vitamin D decreases risk of diabetes

 A systematic review and meta-analysis published last month found that vitamin D supplementation decreased the risk of prediabetes progressing to Type 2 diabetes (see vit d dec risk of diabetes AIM2023 in dropbox, or doi:10.7326/M22-3018)

 

Details:

-- 3 high-quality randomized controlled trials were found by a detailed search for studies designed to answer the specific question of prediabetes progression with vitamin D supplementation in adults, had low risk of bias, lasted at least two years, and evaluated oral vitamin D formulations. These trials had nearly identical design, and had remarkably similar methodology and patient selection. exclusion criteria included pregnant or lactating women, hospitalized patients including those in long-term care facilities, patients with end-stage renal disease, those with known diabetes or HIV, or those who were already on supplemental vitamin D 1000IU or more

    -- Tromso study (Norway): 511 participants recruited from 2008-2011, randomized to vitamin D 20,000 IU/week versus placebo

    -- Vitamin D and type II diabetes study (D2d study, US): 2423 participants recruited from 2013-2016, randomized to cholecalciferol 4000 IU per day versus placebo

    -- Diabetes Prevention with Active Vitamin D (DPVD, Japan): 1256 participants recruited from 2013-2015, randomized to eldecalcitol 0.75 µg/ day versus placebo (eldecalcitol is a synthetic form of 1,25(OH)2D)

-- total of 4190 patients, 2097 on vitamin D on 2093 on placebo

-- mean age 61, 44% women, 51% white/15% Black, current smokers 9%

-- participants taking vitamin D supplement at baseline 29% with median intake 331 IU/d; only 3 patients were taking more than 1000 IU. 21% were taking calcium supplements, median intake 134 mg/d

-- BMI <30 in 56%/30-35 in 27%/>35 in 17%

-- baseline serum 25(OH)D level: mean 25ng/mL;  <12 ng/mL in 5%/12-20 ng/mL in 24%/20-30 ng/mL in 41%/>30 ng/mL in 29%. mean serum calcium 9.2 mg/dL

--  mean fasting glucose 108 mg/dL, mean glucose 2 hours after 75 g oral load 146 mg/dL, mean hemoglobin A1c 5.9%

    --  48% met all 3 of the prediabetes glycemic criteria, 52% met 1 or 2

 

-- Primary outcome: development of new onset diabetes

-- secondary outcomes: regression to normal glucose regulation, and adverse events

 

-- median follow-up 3.0 years

-- results below were adjusted for baseline age, gender, BMI, race, and hemoglobin A1c levels

 

Results:

-- new onset diabetes:

    -- 475 of 2097 participants (22.7%) in the vitamin D group and 524 of 2093 (25.0%) in the placebo group

        -- 15% lower incidence of new diabetes with vitamin D, adjusted HR 0.85 (0.75-0.96)

        -- 8.42 versus 9.50 events per 100 person-years: absolute risk reduction by vitamin D was 3.3% (0.6%-6.0%), leading to the number needed to treat of 30 to prevent prediabetes from becoming diabetes

-- secondary analysis (which did not include those who stopped taking their pills, started using diabetes or weight-loss medication, or took supplemental vitamin D above 1000 IU/d):

    -- this outcome occurred in 447 participants on vitamin D and 505 in the placebo group, 8.26 and 9.61 events per 100 person-years, a 17% reduction with adjusted HR 0.83 (0.73-0.94)

-- subgroup analysis: the effect of vitamin D on new onset diabetes did not change based on age, gender, BMI, race, glycemic risk (high risk = at least 2 or 3 prediabetes criteria), or total calcium intake from supplements

 

-- assessment of new onset diabetes by baseline 25(OH) vitamin D levels:

    -- patients with <12 ng/mL 25(OH)D: 42% increased risk of diabetes, HR 0.58 (0.35-.97)

    -- patients with BMI below the median of 31.3, had a 36% decreased risk, HR 0.74 (0.60- 0.90), but those with BMI above the median had no benefit with HR 1.01 (0.82-1.20)

        -- BMI did not matter in those on eldecalcitol in the Japanese study, only the other 2 studies with cholecalciferol

 

-- assessment of new onset diabetes by mean serum 25(OH)D levels during the trial, as compared to those who had 25(OH)D levels of 20-29 ng/mL

    -- 25(OH)D 40-50 ng/mL: 400 people, 62% reduction of diabetes: HR 0.38 (0.27-0.55)

        -- absolute risk reduction at 3 years was 11.4% (4.6%-18.3%)

    -- 25(OH)D >50 ng/mL: 472 people, 76% reduction of diabetes: HR 0.24 (0.16-0.36)

        -- absolute risk reduction at 3 years was a 18.1% (11.7%-24.6%)

    -- no significant change in diabetes risk in the placebo group by changes in 25(OH)D levels 

