The relationship between vitamin D status or supplementation and cancer outcomes has been examined in several meta‐analyses. To address remaining knowledge gaps, we conducted a systematic overview and critical appraisal of pertinent meta‐analyses. For meta‐analyses of trials, we assessed their quality using AMSTAR‐2 (A Measurement Tool to Assess Systematic Reviews), strength of associations using umbrella review methodology and credibility of evidence using GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) criteria. Meta‐analyses of observational studies reported inverse associations of 25OHD with risk of cancer incidence and cancer mortality and, particularly for colorectal cancer, fulfilled some of Bradford‐Hill's causation criteria. In meta‐analyses of trials, vitamin D supplementation did not affect cancer incidence. However, we found credible evidence that vitamin D supplementation reduced total cancer mortality risk, with five out of six meta‐analyses reporting a relative risk (RR) reduction of up to 16%: RR, 0.84 (95% CI, 0.74–0.95). The strength of the association, however, was classified as weak. This was true among meta‐analyses of high, moderate, and lower quality (AMSTAR‐2–rated). Trials did not include large numbers of vitamin D‐deficient participants; many tested relatively low doses and lacked sufficiently powered data on site‐specific cancers. In conclusion, meta‐analyses show that, although observational evidence indicates that low vitamin D status is associated with a higher risk of cancer outcomes, randomized trials show that vitamin D supplementation reduces total cancer mortality, but not cancer incidence. However, trials with larger proportions of vitamin D‐insufficient participants and longer durations of follow‐up, plus adequately powered data on site‐specific common cancers, would provide further insight into the evidence base. © 2020 The Authors.
Ecological studies of cancer incidence and mortality have shown that sun exposure, especially solar UV B (and hence vitamin D production), is associated with reduced risk of many cancer types.(
Meta‐analyses provide accurate, succinct, credible, and comprehensive summaries of the best available evidence, and act as a key tool for health care professionals to achieve evidence‐based decisions.(
Thus, we decided to conduct a systematic overview and critical appraisal of pertinent meta‐analyses to better characterize the evidence on vitamin D status or supplementation in relation to cancer outcomes. In our critical appraisal, particular attention was given to intervention studies as this study design, as the gold standard for effectiveness research, provides the highest relevance for evidence‐based decision‐making. For this, we systematically assessed the quality of meta‐analyses, plus strength and credibility of the evidence from these studies across multiple cancer outcomes, and discussed differences between meta‐analyses. Finally, we discussed limitations of the evidence presented and provided some future research directions.
We searched Medline and PubMed for articles published up until 12 May 2020, using the following search terms: vitamin D, cancer, and meta‐analysis. No language restrictions were applied. This was supplemented by a manual search of reference lists from identified articles.
The quality of meta‐analyses was assessed using AMSTAR‐2 (A Measurement Tool to Assess Systematic Reviews), a 16‐point assessment tool of the methodological quality of systematic reviews.(
The strength of associations was evaluated based on umbrella review criteria.(
Umbrella Review Assessment Grades
Strength of association | Criteria |
---|---|
Convincing (class I) | >1000 cases |
Significant summary associations ( | |
No evidence of small‐study effects | |
No evidence of excess of significance bias | |
Prediction intervals not including the null value | |
Largest study nominally significant ( | |
Not large heterogeneity (I2 < 50%) | |
Highly suggestive (class II) | >1000 cases |
Significant summary associations ( | |
Largest study nominally significant ( | |
Suggestive (class III) | >1000 cases |
Significant summary associations ( | |
Weak (class IV) | Significant summary associations ( |
Nonsignificant association | Nonsignificant summary associations ( |
Total for the meta‐analysis.
The credibility of pooled estimates of meta‐analyses was qualitatively assessed using the GRADE (the Grading of Recommendations, Assessment, Development, and Evaluation) method.(
GRADE Assessment Scoring
Study design | Quality of evidence | Lower if | Higher if |
---|---|---|---|
Randomized trial | High |
Risk of bias: ‐1 Serious ‐2 Very serious Inconsistency ‐1 Serious ‐2 Very serious Indirectness ‐1 Serious ‐2 Very serious Imprecision ‐1 Serious ‐2 Very serious Publication bias ‐1 Likely ‐2 Very likely |
Large effect +1 Large +2 Very large Dose response +1 Evidence of a gradient All plausible confounding +1 Would reduce a demonstrated effect or +1 Would suggest a spurious effect when results show no effect |
Moderate | |||
Observational study | Low | ||
Very low |
From: Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, et al. GRADE guidelines: 1. Introduction—GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011;64(4):383–94.
GRADE = Grading of Recommendations, Assessment, Development, and Evaluation.
