Vitamin D and cancer: RCT

2. This is the second of 2 blogs on vitamin D and cancer, a randomized control trial of postmenopausal women treated with vitamin D and calcium, finding a nearly statistically significant benefit in preventing cancer (see vit d and cancer older women jama2017 in dropbox, or doi:10.1001/jama.2017.2115). see http://gmodestmedblogs.blogspot.com/2018/04/vitamin-d-and-cancer-risk-cohort-study.html for the first blog.

Details:

--2303 postmenopausal women, randomized to 2000 IU vit D3 plus 1500 mg calcium per day vs placebo for 4 years. 2064 (90%) completed the study. women were from 31 rural counties in Nebraska

--mean age 65, BMI 30, 99.5% white, 35% surgical menopause, 6% current smokers/67% never smokers, 15% estrogen therapy

--baseline 25(OH)D level 32.8 ng/ml [pretty high at baseline, being already vitamin D replete by the guidelines!!!], baseline median calcium supplement 600 mg/d, vitamin D 720 IU/d, dietary calcium 640 mg/d, dietary vitamin D 105 IU/d;

--primary outcome: first diagnosis of cancer (excluding non-melanoma skin cancer)

--secondary outcomes: specific cancers (breast, lung, colon, lymphoma, leukemia, myeloma); and hypertension, cardiovascular disease, osteoarthritis, colonic adenomas, diabetes, URIs and falls [the latter ones chosen because of prior studies finding mixed results on vitamin D providing benefit, see below]

Results:

--25(OH)D levels were 43.9 ng/ml in those on active meds (76% adherence rate) vs 31.6 ng/ml

--new diagnosis of cancer found in 109 people:

    --45 (3.89%) in the intervention group, vs 64 (5.58%) in the placebo group: difference 1.69% (-0.06% vs 3.46%), p=0.06

    --Kaplan-Meier incidence over 4 yrs (which excluded those who withdrew from the study after randomization): 4.2% (3.2 to 5.6%) vs 6.0% (4.8 to 7.6%), p=0.06

--review of the cancer incidence graph showed no benefit of the vitamin D/calcium supplementation in the first year, then curves of benefit splayed over the next 3 years [ie, after the first year, the benefit seemed to continue to increase over time until the end of the study]

--types of cancer (none of the differences were significant statistically): breast 16 in treatment group vs 23 placebo (breast in-situ, additional 3 vs 1), colorectal 4 vs 4, lung 5 vs 2, neuroendocrine 2 vs 4, all others total of 5 or less

--post-hoc analysis (excluding those who withdrew from the study, died, or developed cancer prior to being in the study for 12 months): 34 in intervention group vs 52 on placebo developed cancer in years 2-4, 3.17% vs 4.86%, p=0.046

--post-hoc analysis based on achieved 25(OH)D levels: the achieved level was inversely associated with cancer at p=0.03; comparing a level of 30 ng/ml vs 55 ng/ml: 35% reduction in cancer, HR 0.65 (0.44-0.97)

--adverse events: renal calculi (16 in intervention group and 10 in placebo), and elevated calcium levels (6 in intervention and 2 in placebo). all differences not statistically significant

--none of the secondary outcomes were reported in this article

Commentary:

--there has been concern in some studies about calcium (with or without vitamin D) and increased cardiovascular risk, with studies pretty much all over the place. the good news in this study is that they looked at this as a secondary outcome. the bad news is that the secondary analyses were not reported (with statement suggesting a subsequent report would be coming out. but none so far, to my perusal). in terms of some prior studies: see calcium intake and dec risk of CAD JAmHrtAssn2016 in dropbox, http://jaha.ahajournals.org/content/5/10/e003815, or doi: 10.1161/JAHA.116.003815, which found decreased incident atherosclerosis with increased calcium intake in the 10 year analysis of the Multi-Ethnic Study of Atherosclerosis, but also calcium and cad WHI bmj 2011​ in  dropbox, or doi:10.1136/bmj.d2040, which found that calcium with or without vitamin D modestly increased CAD risk in the 7-year Women's Health Initiative study.

--a meta-analysis (see Keum N. Brit J Cancer 2014: 11; 976) found that 2-7 years of vitamin D supplementation did not affect total cancer incidence but did significantly reduce total cancer mortality by 12% when taking 400-833 IU/day.  they did not find any difference in the incidence of cancer with attained 25(OH)D levels of ​20-30 ng/ml. ??if there might be benefit from higher achieved levels. and as a meta-analysis, they are combining quite disparate studies, limited by the actual specific data and targets reported in the very different trials.

