because two ways of expressing the published result give slightly different percentages.

How reliable was the two-year inflammation finding?

The two-year result remained significant after allowing for tests of four markers in the main analysis. A separate analysis excluding some participants based on pill use did not meet that same standard. Fewer blood samples were available after four years, and many measurements of one marker could not be used. The study did not establish a group of people who clearly benefited more based on their starting vitamin D level.

Do these blood changes show fewer heart attacks or strokes?

The separate one-year analysis found no overall reduction in its three markers, and one rose. These were different participant samples, so the results do not prove a sequence of effects that starts, reverses and stops. Observational links between low vitamin D and later heart and circulation problems do not prove cause and effect. Changes in blood markers alone do not show fewer heart attacks or strokes. This evidence does not establish a broad, lasting reduction in inflammation or a supplement dose or blood target to protect the heart and circulation.

Read the full article below ↓

Does a low vitamin D blood concentration identify cardiovascular risk, and does raising vitamin D reduce inflammatory markers? Those are separate questions. This page principally examines the two: and four-year VITAL biomarker analysis by Dong and colleagues, alongside an earlier VITAL analysis and observational background. The findings vary by marker and follow-up time, and do not establish that a biomarker change prevents cardiovascular events. For the shorter account, read Vitamin D Status, Inflammation and Cardiovascular Risk Markers: Study Summary.

Three measurements that should not be confused

Vitamin D status is commonly assessed using circulating 25-hydroxyvitamin D, or 25(OH)D. Inflammatory biomarkers measure selected features of biological activity. Clinical outcomes, such as myocardial infarction or stroke, measure events experienced by people. The heart physiology overview provides the wider setting, but a result at one of these levels cannot automatically stand in for a result at another.

The distinction between vitamin D status and vitamin D effect is central here. People with low concentrations may differ in health, behaviour and other exposures from people with higher concentrations. A randomised trial instead asks what changes when an intervention is assigned. Even then, the answer is specific to the population, dose, duration and outcome studied.

Observational background: low status and later cardiovascular events

Wang and colleagues followed 1,739 Framingham Offspring participants without prior cardiovascular disease. Their mean age was 59 years, 55% were women and all were white. Over a mean 5.4 years, 120 participants developed a first cardiovascular event. A baseline 25(OH)D concentration below 15 ng/mL was associated with a multivariable-adjusted hazard ratio of 1.62, with a 95% confidence interval of 1.11-2.36, compared with concentrations at or above 15 ng/mL.

This is a relative association over follow-up. It does not mean that 62% of deficient participants experienced an event, and it is not an estimate of the benefit of supplementation. Adjustment for measured risk factors improves the comparison but cannot remove every unmeasured difference or establish that low vitamin D itself caused the events. The guide to correlation and causation explains why this observational finding motivates intervention research without determining its result.

The principal VITAL analysis: who was studied and what was assigned

Dong and colleagues analysed 1,054 Boston-area participants from VITAL who attended in-person assessments. VITAL was a randomised, double-blind, placebo-controlled trial using a two-by-two factorial design. Participants could receive vitamin D3, marine omega-3 fatty acids, both active agents or both placebos. The vitamin D comparison therefore pooled people assigned active vitamin D across both omega-3 assignments and compared them with those assigned vitamin D placebo.

Vitamin D3 was given at 2,000 IU daily. The separate marine omega-3 capsule contained 1 g of oil, including 460 mg EPA and 380 mg DHA. These were the study regimens, not personalised recommendations. Randomisation was centrally generated within sex, race and age strata. Participants, investigators, care providers and endpoint adjudicators were blinded to assignment.

The parent trial enrolled men aged at least 50 and women aged at least 55 without a history of major cardiovascular events or most cancers. In the biomarker cohort, mean age was 64.9 years, 49% were women and approximately 84% were white. Mean baseline 25(OH)D was 28.1 ng/mL, and 17% had concentrations below 20 ng/mL. This was not a trial restricted to people with severe vitamin D deficiency or an inflammatory disease.