We found 35 meta‐analyses that investigated relationships between vitamin D status, as measured by circulating 25OHD and cancer outcomes: 29 on cancer incidence,(
Meta‐Analyses of Cohort Studies on the Association Between 25OHD and Cancer Outcomes
First author, publication year | Studies & design ( | Participants ( | Events ( | Unit of 25OHD comparison | Pooled association (95% CI) | I2 (%) or |
---|---|---|---|---|---|---|
Cancer incidence | ||||||
All | ||||||
Han, 2019(
| 8 PC | 70,018 | 7511 |
Highest vs. lowest group Per 20 nmol/L (dose–response) |
RR, 0.86 (0.73–1.02) RR, 0.93 (0.91–0.96) |
71 NR |
Breast | ||||||
Chen, 2010(
| 7 (4 CC, 3 NCC) | 11,330 | 5489 | Highest vs. lowest quartile | OR, 0.55 (0.38–0.80) | 86 |
Yin, 2010(
|
9 (5 CC, 4 NCC) 4 NCC |
12,901 6327 |
6147 3117 | Per 20 ng/mL |
OR, 0.73 (0.60–0.88) OR, 0.92 (0.83–1.04) |
84 NR |
Chung, 2011(
| 4 NCC | 4726 | 2363 | Per 10 nmol/L (dose–response) | OR, 0.99 (0.97–1.01) | NR |
Gandini, 2011(
|
10 (1 PC, 4 NCC, 5 CC) 5 (1 PC, 4, NCC) |
29,742 23,078 |
6175 3145 | Per 10 ng/mL |
RR, 0.89 (0.81–0.98) RR, 0.97 (0.92–1.03) |
88 54 |
Mohr, 2011(
|
11 (6 NCC, 5 CC) 6 NCC 5 CC |
16,337 9673 6664 |
7547 4517 3030 |
Highest vs. lowest quintile Highest vs. lowest quintile Highest vs. lowest quintile |
POR, 0.61 (0.47–0.80) POR, 0.87 (0.77–0.99) POR, 0.41 (0.31–0.56) |
|
Bauer, 2013(
|
Premenopause: 6 PC Postmenopause: 9 PC |
1613 3929 |
2890 8766 |
Per 5 ng/mL (dose–response) |
RR, 0.99 (0.97–1.04) RR, 0.97 (0.93–1.00) |
NR NR |
Chen, 2013(
|
21 (10 NCC, 1 RSP, 10 CC) 11 NCC/RSP |
26,317 6811 |
11771 15852 |
Highest vs. lowest quartile Highest vs. lowest quartile |
OR, 0.52 (0.40–0.68) OR, 0.86 (0.75–1.00) |
89 40 |
Wang, 2013(
|
14 (1 PC, 13 NCC) 11 (1 PC, 10 NCC) |
25,354 20,252 |
9110 6715 |
Highest vs. lowest group Per 10 ng/mL (dose response) |
RR, 0.84 (0.75–0.95) RR, 0.97 (0.94–0.99) |
38
|
Kim, 2014(
| 14 (1 PC, 13 NCC) | 27,534 | 9526 |
Highest vs. lowest group Per 10 ng/mL (dose response) |
RR, 0.92 (0.83–1.02) RR, 0.98 (0.96–1.00) |
27
|
Estébanez, 2018(
| 29 (14 NCC, 15 CC) | 58,855 | 18358 | High vs. low group | OR, 0.66 (0.57–0.76) | 41 |
14 NCC | 24,271 | 10266 | High vs. low group | OR, 0.92 (0.83–1.01) | 16 | |
4 PC | 16,875 | 3350 | Variable group comparisons | OR, 0.85 (0.74–0.98) | 4 | |
Hossain, 2019(
| 14 (12 NCC, 1 CC, 1 MR) | 123,044 | 25515 | Per 10 ng/mL | OR, 0.99 (0.98–1.00) | 79 |
5 CC | 2796 | 1306 | <10 ng/mL vs. ≥10 ng/mL | OR, 1.91 (1.51–2.41) | 83 | |
Song, 2019(
| 40 (4 PC, 36 CC) | 162,322 | 31157 | Per 5 nmol/L | OR, 0.94 (0.93–0.96) | 91 |
Colon | ||||||
