Vitamin D supplementation for prevention of cancer in adults
Abstract
Background
The evidence on whether vitamin D supplementation is effective in decreasing cancers is contradictory.
Objectives
To assess the beneficial and harmful effects of vitamin D supplementation for prevention of cancer in adults.
Search methods
We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, LILACS, Science Citation Index Expanded, and the Conference Proceedings Citation Index‐Science to February 2014. We scanned bibliographies of relevant publications and asked experts and pharmaceutical companies for additional trials.
Selection criteria
We included randomised trials that compared vitamin D at any dose, duration, and route of administration versus placebo or no intervention in adults who were healthy or were recruited among the general population, or diagnosed with a specific disease. Vitamin D could have been administered as supplemental vitamin D (vitamin D₃ (cholecalciferol) or vitamin D₂ (ergocalciferol)), or an active form of vitamin D (1α‐hydroxyvitamin D (alfacalcidol), or 1,25‐dihydroxyvitamin D (calcitriol)).
Data collection and analysis
Two review authors extracted data independently. We conducted random‐effects and fixed‐effect model meta‐analyses. For dichotomous outcomes, we calculated the risk ratios (RRs). We considered risk of bias in order to assess the risk of systematic errors. We conducted trial sequential analyses to assess the risk of random errors.
Main results
Eighteen randomised trials with 50,623 participants provided data for the analyses. All trials came from high‐income countries. Most of the trials had a high risk of bias, mainly for‐profit bias. Most trials included elderly community‐dwelling women (aged 47 to 97 years). Vitamin D was administered for a weighted mean of six years. Fourteen trials tested vitamin D₃, one trial tested vitamin D₂, and three trials tested calcitriol supplementation. Cancer occurrence was observed in 1927/25,275 (7.6%) recipients of vitamin D versus 1943/25,348 (7.7%) recipients of control interventions (RR 1.00 (95% confidence interval (CI) 0.94 to 1.06); P = 0.88; I² = 0%; 18 trials; 50,623 participants; moderate quality evidence according to the GRADE instrument). Trial sequential analysis (TSA) of the 18 vitamin D trials shows that the futility area is reached after the 10th trial, allowing us to conclude that a possible intervention effect, if any, is lower than a 5% relative risk reduction. We did not observe substantial differences in the effect of vitamin D on cancer in subgroup analyses of trials at low risk of bias compared to trials at high risk of bias; of trials with no risk of for‐profit bias compared to trials with risk of for‐profit bias; of trials assessing primary prevention compared to trials assessing secondary prevention; of trials including participants with vitamin D levels below 20 ng/mL at entry compared to trials including participants with vitamin D levels of 20 ng/mL or more at entry; or of trials using concomitant calcium supplementation compared to trials without calcium. Vitamin D decreased all‐cause mortality (1854/24,846 (7.5%) versus 2007/25,020 (8.0%); RR 0.93 (95% CI 0.88 to 0.98); P = 0.009; I² = 0%; 15 trials; 49,866 participants; moderate quality evidence), but TSA indicates that this finding could be due to random errors. Cancer occurrence was observed in 1918/24,908 (7.7%) recipients of vitamin D₃ versus 1933/24,983 (7.7%) in recipients of control interventions (RR 1.00 (95% CI 0.94 to 1.06); P = 0.88; I² = 0%; 14 trials; 49,891 participants; moderate quality evidence). TSA of the vitamin D₃ trials shows that the futility area is reached after the 10th trial, allowing us to conclude that a possible intervention effect, if any, is lower than a 5% relative risk reduction. Vitamin D₃ decreased cancer mortality (558/22,286 (2.5%) versus 634/22,206 (2.8%); RR 0.88 (95% CI 0.78 to 0.98); P = 0.02; I² = 0%; 4 trials; 44,492 participants; low quality evidence), but TSA indicates that this finding could be due to random errors. Vitamin D₃ combined with calcium increased nephrolithiasis (RR 1.17 (95% CI 1.03 to 1.34); P = 0.02; I² = 0%; 3 trials; 42,753 participants; moderate quality evidence). TSA, however, indicates that this finding could be due to random errors. We did not find any data on health‐related quality of life or health economics in the randomised trials included in this review.