Of these participants, 520 were assigned active vitamin D3 and 534 vitamin D placebo. Background medication use was common: about 41% used hypertension medication and 35% cholesterol medication. The results therefore concern older community-dwelling adults with varied risk factors, rather than a laboratory experiment that isolates a single inflammatory pathway.

Which inflammatory markers were measured?

The study measured four serum markers. They describe different parts of the inflammatory system and should not be merged into one score without justification. The inflammatory signalling overview explains the broader communication network within which such blood measurements sit.

MarkerMeaning in this study
High-sensitivity C-reactive protein, hs-CRPA sensitive measurement of a circulating acute-phase protein associated with systemic inflammation
Interleukin-6, IL-6A cytokine involved in inflammatory signalling
Tumour necrosis factor-α, TNF-αA cytokine involved in several inflammatory pathways
Interleukin-10, IL-10A cytokine with important anti-inflammatory actions

A blood concentration is not a direct reading of every tissue’s inflammatory activity. The experiment did not measure endothelial barrier leakage, plaque stability or heart-muscle contractility. The endothelial function pathway is a related physiological domain, but it requires its own measurements. An inflammatory-marker result cannot substitute for a vascular-function experiment.

Sample collection and analysis affect the interpretation

Fasting blood was collected at baseline, year two and year four, processed promptly and stored at −80°C. The analysis used 2,713 participant samples: 1,051 at baseline, 977 at year two and 685 at year four. These counts are blood samples at different visits, not 2,713 independently randomised people. There were also separate quality-control samples used to assess assay performance.

The assays did not yield usable values for every sample. Results outside the quantification limits were excluded: 0.5% for hs-CRP, 29% for IL-6, 2% for IL-10 and 0.1% for TNF-α. The particularly large IL-6 exclusion matters when judging precision and comparability between markers. A null estimate based on fewer usable measurements does not have exactly the same evidential weight as one based on almost all samples.

Because marker concentrations were skewed, the researchers analysed their natural logarithms. Mixed-effects models accounted for repeated samples from the same participant. The central treatment estimate was an interaction between treatment assignment and time, testing whether the trajectory differed between groups. The fully adjusted model also included the other randomised treatment, age, sex, race, body mass index, smoking and several medication categories.

These statistical adjustments do not turn the analysis into a different randomised trial. They address measured imbalance and improve the specified comparison within this cohort. The authors described the work as a post hoc biomarker analysis of a parent trial that was originally designed around cancer and cardiovascular outcomes. Its purpose and inferential limits should be understood at that level.

The positive result was confined to hs-CRP at two years

At year two, the vitamin D3 group had a lower hs-CRP trajectory than placebo. In the fully adjusted model, the log-scale treatment-by-time coefficient was −0.17 with a standard error of 0.06. The nominal P value was 0.007. After adjustment for testing four biomarkers, the reported P value was 0.028. The other three inflammatory markers did not show significant vitamin D effects at this time point.

The paper describes the hs-CRP finding as a 19% reduction. There is a numerical reporting nuance: directly exponentiating the rounded coefficient of −0.17 gives a ratio of approximately 0.84, or about 16% lower, under the usual natural-log interpretation. That calculation is an appraisal of the published coefficient, not a reanalysis of participant data. The precise percentage should therefore be treated cautiously; the supported finding is a lower hs-CRP trajectory at two years.

At year four, the fully adjusted hs-CRP coefficient was −0.02 with a standard error of 0.07 and P=0.753. No other marker showed a significant vitamin D effect at year four. The study consequently does not demonstrate a sustained reduction across all four years or a general lowering of every inflammatory signal. A statistically nonsignificant estimate also does not establish exact equivalence between groups.