Yin, 2009(
| 6 NCC | 2081 | 759 | Per 20 ng/mL | OR, 0.78 (0.54–1.13) | 45 |
Lee, 2011(
| 8 PC | 4578 | 1822 | Highest vs. lowest group | OR, 0.77 (0.56–1.07) |
|
Touvier, 2011(
| 6 NCC | 3550 | 1477 | Per 100 IU/L (dose–response) | RR, 0.95 (0.92–0.995) | 48 |
Rectum | ||||||
Yin, 2009(
| 4 NCC | 719 | 258 | Per 20 ng/mL | OR, 0.41 (0.11–1.49) | 63 |
Lee, 2011(
| 9 NCC | 4578 | 868 | Highest vs. lowest group | OR, 0.50 (0.28–0.88) |
|
Touvier, 2011(
| 5 NCC | 1645 | 721 | Per 100 IU/L | RR, 0.95 (0.89–1.05) | 67 |
Colorectal | ||||||
Gorham, 2007(
| 5 NCC | 1448 | 535 | Highest vs. lowest group | POR, 0.49 (0.35–0.68) |
|
Yin, 2009(
| 5 NCC | 3286 | 1199 | Per 20 ng/mL | OR, 0.57 (0.43–0.76) | 9 |
Chung, 2011(
| 9 NCC | 2249 | 1127 | Per 10 nmol/L | OR, 0.94 (0.91–0.97) | NR |
Gandini, 2011(
|
9 (1 PC, 7 NCC, 1CC) 8 (1 PC, 7 NCC) |
22,948 22,870 |
2630 2604 |
Per 10 ng/mL (dose–response) Per 10 ng/mL (dose–response) |
RR, 0.85 (0.79–0.91) RR, 0.85 (0.79–0.92) |
55 59 |
Lee, 2011(
| 8 NCC | 4578 | 2690 | Highest vs. lowest group | OR, 0.66 (0.54–0.81) |
|
Ma, 2011(
| 9 (7 PC, 2 NCC) | 6715 | 2767 |
Highest vs. lowest group Per 10 ng/mL (dose–response) |
RR, 0.67 (0.54–0.80) RR, 0.74 (0.63–0.89) |
0 NR |
Touvier, 2011(
| 6 NCC | 5833 | 2370 | Per 100 IU/L | RR, 0.96 (0.94–0.97) | 0 |
Huang, 2019(
| 30 (6 PC, 23 NCC, 1 CC) | 204,544 | 13051 | Highest vs. lowest group | RR, 0.68 (0.60–0.78) | 56 |
Zhang, 2019(
| 8 (1 NCC, 7 CC) | 9594 | 2916 |
Highest vs. lowest group Per 16 ng/mL (dose–response) |
OR, 0.75 (0.58–0.97) OR, 0.79 (0.64–0.97) |
54 54 |
Colorectal adenoma | ||||||
Wei, 2008(
|
All adenomas: 7 (1 CS, 3 CC, 3 NCC/PC) Advanced adenomas: 2 NCC |
3787 1023 |
2628 2347 |
Highest vs. lowest quintile High vs. low groups |
OR, 0.70 (0.56–0.87) OR, 0.64 (0.45–0.90) |
54 NR |
Fedirko, 2010(
| 3 CC | 1386 | 616 | Highest vs. lowest quartile | OR, 0.59 (0.41–0.84) | NR |
Yin, 2011(
|
Incident events: 9 (5 CC, 1 CS, 3 NCC) Recurrent events: 3 PC |
7654 2169 |
3539 984 |
Per 20 ng/mL Per 20 ng/mL |
OR, 0.82 (0.69–0.97) OR, 0.87 (0.56–1.35) |
66 57 |
Huang, 2019(
| 22 (5 PC, 2 NCC, 14 CC, 1CS) | 13652 | 6445 | Highest vs. lowest group | RR, 0.80 (0.71–0.89) | 34 |
Prostate | ||||||
Yin, 2009(
| 10 (1 PC, 9 NCC) | 7806 | 3124 | Per 10 ng/mL | OR, 1.03 (0.96–1.11) | 23 |
Chung, 2011(
| 8 NCC | 5609 | 2399 | Per 10 nmol/L (dose–response) | OR, 1.01 (0.99–1.04) | NR |
Gandini, 2011(
| 11 PC | 26,575 | 3956 | Per 10 ng/mL(dose–response) | RR, 0.99 (0.95–1.03) | 37 |
Gilbert, 2011(
| 14 (5 PC, 9 NCC) | 12,051 | 4353 | Per 10 ng/mL | OR, 1.04 (0.99–1.10) | 0 |