Authors' conclusions
There is currently no firm evidence that vitamin D supplementation decreases or increases cancer occurrence in predominantly elderly community‐dwelling women. Vitamin D₃ supplementation decreased cancer mortality and vitamin D supplementation decreased all‐cause mortality, but these estimates are at risk of type I errors due to the fact that too few participants were examined, and to risks of attrition bias originating from substantial dropout of participants. Combined vitamin D₃ and calcium supplements increased nephrolithiasis, whereas it remains unclear from the included trials whether vitamin D₃, calcium, or both were responsible for this effect. We need more trials on vitamin D supplementation, assessing the benefits and harms among younger participants, men, and people with low vitamin D status, and assessing longer duration of treatments as well as higher dosages of vitamin D. Follow‐up of all participants is necessary to reduce attrition bias.
PICOs
Plain language summary
Vitamin D supplementation for prevention of cancer in adults
Review question
Does vitamin D supplementation prevent cancer?
Background
The available evidence on vitamin D and cancer occurrence is intriguing but inconclusive. Many observational studies as well as randomised trials suggest that high vitamin D levels in the blood are related to reduced cancer occurrence. However, results of randomised trials testing the effect of vitamin D supplementation for cancer prevention are contradictory.
Study characteristics
The aim of this systematic review was to analyse the benefits and harms of the different forms of vitamin D especially on cancer occurrence. A total of 18 trials provided data for this review; 50,623 participants were randomly assigned to either vitamin D or placebo or no treatment. All trials were conducted in high‐income countries.
Key results
The age range of the participants was 47 to 97 years and on average 81% were women. The majority of the included participants did not have vitamin D deficiency. Vitamin D administration lasted on average six years and most trial investigators used vitamin D₃ (cholecalciferol). We did not find firm evidence that vitamin D supplementation decreases or increases cancer occurrence in predominantly elderly community‐dwelling women. We observed decreases in all‐cause mortality and cancer‐related mortality among the vitamin D/D₃ treated participants in comparison with the participants in the control groups. However, using trial sequential analysis, a statistical approach to reconfirm or question these findings, we conclude that these results could be due to random errors (play of chance). We also found evidence that combined vitamin D₃ and calcium supplements increased renal stone occurrence, but it remains unclear from the included trials whether vitamin D₃, calcium, or both were responsible for this effect. Moreover, these results could also be due to random errors (play of chance).
Quality of the evidence
A large number of the study participants left the trials before completion, and this raises concerns regarding the validity of the results. Most of the trials were judged not to be well and fairly conducted so that the results were likely to be biased (that is, possibly an overestimation of benefits and underestimation of harms).
Currentness of evidence
This evidence is up to date as of February 2014.
Authors' conclusions
Summary of findings
| Vitamin D versus placebo or no intervention for prevention of cancer in adults | ||||||
| Patient or population: healthy participants or recruited among the general population; individuals diagnosed with a specific disease in a stable phase or with vitamin D deficiency Settings: outpatients | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Control | Vitamin D versus placebo or no intervention | |||||
| Cancer occurrence Follow‐up: 0.5 to 7 years | Study population | RR 1.00 | 50623 | ⊕⊕⊕⊝ moderatea | Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. | |
| 77 per 1000 | 77 per 1000 | |||||
| Moderate | ||||||
| 28 per 1000 | 28 per 1000 | |||||