 

 

-- Regression to normal glucose regulation:

    -- 271 of 1881 (14.4%) of the vitamin D group experienced regression to normal glucose regulation, versus 209 of 1889 (11.1%) in the placebo group

        -- the rate ratio for the vitamin D group versus placebo, at the last visit, was 1.30 (1.16-1.46)

 

-- safety:

    -- frequency of prespecified adverse events (kidney stones, hypercalcemia, and hypercalciuria) found nonsignificant trends of increases with vitamin D

 

Commentary:

-- several observational studies were consistent in finding a strong relationship between 25(OH)D levels and decreased risk of type II diabetes

-- prior meta-analyses did not find a significant relationship between vitamin D and diabetes, but they included trials with nondiabetes outcomes (post-hoc analyses), were of short duration, or had high risk of bias

-- the 3 trials analyzed in this meta-analysis were the studies that had low risk of bias, and had individual participant data which allowed more of a more precise characterization of their participants and more granular, evaluable data

    -- in these 3 trials, the reported risks of diabetes decreased 10-13% with the vitamin D supplementation, however this decrease was smaller than what these trials were powered  statistically to detect benefit (projected at 25% to 30% decrease)

-- The major conclusions of the study were:

    -- an overall decrease in prediabetes progression to diabetes in those on vitamin D supplementation

    -- those who achieved a higher 25(OH)D level had the largest benefit (a strong dose-response relationship), and this 25(OH)D level was much higher than the typical vitamin D goals of 600-800 IU/d

    -- those who had a low initial 25(OH)D level had a much higher benefit from vitamin D

    -- there was an increased rate of reversion to a non-prediabetic state in those on vitamin D

    -- there were no significant safety issues with this therapy

-- the overall relative risk reduction for diabetes was only 15%, which was significantly less than the 58% found with intensive lifestyle modification in the Diabetes Prevention Program, or with metformin having a 31% reduction in that trial

    -- the number needed to treat was 30 in the current trial, but 7 with intensive lifestyle modification and 14 with metformin

    -- even with the number needed to treat being 30 in the current trial, this would extrapolate to 10 million people not developing diabetes, of the 374 million adults worldwide with prediabetes

    -- though those achieving significantly higher 25(OH)D levels in this trial did have significantly more improvement, up to 76%, much more than in the Diabetes Prevention Program

    -- they did find a difference in 2 of the studies by BMI. the presumed reason is that these 2 studies used cholecalciferol supplements, and these require a two-step hydroxylation process to become the metabolically active chemical calcitriol; there is evidence that obesity suppresses vitamin D bioactivation (per CYP2R1) and weight loss upregulates it; and the one study of the 3 in this analysis that did not find any difference by BMI used eldecalcitol, a synthetic analog of calcitriol that does not need this enzyme for hydroxylation

 

-- benefits of vitamin D have been conflicting in the studies done to date. Several of these studies did not achieve a very high 25(OH)D level (in the low 20s), or did not even get much of a difference between the placebo in the vitamin D groups.

-- Argument for having higher vitamin D levels is basically:

    -- those living in areas such as Boston have very high levels of vitamin D deficiency (my anecdotal information confirms this, since I have been checking vitamin D levels for a few decades now, and when first checked 10% of my patients had undetectable levels, 10% had levels at least in the high 20s, and the rest mostly had levels from 10 to 20). The issue is that people in Boston need lots of sun from 10 AM to 2 PM and only in the middle of the summer, without a lot of clothing preventing the sunrays from hitting the skin. And people wearing sunscreen, or those with darker complexions, have a harder time getting sufficient sunlight. People older than 50 also have less sun-induced vitamin D synthesis

    -- there are vitamin D receptors on essentially all of cells of the body (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5031366/ ), including immunologic cells (eg monocytes, macrophages, T lymphocytes, antigen-presenting cells). This review article goes through many of the other potential benefits, including cancer (esp breast, prostate, nonmelanoma skin cancer, colorectal) finding vitamin D helps with inhibition of angiogenesis and metastases, DNA damage repair, decreased proliferation/increased apoptosis, etc;  cardiovascular benefits (inhibits renin production as well as atrial natriuretic peptide); skin conditions (eg psoriasis), immunologic diseases (eg multiple sclerosis, type 1 diabetes, Crohn's disease, rheumatoid arthritis, lupus, Grave's  dz). [of note, many of these findings are from the disease associations with vitamin D levels in retrospective analysis and not from clinical intervention trials, but are supported by animal and human cell studies in the lab: but to see a large number of prior blogs on vitamin d, see https://gmodestmedblogs.blogspot.com/search?q=vitamin+d ]. (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3738984/ ). But the retrospective trials cannot determine causality (eg, that actually increasing the vitamin D intake itself leads to improved outcomes). However, animal and human studies have shown that increasing vitamin D levels benefits immune function