Kidney | ||||||
Gallicchio, 2010(
| 8 PC | 1550 | 775 | 50‐ < 75 vs. ≥100 nmol/L | OR, 0.92 (0.44–1.92) | P‐value: NS |
Liver | ||||||
Guo, 2020(
| 6 (1 PC, 5 NCC) | 60,811 | 992 | High vs. low group | RR, 0.78 (0.63–0.95) | 54 |
Per 10 nmol/L (dose–response) | RR, 0.92 (0.89–0.95) | NR | ||||
Lung | ||||||
Feng, 2017(
| 9 (6 PC, 3 CC) | 111,148 | 1511 |
Variable group comparisons Per 10 nmol/L (dose–response) |
RR, 0.84 (0.74–0.95) RR, 0.92 (0.87–0.96) |
50 NR |
Non‐Hodgkin lymphoma | ||||||
Purdue, 2010(
|
Males: 6 PC Females: 4 PC |
923 923 |
733 733 |
>100 vs. 50–75 nmol/L >100 vs. 50–75 nmol/L |
OR, 0.67 (0.37–1.20) OR, 0.81 (0.39–1.69) |
NR NR |
Ovarian | ||||||
Yin, 2011(
| 10 NCC | 3373 | 884 | Per 20 ng/mL | 0.83 (0.63–1.08) | 0 |
Pancreatic | ||||||
Stolzenberg‐Solomon, 2010(
| 6 NCC | 833 | 345 | ≥100 vs. 50–75 nmol/L | OR, 2.14 (0.93–4.92) |
|
Thyroid | ||||||
Hu, 2018(
| 9 (7 CC, 1 CS, 1 RSP) | 7099 | 1172 | <20 vs. ≥20 ng/mL | OR, 1.42 (1.17–1.73) | 27 |
7 (5 CC, 2 CS) | 6498 | 775 | Cases vs. controls | SMD, –0.20 (−0.36 to −0.03) | 55 | |
Zhao, 2019(
| 6 CC | 6241 | 711 | Deficient vs. non‐deficient | OR, 1.30 (1.00–1.69) | 38 |
12 CC | 7278 | 1239 | Cases vs. controls | SMD, –0.37 (−0.45 to –0.28) | 93 | |
Cancer mortality | ||||||
All | Patients with | |||||
Li, 2014(
|
breast cancer: 4 PC CRC: 3 (2 PC, 1 NC) lymphoma: 7 PC |
4813 1558 1234 |
661 883 511 |
Highest vs. lowest quartile Highest vs. lowest quartile Highest vs. lowest quartile |
HR, 0.65 (0.44–0.98) HR, 0.65 (0.47–0.88) HR, 0.50 (0.36–0.68) |
45 6 0 |
Han, 2019(
| 16 PC | 101,794 | 8729 |
Highest vs. lowest group Per 20 nmol/L (dose–response) |
RR, 0.81 (0.71–0.93) RR, 0.98 (0.97–0.99) |
49 NR |
Breast | ||||||
Kim, 2014(
|
4 PC 3 PC |
400 NR |
4556 NR |
Highest vs. lowest group Per 10 ng/mL (dose response) |
RR, 0.58 (0.40–0.85) RR, 0.88 (0.79–0.98) |
27 23 |
Maalmi, 2014(
| 3 PC | 2636 | 194 | High vs. low group | HR, 0.57 (0.38–0.84) | 17 |
Colorectal | ||||||
Maalmi, 2014(
| 3 PC | 1558 | 566 | High vs. low group | HR, 0.65 (0.49–0.86) | 0 |
Xu, 2018(
| 5 (3 PC, 2 NCC) | 4126 | 982 | High vs. low group | HR, 0.73 (0.55–0.97) | 69 |
Huang, 2019(
| 12 PC | 53,910 | 2021 | High vs. low group | HR, 0.64 (0.56–0.73) | 3 |
DerSimonian‐Laird Q statistic.
CC = Case control; CRC = colorectal cancer; CS = cross sectional; HR = hazard ratio; MR = Mendelian randomization; NCC = nested case–control; NR = not reported; NS = not significant; OR = odds ratio; PC = prospective cohort; POR = Peto odds ratio; RR = relative risk; RSP = retrospective; SMD = standardized mean difference.