| Cancer occurrence in trials using vitamin D₃ (cholecalciferol) Follow‐up: 0.5 to 7 years | Study population | RR 1.00 | 49891 | ⊕⊕⊕⊝ moderatea | Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. | |
| 77 per 1000 | 77 per 1000 | |||||
| Moderate | ||||||
| 28 per 1000 | 28 per 1000 | |||||
| All‐cause mortality Follow‐up: 0.5 to 7 years | Study population | RR 0.93 | 49866 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 80 per 1000 | 75 per 1000 | |||||
| Moderate | ||||||
| 16 per 1000 | 15 per 1000 | |||||
| Cancer mortality in trials using vitamin D₃(cholecalciferol) Follow‐up: 5 to 7 years | Study population | RR 0.88 | 44492 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 29 per 1000 | 25 per 1000 | |||||
| Moderate | ||||||
| 37 per 1000 | 33 per 1000 | |||||
| Adverse events: nephrolithiasis in trials using vitamin D₃(cholecalciferol) combined with calcium Follow‐up: 0.5 to 7 years | Study population | RR 1.17 | 42753 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 18 per 1000 | 21 per 1000 | |||||
| Moderate | ||||||
| 1 per 1000 | 1 per 1000 | |||||
| Health‐related quality of life | See comment | Not investigated. | ||||
| Health economics | See comment | Not investigated. | ||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded by one level because of risk of attrition bias bDowngraded by two levels because of risk of attrition bias and imprecision | ||||||
Background
Vitamin D is a fat‐soluble vitamin, which maintains calcium and phosphorus homeostasis (Holick 2007; Horst 2005; Lips 2006). Vitamin D₃ (cholecalciferol) is synthesised in the skin from 7‐dehydrocholesterol during exposure to sunlight. Alternatively, vitamin D, in the form of either vitamin D₃ or D₂, can be obtained from dietary sources. Vitamin D₃ (cholecalciferol) is the predominant form of vitamin D in humans. Vitamin D₂ (ergocalciferol) is derived mainly from irradiated plants. Vitamin D, as either D₃ or D₂, does not have biological activity. It must be metabolised within the liver to 25‐hydroxycalciferol (calcidiol) and in the kidney to the biologically active form known as 1,25‐dihydroxycalciferol (calcitriol) (Holick 2007; Horst 2005; Lips 2006). Current interest in vitamin D is engendered by the hypothesis that it may prevent cancer and prolong life (Bjelakovic 2014; Davis 2007; Giovannucci 2005). The evidence on whether vitamin D is effective in decreasing cancers is contradictory.
Description of the condition
Vitamin D status is determined by the measurement of the serum 25‐hydroxyvitamin D level (Bischoff‐Ferrar 2009c; Dawson‐Hughes 2005; Lips 2004). There is controversy about the definition of optimal vitamin D status (Bouillon 2013; Hilger 2014). The Institute of Medicine recently recommended a target serum 25‐hydroxyvitamin D level of 20 ng/ml (50 nmol/L) (IOM 2011). Based upon the systematic review prepared by the Institute of Medicine there are insufficient data to determine the safe upper limit of serum 25‐hydroxyvitamin D level (IOM 2011). However, serum 25‐hydroxyvitamin D level concentrations above 50 ng/mL (125 nmol/L) were considered potentially harmful (IOM 2011). The International Osteoporosis Foundation and the Endocrine Society Task Force recommend a target serum 25‐hydroxyvitamin D level of 30 ng/ml (75 nmol/L) (Dawson‐Hughes 2010; Holick 2011). The worldwide prevalence of suboptimal vitamin D status is estimated to be high (Hilger 2014; Holick 2007; Lamberg‐Allardt 2006; Lips 2010: Van Schoor 2011; Zittermann 2003). The major causes of vitamin D deficiency are insufficient exposure to sunlight, decreased dietary intake, skin pigmentation, obesity, and advanced age (Lips 2006). Vitamin D deficiency in childhood results in rickets, while in adults it precipitates or exacerbates osteopenia and osteoporosis, and induces osteomalacia (Holick 2007). It has been speculated that vitamin D insufficiency is related to increased risks of cancer and cardiovascular diseases (Chiang 2013; Freedman 2007; Garland 2007; Giovannucci 2005; Gorham 2007; Michos 2008; Norman 2014; Schwartz 2007; Zittermann 2005; Zittermann 2006), the leading causes of death in middle‐ and high‐income countries (Mathers 2006). Vitamin D insufficiency might also be the consequence of a disease, but not the cause (Marshall 2008).