Vitamin D comparisonFully adjusted hs-CRP coefficient, standard errorInterpretation
Two years−0.17, 0.06; nominal P=0.007Lower trajectory; remained significant after the reported four-marker correction
Four years−0.02, 0.07; nominal P=0.753No statistically significant separation in the reported analysis

Why adherence and subgroup findings need care

The investigators also analysed data after censoring follow-up from participants reporting low study-pill use or substantial outside supplementation. The two-year hs-CRP result retained its direction, but its nominal P value was 0.027 and its multiple-comparison-adjusted P value was 0.11. Thus, this sensitivity analysis did not meet the same corrected significance standard. It should not be described as equally strong confirmation of the primary analysis.

Removing participants according to behaviour after randomisation can also weaken the protection provided by random assignment. These analyses help assess robustness, but they are not automatically more reliable than the assigned-treatment comparison. The year-four result remained nonsignificant. The smaller number of available year-four samples adds another reason to examine uncertainty rather than assume a precise biological disappearance at a particular date.

The supplementary subgroup figure explored age, sex, race, smoking, body mass index, baseline nutrient status and medication use. Some within-subgroup comparisons reached nominal significance and others did not. However, the interaction tests did not establish that the vitamin D effect differed between the investigated subgroups. For baseline 25(OH)D below versus at least 20 ng/mL, the interaction P value was 0.47.

It would therefore be misleading to use that figure to identify a proven responder group or a blood concentration at which the intervention starts working. Statistical significance in one subgroup and nonsignificance in another is not itself a statistically significant difference between their effects. These analyses were exploratory and had limited power.

An earlier VITAL analysis found no overall anti-inflammatory reduction

Costenbader and colleagues examined a separate selected sample of 1,561 VITAL participants with paired baseline and one-year blood samples. They measured IL-6, hs-CRP and tumour necrosis factor receptor 2, or TNFR2. TNFR2 is a receptor-related marker of the TNF system; it is not the TNF-α concentration measured in Dong’s later analysis. The vitamin D regimen was again 2,000 IU daily within the factorial trial.

At one year, none of these three markers was significantly reduced by vitamin D. The reported net relative change in IL-6 was an increase of 8.19%, with a 95% confidence interval of 1.52%-15.31%. TNFR2 and hs-CRP did not show significant overall treatment effects. This finding does not establish clinical harm, just as a lower marker concentration would not by itself establish cardiovascular protection.

An exploratory pattern suggested a greater hs-CRP increase among participants with baseline 25(OH)D below 20 ng/mL. The study conducted numerous subgroup comparisons without a strict multiplicity correction. Such findings need confirmation and do not supply a treatment rule. The earlier samples were shipped for processing, whereas the Boston clinic samples were processed promptly on site; cohorts, assays and observation times also differed.

What these findings mean for cardiovascular interpretation

The one-, two: and four-year results are not a single uninterrupted measurement series in an identical sample. They should not be narrated as proof that vitamin D first raises inflammation, then lowers it, then stops working. Each analysis answers a defined question with its own participants, laboratory measurements and statistical uncertainty.

Biological hypotheses involving vitamin D receptor signalling remain relevant, but these trials did not directly demonstrate the molecular pathway responsible for an individual blood-marker change. They also did not establish that the hs-CRP finding mediated a change in cardiovascular-event risk. Demonstrating mediation would require additional evidence linking assignment, marker change and clinical outcomes within an appropriate analysis.

The evidence hierarchy helps keep the conclusions proportional. Observational low-status associations identify patterns worth investigating. Randomisation strengthens inference about assigned supplementation and the outcomes actually measured. Neither approach permits replacing a measured biomarker with an unmeasured clinical benefit.

The defensible conclusion is specific: a VITAL biomarker analysis found lower hs-CRP at two years, with qualifications about magnitude and sensitivity analyses, but not a sustained or broad anti-inflammatory effect across the measured markers. Together with the earlier null or upward findings, this evidence does not establish a cardiovascular supplement dose, a target vitamin D concentration or prevention of heart attacks and strokes.