With total cancer incidence as the outcome, one of these meta‐analyses combined data from eight prospective cohort studies (70,018 participants and 7511 events).(
There were numerous meta‐analyses (43 from 32 articles) on site‐specific cancer outcomes: breast (
A recent meta‐analysis had total cancer mortality as the outcome.(
Four meta‐analysis articles of prospective studies focused on patients with cancer.(
We identified eight meta‐analyses of clinical trials that evaluated the impact of vitamin D supplementation on cancer outcomes (incidence and mortality; Table
Meta‐Analyses of Intervention Studies
Study | Studies ( | Sample size ( | Events ( | Pooled RR effect (95% CI) | NNT (95% CI) | I2 (%) | Quality of meta‐analysis (AMSTAR‐2 rating) | Strength of association (umbrella review class) | GRADE credibility of evidence |
---|---|---|---|---|---|---|---|---|---|
Cancer incidence | |||||||||
All | |||||||||
Bolland, 2014(
| 7 | 48,167 | 3979 | 0.99 (0.93–1.05) | — | 0 | Critically low | NS | Moderate |
Bjelakovic, 2014(
| 14 | 49,891 | 3851 | 1.00 (0.94–1.06) | — | 0 | High | NS | Moderate |
Keum, 2014(
| 4 | 45,151 | 4333 | 1.00 (0.94–1.06) | — | 0 | Critically low | NS | High |
Goulão, 2018(
| 24 | 18,440 | 1061 | 1.03 (0.91–1.15) | — | 0 | Critically low | NS | High |
Haykal, 2019(
| 9 | 42,773 | 3022 | 0.96 (0.86–1.07) | — | 31 | Critically low | NS | High |
Keum, 2019(
| 10 | 83,353 | 6537 | 0.98 (0.93–1.03) | — | 0 | Critically low | NS | High |
Breast | |||||||||
Sperati, 2013(
| 2 | 5372 | 91 | 1.11 (0.74–1.68) | — | 0 | Moderate | NS | Low |
Bjelakovic, 2014(
| 7 | 43,669 | 1135 | 0.97 (0.86–1.09) | — | 0 | High | NS | Moderate |
Colorectal | |||||||||
Bjelakovic, 2014(
| 5 | 45,598 | 436 | 1.11 (0.92–1.34) | — | 0 | High | NS | Moderate |
Lung | |||||||||
Bjelakovic, 2014(
| 5 | 45,509 | 329 | 0.86 (0.69–1.07) | — | 0 | High | NS | Moderate |
Pancreatic | |||||||||
Bjelakovic, 2014(
| 2 | 36,405 | 69 | 0.91 (0.57–1.46) | — | 0 | High | NS | Moderate |
Cancer mortality | |||||||||
All | |||||||||
Bjelakovic, 2014(
| 4 | 44,492 | 1192 | 0.88 (0.78–0.98) | 292 (159–1751) | 0 | High | Weak | High |
Keum, 2014(
| 3 | 44,290 | 1190 | 0.88 (0.78–0.98) | 86 (47–515) | 0 | Critically low | Weak | High |
Goulão, 2018(
| 7 | 11,202 | 320 | 0.88 (0.70–1.09) | — | 0 | Critically low | NS | Moderate |
Haykal, 2019(
| 5 | 70,547 | 1533 | 0.87 (0.79–0.96) | 381 (236–1238) | 0 | Low | Weak | High |
Keum, 2019(
| 5 | 75,239 | 1591 | 0.87 (0.79–0.96) | 294 (182–957) | 0 | Critically low | Weak | High |
Zhang, 2019(
| 12 | 45,578 | 939 | 0.84 (0.74–0.95) | 279 (171–892) | 0 | Moderate | Weak | High |
All were randomized controlled trials.
AMSTAR = A Measurement Tool to Assess Systematic Reviews; GRADE = Grading of Recommendations, Assessment, Development, and Evaluation; NNT = number needed to treat; NS = nonsignificant association; RR = relative risk.
Study Selection Criteria of Meta‐Analyses of Intervention Studies
Meta‐analysis | Inclusion criteria | Exclusion criteria |
---|---|---|
Sperati, 2013(
|
1. Compared with placebo/no treatment 2. Vitamin D as single agent 3. Combined regimens including supplements & lifestyle modifications if used equally in all groups | Pregnant or lactating women |
Bjelakovic, 2014(
|
1. RCTs, irrespective of blinding, publication, status, or language. 2. Any dose, duration, and route of administration 3. Monotherapy or in combination with calcium 4. Concomitant interventions if used equally in all intervention groups |
1. Secondary induced osteoporosis (eg, glucocorticoid‐induced osteoporosis, thyroidectomy, primary hyperparathyroidism, chronic kidney disease, liver cirrhosis, Crohn disease, gastrointestinal bypass surgery) 2. Pregnant or lactating women 3. People with cancer |
Bolland, 2014(
| Cholecalciferol or ergocalciferol |
1. Cluster randomized trials 2. Trials of hydroxylated vitamin D or vitamin D analogues 3. Other interventions only in vitamin D group 4. Trials of fortified dairy products 5. Chronic comorbidity other than osteoporosis or frailty |
Keum, 2014(
| With or without calcium supplementation |
1. Non‐English articles 2. Abstracts & unpublished reports |
Goulão, 2018(
|
1. Mean or median age of ≥60 years 2. Follow‐up ≤1 year 3. Any vitamin D or vitamin D analog 4. Coadministration of other medications (eg, calcium) if the comparator group received the same medication 5. All languages |
1. Renal impairment, steroid‐induced osteoporosis, or psoriasis 2. Nonmelanoma skin cancers not counted as events |
Haykal, 2019(
|
1. Primary prevention 2. Vitamin D compared with placebo 3. Vitamin D for ≥3 years | |
Keum, 2019(
| Cholecalciferol or ergocalciferol, with or without other nutrients |
1. Number of outcomes ≤10 2. Follow‐up ≤1 year |
Zhang, 2019(
|
1. Age ≥ 18 years 2. Any health conditions 3. Vitamin D (any dose) vs. placebo or no treatment 4. Concomitant agents had to be same dose in all groups |
1. Case reports, case series, observational studies 2. All participants received vitamin D 3. Pregnant or lactating women 4. Critically patients 5. Hydroxylated vitamin D or vitamin D analogues |
All were randomized controlled trials (RCTs).