Description of the intervention
Vitamin D supplementation prevents osteoporosis, osteomalacia, and fractures (Holick 2007; Lamberg‐Allardt 2006). It has been postulated that vitamin D may have additional health effects beyond prevention of bone diseases (Bikle 2009; Sutton 2003). Ecologic studies have found that living at higher altitudes with lower exposure to sunlight is linked to increased cancer risk (Apperly 1941; Garland 1980). Most observational studies have associated increased vitamin D intake with decreased risk of cancer (Garland 2007; Gorham 2007; Schwartz 2007). Although high vitamin D status was connected with increased pancreatic cancer risk (Helzlsouer 2010; Stolzenberg‐Solomon 2006; Stolzenberg‐Solomon 2010), the range of 25‐hydroxyvitamin D levels associated with the cancer risk was below that considered to reflect hypervitaminosis D (Krstic 2011).
How the intervention might work
The active form of vitamin D functions as a steroid‐like hormone (Horst 2005). The effects of vitamin D are mediated by its binding to a vitamin D receptor (Norman 2006; Wesley Pike 2005), which is present in most tissues and cells in the body (Lips 2006). Upon binding to its receptors, vitamin D may enhance cell differentiation and cell apoptosis, and may inhibit cell proliferation in a variety of cell types (Flynn 2006). In addition, vitamin D is required for a functional immune system which is important for an adequate physiological response to infections, inflammatory diseases, immune‐system‐mediated diseases, and cancer (Bikle 2009; Cutolo 2009; Hart 2011; Sutton 2003). Thus, vitamin D supplementation could reduce cancer development.
Adverse effects of the intervention
Vitamin D toxicity is the result of excessive vitamin D intake. There is sparse evidence that ingestion of high quantities of vitamin D is harmful. The majority of the trials that reported hypercalcaemia, hypercalciuria, or nephrocalcinosis were conducted in participants with renal failure (Cranney 2007). We have shown that vitamin D₃ combined with calcium may increase nephrolithiasis, and that alfacalcidol and calcitriol may increase hypercalcaemia (Bjelakovic 2014).
Why it is important to do this review
The available evidence on vitamin D and cancer occurrence is intriguing but inconclusive. Results of recently completed randomised trials testing the influence of vitamin D supplementation for cancer prevention are contradictory. Lappe 2007 found that vitamin D supplementation is associated with significantly decreased cancer incidence. In contrast, another large randomised trial found no effect of vitamin D and calcium supplementation on cancer incidence (Brunner 2011). Although Chung 2011 found inconclusive evidence regarding vitamin D supplementation for the prevention of cancer in a recently updated meta‐analysis, the same review suggests that hazard ratios and risk ratios are correlated better with the cancer mortality when compared to the cancer incidence rates (Krstic 2012). Our aim was to systematically review and statistically analyse the available evidence in order to assess the effects of vitamin D supplementation on cancer prevention in adults.
Objectives
To assess the beneficial and harmful effects of vitamin D supplementation for prevention of cancer in adults.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised trials, irrespective of blinding, publication status, or language.
Types of participants
Adult participants (aged 18 years or over) who were:
-
healthy or were recruited from the general population;
-
diagnosed with a specific disease in a stable phase;
-
diagnosed with vitamin D deficiency.
We excluded trials including:
-
people with secondary induced osteoporosis (e.g., glucocorticoid‐induced osteoporosis, thyroidectomy, primary hyperparathyroidism, chronic kidney disease, liver cirrhosis, Crohn's disease, gastrointestinal bypass surgery);
-
pregnant or lactating women (as they are usually in need of vitamin D);
-
people with cancer.
Types of interventions
We considered for inclusion randomised trials that compared vitamin D at any dose, duration, and route of administration versus placebo or no intervention.
The vitamin D was administered.
-
As monotherapy.
-
In combination with calcium.
Concomitant interventions were allowed if used equally in all intervention groups of the trial.
Types of outcome measures
Primary outcomes
-
Cancer occurrence.
-
All‐cause mortality.
-
Cancer mortality.