    -- and there was even a study finding a graded benefit of 25(OH)D levels up to >40 ng/dL being associated with less severe Covid infections: https://gmodestmedblogs.blogspot.com/2022/02/covid-high-vit-d-prior-to-covid.html

        -- so it does seem to me that evolutionarily there is a role for higher vitamin D levels than we achieve in much of the world (after all, we humans evolved in the very sunny Fertile Crescent, but then moved on to less sun-friendly realms)

-- so this all brings up the real question: what is the optimal vitamin D level?? It is not clear...though this study adds to the argument. For bone health the minimum seems to be 20 ng/dL, though 30 ng/mL may be better. Observational studies for cancer have suggested that the level should be more like 50 ng/mL. And, this study does add that diabetes may be prevented best in prediabetics with levels >50ng/mL (which clearly exceeds the likely 25(OH)D level achieved by the guidelines suggesting 400-800 IU of vitamin D, except in very sunny areas near the equator and having lots of sun exposure and running around without a lot of clothes on, and without sunscreen, and (for vitamin D levels) having lighter skin complexion. And best to be younger than 50

 

Limitations:

-- typically, meta-analyses combine different studies with different mythologies and different interventions into a single result, making it hard to be sure of the results validity. one of the remarkable aspects of this diabetes study is that they found 3 RCTs with low likelihood of bias that were very similar, despite the fact they were in 3 very different countries: they all used modern definitions of prediabetes, had very similar methodology, had patient-level data, and each had very similar results

-- it is statistically problematic to assume that secondary analyses necessarily produce accurate conclusions: it might well be that the people who achieved the highest 25(OH)D levels (and had the best outcomes) were more adherent to their medications and perhaps did other things that were more healthful, such as dietary changes, exercise, weight loss, etc. So, there are likely some unmeasured confounders (and different in the different studies) that limits the validity of the study results

-- one may not be able to generalize the conclusions of the study to those who were excluded: children, lactating or pregnant women, hospitalized patients, those with end-stage renal disease, or those with HIV

-- and one cannot generalize the results to those without diabetes or who are at average risk for developing diabetes (vs the high risk of being prediabetic)

-- the issue of safety concerns is somewhat problematic in this and many studies. Typically, studies are powered statistically to produce a result that is likely to be statistically significant. Most studies are not powered sufficiently to produce statistically significant assessment of safety features

-- the mean A1C was 5.7%, which is low in the prediabetes range. so, the likelihood of progression to diabetes is low for the majority of patients in this 3-year study. Longer followup might well show more benefit from the vitamin D supplementation

 

So, this trial does bring up several interesting issues:

-- there may well need to be a higher 25(OH)D goal to achieve the diverse potential benefits of vitamin D

    -- prior studies that set a much less aggressive vitamin D supplementation and a lower 25(OH)D goal may not have sufficient vitamin D to produce positive results

    -- importantly, there did not appear to be many adverse effects to higher levels of vitamin D (unless the 25(OH)D levels are greater than 150 ng/mL), and vitamin D pills are really cheap

        -- but those not achieving really high 25(OH)D levels on 4000 IU or more should be evaluated for celiac disease. And there should be an evaluation for hypercalcemia and perhaps hypercalciuria, especially when on doses >800IU

-- this study also adds the argument we should be checking achieved 25(OH)D levels and adjusting supplements to achieve a higher level (perhaps in the 50 ng/mL range to get all the real and likely benefits), as opposed to just assuming that everyone in Boston needs vitamin D supplementation, and suggesting that they take 400 to 1000 IU of vitamin D. Sort of like tailoring statin doses to an achieved LDL goal

-- the increased likelihood of reversion to normoglycemia in those on vitamin D with prediabetes in this study also brings up a few points:

    -- restoring normoglycemia clearly seems to help prevent microvascular complications of diabetes

    -- and, prediabetes itself is associated with adverse effects, as per a few recent blogs (eg see http://gmodestmedblogs.blogspot.com/2022/11/prediabetes-increases-risk-heart.html for relationship with heart failure in those with atrial fibrillation, http://gmodestmedblogs.blogspot.com/2023/02/pancreatic-cancer-increases-with-higher.html  noting a higher incidence of pancreatic cancer and cardiac disease even down to A1c levels under 6%

-- and, it is likely that those who are non-prediabetic may well have less incidence of developing prediabetes or diabetes, perhaps especially in those with positive family histories of diabetes...

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs (2 options):

 

1. go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

2. go to https://bucommunitymed.wpengine.com/, a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs and is the easiest to view blogs and displays more at a time.

 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list

Comments

Popular posts from this blog

air pollution and heart disease

UPDATE: ASCVD risk factor critique

visceral fat loss is sustained long-term