The AMSTAR‐2 ratings of the eight meta‐analyses are summarized in Table
AMSTAR‐2 Ratings of Meta‐Analyses of Intervention Studies on the Effect of Vitamin D Supplementation on Cancer Outcomes
AMSTAR‐2 item | First author, publication year (citation) | ||||||||
---|---|---|---|---|---|---|---|---|---|
Item | Description | Bolland, 2014(
| Sperati, 2013(
| Bjelakovic, 2014(
| Keum, 2014(
| Goulão, 2018(
| Haykal, 2019(
| Keum, 2019(
| Zhang, 2019(
|
1 | Did the research questions and inclusion criteria include the components of PICO? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
2 | Did the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol? | × | ✓ | ✓ | × | × | × | × | ✓ |
3 | Did the review authors explain their selection of the study designs for inclusion in the review? | × | × | × | ✓ | × | × | ✓ | × |
4 | Did the review authors use a comprehensive literature search strategy? | × | ✓ | ✓ | × | ✓ | Partial ✓ | Partial ✓ | ✓ |
5 | Did the review authors perform study selection in duplicate? | × | ✓ | ✓ | ✓ | × | ✓ | ✓ | ✓ |
6 | Did the review authors perform data extraction in duplicate? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
7 | Did the review authors provide a list of excluded studies and justify the exclusions? | ✓ | ✓ | ✓ | × | × | × | × | ✓ |
8 | Did the review authors describe the included studies in adequate detail? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | Partial ✓ |
9 | Did the review authors use a satisfactory technique to assess the RoB in studies that were included in the review? | × | ✓ | ✓ | × | ✓ | ✓ | × | ✓ |
10 | Did the review authors report on the sources of funding for the studies included in the review? | × | × | ✓ | × | × | × | × | × |
11 | If meta‐analysis was performed did the review authors use appropriate methods for statistical combination of results? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
12 | If meta‐analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta‐analysis or other evidence synthesis? | × | ✓ | ✓ | × | ✓ | ✓ | ✓ | ✓ |
13 | Did the review authors account for RoB in individual studies when interpreting/discussing the results of the review? | × | ✓ | ✓ | × | ✓ | ✓ | × | ✓ |
14 | Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
15 | If they performed quantitative synthesis did the review authors investigate publication bias and discuss its likely impact on the results? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
16 | Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review? | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Rating of overall confidence in the results of the review | CL | Moderate | High | CL | CL | CL | CL | Moderate |
Critical domains.
AMSTAR = A Measurement Tool to Assess Systematic Reviews; CL = critically low; PICO components: P = patient population/problem, I = intervention (issue of interest, considered for implementation), C = comparison (comparitor or current practice), O = outcome (how this is measured); RoB = risk of bias.
Six of these meta‐analyses had total cancer incidence as an outcome.(
Two of these meta‐analyses had specific cancer incidence as an outcome.(
Six of these meta‐analyses had cancer mortality as an outcome.(
Five intervention meta‐analyses of total cancer mortality had weak strength of association according to umbrella review criteria,(
Umbrella Review Assessment of Meta‐Analyses of Intervention Studies on the Effect of Vitamin D Supplementation on Cancer Outcomes
Reference | RR Effect (95% CI) | >1000 events (cases) | Significant summary associations per random‐effects calculations | No evidence of small‐study effects | No evidence of excess of significance | Prediction intervals excluding null value | Largest study nominally significant ( | Not large heterogeneity (I2 <50%) | Umbrella review class | ||
---|---|---|---|---|---|---|---|---|---|---|---|
| <10−3 |
| |||||||||
All cancers | |||||||||||
Bolland, 2014(
| 0.99 (0.93–1.05) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Bjelakovic, 2014(
| 1.00 (0.94–1.06) | ✓ | × | × | × | × | ✓ | × | × | ✓ | NS association |
Keum, 2014(
| 1.00 (0.94–1.06) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Goulão, 2018(
| 1.03 (0.91–1.15) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Haykal, 2019(
| 0.96 (0.86–1.07) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Keum, 2019(