Secondary outcomes
-
Adverse events depending on the availability of data, we attempted to classify adverse events as serious or non‐serious. Serious adverse events were defined according to the International Conference on Harmonisation (ICH) Guidelines for Good Clinical Practice as any untoward medical occurrence that at any dose resulted in death, was life‐threatening, required inpatient hospitalisation or prolongation of existing hospitalisation, resulted in persistent or significant disability or incapacity, or was a congenital anomaly/birth defect, or any medical event, which might have jeopardised the patient, or required intervention to prevent it (ICH‐GCP 1997). All other adverse events (that is, any medical occurrence not necessarily having a causal relationship with the treatment, but causing a dose reduction or discontinuation of the treatment) were considered as non‐serious.
-
Health‐related quality of life.
-
Health economics.
Covariates, effect modifiers, and confounders
We noted and recorded any possible covariates, effect modifiers, and confounders (for example, compliance or other medications).
Timing of outcome measurement
We calculated the outcome effects at the end of the follow‐up period. We did not apply any restrictions regarding the length of intervention or the length of follow‐up.
Search methods for identification of studies
Electronic searches
We used the following sources from inception to the specified date for identification of the trials.
-
The Cochrane Central Register of Controlled Trials (CENTRAL) (until February 2014).
-
MEDLINE (until February 2014).
-
EMBASE (until February 2014).
-
LILACS (until February 2014).
-
Science Citation Index Expanded (until February 2014).
We also searched databases of ongoing trials (www.clinicaltrials.gov/ and www.controlled‐trials.com/ (with links to several databases) and the World Health Organization International Clinical Trials Registry Platform (ICTRP 2011)). For detailed search strategies, see Appendix 1. We included trials published in any language.
Searching other resources
We have contacted the main manufacturers of vitamin D to ask for unpublished randomised trials. We tried to identify additional trials by searching the reference lists of included trials and (systematic) reviews, meta‐analyses, and health technology assessment reports during the review preparation.
Data collection and analysis
Selection of studies
To determine the studies to be assessed further, two review authors (GB and DN) independently scanned the abstract, title, or both sections of every record retrieved. We investigated all potentially relevant articles as full text. Where differences in opinion existed, we resolved them by recourse to a third party (CG). If resolving disagreement was not possible, we added the article to those 'awaiting assessment' and contacted the authors for clarification. An adapted PRISMA (preferred reporting items for systematic reviews and meta‐analyses) flow‐chart of study selection is attached (Figure 1) (Liberati 2009).
Study flow diagram.
Data extraction and management
For studies that fulfilled our inclusion criteria, two review authors (GB and DN) independently abstracted relevant population and intervention characteristics using standard data extraction templates (for details, see 'Characteristics of included studies'; Table 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6), resolving any disagreements by discussion, or if required by recourse to a third party. We sought any relevant missing information on the trial from the study author(s) of the article, if required.