--one real strength of this study is that there was pretty rigorous vitamin D supplementation (vs the Women's Health Initiative, which found no cancer-reducing effect of supplementation but used only 400 IU/d and had 50% adherence to that, though they still found a statistically significant inverse relationship between baseline 25(OH)D levels and colon cancer). and this current study included calcium supplementation, which by itself is important for vitamin D signaling, and animal and human studies find that calcium seems to lower levels of colon cancer and colon adenomas, though these results have been inconsistent in the literature. similarly, vitamin D has several potential cancer-reducing mechanisms, including promoting cell differentiation, inhibiting cancer cell proliferation, decreasing inflammation, being pro-apoptotic, and anti-angiogenic (see prior blog http://gmodestmedblogs.blogspot.com/2018/04/vitamin-d-and-cancer-risk-cohort-study.html​ ).

--the study has several limitations: 

    --the baseline 25(OH)D levels were enviably high (and certainly dramatically higher than i ever see in boston, and much higher than the US NHANES report finding that 75-80% of the population had levels <30 ng/ml and 30% below 20 ng/ml)

    --it was a pretty short study for cancer (which typically takes years to transition from cancerous cells to manifest disease)

    --there was no control for diet and exercise (both of which affect many cancers)

    --the baseline level of supplements and dietary intake of calcium/vitamin D do suggest this was a more health-conscious group and results from them may not be generalizable to many other settings

    ​--also the demographics were quite narrow as compared to the rest of the US. And men, who may have benefit in lowering prostate cancer, were not included

--it is very hard to know the optimal vitamin D intake associated with decreased cancer incidence: the Japanese study in the last blog suggested there might be a ceiling effect at the 23 ng/ml level, though this study suggested impressive benefit at the 55 ng/ml level over the 30 ng/ml level (35% less cancer!!)

--this is a bit like the comments in the fish oil study (see http://gmodestmedblogs.blogspot.com/2018/02/fish-oils-cardiovasc-disease-and.html ): supplementing vitamin D is an easy and essentially nontoxic med, and the fact that there was an almost 2% absolute decrease in cancer in those on vitamin D/calcium is pretty impressive despite the fact that it did not quite reach statistical significance (looking at the confidence intervals, the likelihood of actual benefit is statistically much, much higher than the likelihood of harm or non-benefit). This was also true in the last vitamin D blog (see http://gmodestmedblogs.blogspot.com/2018/04/vitamin-d-and-cancer-risk-cohort-study.html ).

--it might be useful to have a longer study (and, as noted above, the cancer-decreasing benefit seemed to be increasing over time in this short study), or one with women who were actually vitamin D deficient at baseline might be useful

so, my concern with vitamin D is that it’s normal blood level is predominantly related to sun exposure (and many of us who eat well, exercise, but live in much of the world with inadequate sunlight exposure have insufficient blood levels), 25(OH)D levels typically decrease in older age through decreased production efficiency by sunlight (and this is when more people develop cancer), there are vitamin D receptors all over the body and in the immune system (and there probably is an evolutionary reason for that, esp since humans seem to have evolved in pretty sun-drenched regions). And, supplementation is pretty nontoxic and cheap, so why not aim to achieve physiologically appropriate levels (though not exactly sure what they are, but my guess is at least 30 ng/ml, based on some small studies showing hyperparathyroidism at lower levels. but perhaps should they be higher for cancer prevention, as suggested in this study???). 

--the only other vitamin which i think is worth checking and correcting is vitamin B12:

    --which is available in foods associated with a healthy diet

    --but serum levels are below “normal” in about 10-15% in those over the ripe old age of 60 (probably multiple factors: more likely to be on meds like metformin which lead to vitamin B12 deficiency in 30%; more likely to have atrophic gastritis decreasing intrinsic factor; and also more likely to have decreased activity of salivary R-factor, also called haptocorrin, and pancreatic proteases, such as chymotrypsin, which are necessary to liberate the B12 from the food containing it)

    --as with vitamin D, vitamin B12 is similarly of potential benefit (esp neurologically/psychologically/hematologically, though I am unaware of any trials testing supplementation in a B12 deficient but asymptomatic population)

    ​--and it is also cheap and essentially nontoxic​ even in very high doses