| 0.98 (0.93–1.03) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Breast cancer | |||||||||||
Sperati, 2013(
| 1.11 (0.74–1.68) | × | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Bjelakovic, 2014(
| 0.97 (0.86–1.09) | ✓ | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Colorectal cancer | |||||||||||
Bjelakovic, 2014(
| 1.11 (0.92–1.34) | × | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Lung cancer | |||||||||||
Bjelakovic, 2014(
| 0.86 (0.69–1.07) | × | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Pancreatic cancer | |||||||||||
Bjelakovic, 2014(
| 0.91 (0.57–1.46) | × | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Total cancer mortality | |||||||||||
Bjelakovic, 2014(
| 0.88 (0.78–0.98) | ✓ | × | × | ✓ | × | ✓ | ✓ | × | ✓ | Weak (class IV) |
Keum, 2014(
| 0.88 (0.78–0.98) | ✓ | × | × | ✓ | ✓ | ✓ | ✓ | × | ✓ | Weak (class IV) |
Goulão, 2018(
| 0.88 (0.70–1.09) | × | × | × | × | ✓ | ✓ | × | × | ✓ | NS association |
Haykal, 2019(
| 0.87 (0.79–1.06) | × | × | × | ✓ | ✓ | ✓ | ✓ | × | ✓ | Weak (class IV) |
Keum, 2019(
| 0.87 (0.79–0.96) | ✓ | × | × | ✓ | ✓ | ✓ | ✓ | × | ✓ | Weak (class IV) |
Zhang, 2019(
| 0.84 (0.74–0.95) | × | × | × | ✓ | ✓ | ✓ | ✓ | × | ✓ | Weak (class IV) |
NS = nonsignificant; RCT = randomized controlled trial; RR = relative risk.
GRADE Summary of Findings for Meta‐Analyses of Intervention Studies on the Effect of Vitamin D Supplementation on Cancer Outcomes
Reference | Studies ( | Study design | Risk of bias | Imprecision | Inconsistency | Indirectness | Publication bias | RR Effect (95% CI) | Certainty (GRADE) |
---|---|---|---|---|---|---|---|---|---|
All cancers | |||||||||
Bolland, 2014(
| 7 | RCT | Not serious | Not serious | Not serious | Not serious | Possible: Egger's | 0.99 (0.93–1.05) | +++ Moderate |
Bjelakovic, 2014(
| 14 | RCT | Not serious | Not serious | Not serious | Not serious | Likely: Egger's | 1.00 (0.94–1.06) | +++ Moderate |
Keum, 2014(
| 4 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 1.00 (0.94–1.06) | ++++ High |
Goulão, 2018(
| 24 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 1.03 (0.91–1.15) | ++++ High |
Haykal, 2019(
| 9 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.96 (0.86–1.07) | ++++ High |
Keum, 2019(
| 10 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.98 (0.93–1.03) | ++++ High |
Breast cancer | |||||||||
Sperati, 2013(
| 2 | RCT | Not serious | Serious: wide CI from benefit to appreciable harm | Not serious | Not serious | Unlikely | 1.11 (0.74–1.68) | ++ Low |
Bjelakovic, 2014(
| 7 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.97 (0.86–1.09) | +++ Moderate |
Lung cancer | |||||||||
Bjelakovic, 2014(
| 5 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.86 (0.69–1.07) | +++ Moderate |
Colorectal cancer | |||||||||
Bjelakovic, 2014(
| 5 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 1.11 (0.92–1.34) | +++ Moderate |
Pancreatic cancer | |||||||||
Bjelakovic, 2014(
| 2 | RCT | Not serious | Serious: wide CI from appreciable benefit to appreciable harm | Not serious | Not serious | Too few studies to assess | 0.91 (0.57–1.46) | +++ Moderate |
Total cancer mortality | |||||||||
Bjelakovic, 2014(
| 4 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.88 (0.78–0.98) | ++++ High |
Keum, 2014(
| 3 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.88 (0.78–0.98) | ++++ High |
Goulão, 2018(
| 7 | RCT | Not serious | Serious: wide CI from appreciable benefit to small harm | Not serious | Not serious | Unlikely | 0.88 (0.70–1.09) | +++ Moderate |
Haykal, 2019 ( | 5 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.87 (0.79–0.96) | ++++ High |
Keum, 2019(
| 5 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.87 (0.79–0.96) | ++++ High |
Zhang, 2019(
| 12 | RCT | Not serious | Not serious | Not serious | Not serious | Unlikely | 0.84 (0.74–0.95) | ++++ High |
RCT = randomized controlled trial; RR = relative risk.
As pooled effects across meta‐analyses on the same outcome were mostly similar (with overlapping 95% CIs; Table
Owing to several limitations, the meta‐analyses and their primary studies have boundaries of applicability. These limitations influence the strength of associations and are key targets for future research. We discuss these for observational and intervention studies separately.