Scientific sources

Dong Y, Zhu H, Chen L, Huang Y, Christen W, Cook NR, Copeland T, Mora S, Buring JE, Lee IM, Costenbader KH, Manson JE. Effects of Vitamin D3 and Marine Omega-3 Fatty Acids Supplementation on Biomarkers of Systemic Inflammation: 4-Year Findings from the VITAL Randomized Trial. Nutrients. 2022;14(24):5307. DOI: 10.3390/nu14245307. PMID: 36558465. Principal appraisal, including the supplementary subgroup figure.

Costenbader KH, MacFarlane LA, Lee IM, Buring JE, Mora S, Bubes V, Kotler G, Camargo CA Jr, Manson JE, Cook NR. Effects of One Year of Vitamin D and Marine Omega-3 Fatty Acid Supplementation on Biomarkers of Systemic Inflammation in Older US Adults. Clinical Chemistry. 2019;65(12):1508-1521. DOI: 10.1373/clinchem.2019.306902. PMID: 31699704. Earlier VITAL biomarker comparison.

Wang TJ, Pencina MJ, Booth SL, Jacques PF, Ingelsson E, Lanier K, Benjamin EJ, D’Agostino RB, Wolf M, Vasan RS. Vitamin D Deficiency and Risk of Cardiovascular Disease. Circulation. 2008;117(4):503-511. DOI: 10.1161/CIRCULATIONAHA.107.706127. PMID: 18180395. Observational background; the cohort finding is used to distinguish association from intervention evidence.

Related reading

Original sources

Source

Effects of Vitamin D3 and Marine Omega-3 Fatty Acids Supplementation on Biomarkers of Systemic Inflammation: 4-Year Findings from the VITAL Randomized Trial

Background: The VITAL study was a nationwide, randomized, double-blind, placebo-controlled, 2 × 2 factorial trial of vitamin D3 (2000 IU/day) and marine n-3 FAs (1 g/day) supplements.

Publication details

Dong Y, Zhu H, Chen L, Huang Y, Christen W, Cook NR, Copeland T, Mora S, Buring JE, Lee IM, Costenbader KH, Manson JE. Effects of Vitamin D3 and Marine Omega-3 Fatty Acids Supplementation on Biomarkers of Systemic Inflammation: 4-Year Findings from the VITAL Randomized Trial. Nutrients. 2022;14(24):5307. DOI: 10.3390/nu14245307. PMID: 36558465. Principal appraisal, including the supplementary subgroup figure.

Read more →
Source

Effects of One Year of Vitamin D and Marine Omega-3 Fatty Acid Supplementation on Biomarkers of Systemic Inflammation in Older US Adults

Published in Clinical Chemistry (2019).

Publication details

Costenbader KH, MacFarlane LA, Lee IM, Buring JE, Mora S, Bubes V, Kotler G, Camargo CA Jr, Manson JE, Cook NR. Effects of One Year of Vitamin D and Marine Omega-3 Fatty Acid Supplementation on Biomarkers of Systemic Inflammation in Older US Adults. Clinical Chemistry. 2019;65(12):1508-1521. DOI: 10.1373/clinchem.2019.306902. PMID: 31699704. Earlier VITAL biomarker comparison.

Read more →
Source

Vitamin D Deficiency and Risk of Cardiovascular Disease

Published in Circulation (2008).

Publication details

Wang TJ, Pencina MJ, Booth SL, Jacques PF, Ingelsson E, Lanier K, Benjamin EJ, D’Agostino RB, Wolf M, Vasan RS. Vitamin D Deficiency and Risk of Cardiovascular Disease. Circulation. 2008;117(4):503-511. DOI: 10.1161/CIRCULATIONAHA.107.706127. PMID: 18180395. Observational background; the cohort finding is used to distinguish association from intervention evidence.

Read more →