A limitation of the observational studies we reviewed is that, although they adjusted for multiple confounders, they are vulnerable to residual confounding; more so when the confounders (eg, smoking behavior, BMI, physical activity, diet) are measured less well. Adding to this issue is that vitamin D status is related to multiple diseases, besides cancer.(
An important issue for intervention studies is the growing randomized controlled trial (RCT) evidence that health benefits of vitamin D supplementation are greatest in vitamin D‐deficient people.(
Insufficient vitamin D dosage may be another limitation. For example, in the Goulão and colleagues meta‐analysis of cancer incidence,(
Third, as carcinogenesis is a long‐term and gradual process (often spanning decades), the need for a long follow‐up period is particularly great.(
A fourth limitation is that, as shown in the meta‐analyses for observational studies, 25OHD was more consistently associated with colorectal and breast cancers than other cancers (Table
Fifth, the vast majority of participants in the intervention studies were White, which restricts applicability of findings to non‐White populations.
Finally, there were quality‐related shortcomings of the meta‐analyses of intervention studies we reviewed. The most common ones, detected by AMSTAR‐2, were the lack of (i) information on funding sources of primary studies (noteworthy as many vitamin D trials are industry‐funded and have a high risk of “for‐profit” bias(
A potential limitation of this review is that, although a comprehensive and systematic literature search was performed, we may have missed some meta‐analyses. Second, our study was a meta‐review, and although this provides an overarching perspective on a research topic, we did not provide granulate analyses at the primary study level. Third, our review focused on meta‐analyses; thus, some primary studies may not have been included either because the meta‐analysis did not identify them or they were too recent to be included. Finally, we did not critically appraise the quality of all primary studies individually. This should have been done in each meta‐analysis; doing this here was beyond the scope of our review.
Several areas of future research would strengthen our understanding of vitamin D effects on cancer. First, given the emerging evidence for threshold effects related to vitamin D status, future trials should aim to recruit participants with vitamin D insufficiency (25OHD <50 nmol/L). There are major logistical and practical barriers to doing this in populations that are vitamin D replete, and trials could be undertaken more easily and cheaply in populations with a high prevalence of vitamin D insufficiency. These trials should have longer follow‐up periods, include more adequately powered data on site‐specific cancers (feasible for common cancers), and study more non‐White populations. However, ethical issues can arise with the conduct of long‐term trials in vitamin D‐deficient participants, as 50% will be randomly assigned to placebo and remain deficient for a prolonged period.
Second, cells that express the cell‐surface receptor proteins megalin and cubulin (eg, those in the kidney, lung, thyroid, mammary gland, gall bladder, and thyroid) can internalize 25OHD bound to vitamin D‐binding protein, with subsequent unbinding of 25OHD intracellularly and conversion to 1,25‐dihydroxyvitamin D, which can exert anticancer effects by activating the vitamin D receptor.(
Third, articles that reported inverse, longitudinal associations between 25OHD and cancer mortality(
Fourth, cancer incidence and mortality, and overall survival, though considered the gold standard endpoint in oncology trials, require a large sample size and long follow‐up time to achieve adequate statistical power.(
Finally, a novel area of research is investigating whether vitamin D pathway genes may alter health effects on vitamin D supplementation. A meta‐analysis of eight prospective studies found that colorectal cancer risk was lower in participants with the BB genotype of the
Observational studies showed that, in many cases, low vitamin D was inversely associated with cancer outcomes. For this, the associations for some outcomes, particularly colorectal cancer, seem to fulfill some (but not all) of Bradford‐Hill's criteria for causation,(
To our knowledge, this review is the first report to systematically compile and appraise clinical evidence—by concurrently using AMSTAR‐2, umbrella review, and GRADE assessment tools—of vitamin D supplementation in relation to cancer outcomes from meta‐analyses. We found highly credible RCT evidence that vitamin D supplementation reduces risk of total cancer mortality, but the magnitude of effect was classified as weak. Our finding of a highly credible weak effect on total cancer mortality is in line with that of a 2017 systematic review of meta‐analyses, which reported that vitamin D supplementation reduces risk of all‐cancer mortality.(
The available research, however, is not without limitations. To address these limitations and to provide clearer and further insight into the role of vitamin D in cancer incidence and related mortality, future research should include trials with more vitamin D‐insufficient participants and of longer follow‐up duration, plus adequately powered data on site‐specific cancers (where feasible).
The authors have no conflicts of interest.
The peer review history for this article is available at
The Health Research Council of New Zealand (HRC) supported JDS with a fellowship.
Authors' roles: JDS conducted the literature search, created the tables, and drafted the manuscript. RS verified the AMSTAR‐2 and GRADE assessments. JEM and RS edited and revised the manuscript. HRC had no role in the design, analysis, interpretation, or presentation of the results.
All authors approved the final version of the manuscript.