| Characteristic | Intervention(s) and comparator(s) | Screened/eligible | Randomised | ITT | Finishing study | Randomised finishing study |
| (1) Avenell 2012
| I1: vitamin D₃ | 15,024 | 1343 | 1343 | 1813 | 68 |
| I2: vitamin D₃ plus calcium | 1306 | 1306 | ||||
| C1: calcium | 1311 | 1311 | 1762 | 67 | ||
| C2: matched placebo tablets | 1332 | 1332 | ||||
| total: | 5292 | 5292 | 3575 | 68 | ||
| (2) Bolton‐Smith 2007
| I1: vitamin D₃ plus calcium | ‐ | 62 | 62 | 50 | 81 |
| C1: matched placebo | 61 | 61 | 56 | 92 | ||
| total: | 123 | 123 | 106 | 86 | ||
| (3) Brunner 2011
| I1: vitamin D₃ plus calcium | 68,132 | 18,176 | 18,176 | 16,936 | 93 |
| C1: matched placebo | 18,106 | 18,106 | 16,815 | 93 | ||
| total: | 36,282 | 36,282 | 33,751 | 93 | ||
| (4) Daly 2008
| I1: calcium‐vitamin D₃‐fortified milk plus calcium | 422 | 85 | 85 | 76 | 89 |
| C1: usual diet | 82 | 82 | 73 | 89 | ||
| total: | 167 | 167 | 149 | 89 | ||
| (5) Gallagher 2001
| I1: calcitriol | 1905 | 123 | 123 | 101 | 82 |
| C1: matched placebo | 123 | 123 | 112 | 91 | ||
| total: | 246 | 246 | 213 | 87 | ||
| (6) Glendenning 2012
| I1: cholecalciferol | 2110 | 353 | 353 | 331 | 94 |
| C1: placebo vitamin D | 333 | 333 | 307 | 92 | ||
| total: | 686 | 686 | 638 | 93 | ||
| (7) Grady 1991
| I1: calcitriol | 98 | 50 | 50 | 49 | 98 |
| C1: placebo vitamin D | 48 | 48 | 48 | 100 | ||
| total: | 98 | 50 | 97 | 99 | ||
| (8) Janssen 2010
| I1: vitamin D₃ plus calcium | 91 | 36 | 36 | 18 | 50 |
| C1:placebo vitamin D₃ plus calcium | 34 | 34 | 31 | 91 | ||
| total: | 70 | 70 | 49 | 70 | ||
| (9) Komulainen 1999
| I1: vitamin D₃ plus calcium | 13,100 | 116 | 116 | 112 | 97 |
| C1: placebo | 116 | 116 | 115 | 99 | ||
| total: | 232 | 232 | 227 | 98 | ||
| (10) Lappe 2007
| I1: vitamin D₃ plus calcium | 1180 | 446 | 446 | 403 | 90 |
| C1: vitamin D₃ placebo plus calcium | 445 | 445 | 416 | 93 | ||
| C2: vitamin D₃ placebo plus calcium placebo | 288 | 288 | 266 | 92 | ||
| total: | 1179 | 1179 | 1085 | 92 | ||
| (11) Larsen 2012 | I1: vitamin D₃ | 136 | 65 | 65 | 55 | 85 |
| C1: vitamin D placebo | 65 | 65 | 57 | 88 | ||
| total: | 130 | 130 | 112 | 86 | ||
| (12) Murdoch 2012 | I1: vitamin D₃ | 351 | 161 | 161 | 148 | 92 |
| C1: vitamin D placebo | 161 | 161 | 146 | 91 | ||
| total: | 322 | 322 | 294 | 91 | ||
| (13) Ott 1989
| I1: calcitriol plus calcium | ‐ | 43 | 43 | 39 | 91 |
| C1: placebo vitamin D plus calcium | 43 | 43 | 37 | 86 | ||
| total: | 86 | 86 | 76 | 88 | ||
| (14) Prince 2008
| I1: vitamin D₂ plus calcium | 827 | 151 | 151 | 144 | 95 |
| C1: placebo vitamin D plus calcium | 151 | 151 | 145 | 96 | ||
| total: | 302 | 302 | 289 | 95 | ||
| (15) Sanders 2010
| I1: vitamin D₃ | 7204 | 1131 | 1131 | 1015 | 90 |
| C1: vitamin D placebo | 1127 | 1127 | 1017 | 90 | ||
| total: | 2258 | 2258 | 2032 | 90 | ||
| (16) Trivedi 2003
| I1: vitamin D₃ | ‐ | 1345 | 1345 | 1262 | 94 |
| C1:placebo vitamin D | 1341 | 1341 | 1264 | 94 | ||
| total: | 2686 | 2686 | 2526 | 94 | ||
| (17) Witham 2013 | I1: vitamin D₃ | 341 | 80 | 80 | 73 | 91 |
| C1: placebo vitamin D | 79 | 79 | 69 | 87 | ||
| total: | 159 | 159 | 142 | 89 | ||
| (18) Wood 2012 | I1: vitamin D₃ | 424 | 102 | 102 | 84 | 82 |
| I2: vitamin D₃ | 101 | 101 | 90 | 89 | ||
| C1: placebo vitamin D | 102 | 102 | 91 | 89 | ||
| total: | 305 | 305 | 265 | 87 | ||
| Grand total | All interventions | 25,275 | 22,799 | 90 | ||
| All controls | 25,348 | 22,827 | 90 | |||
| All interventions and controls | 50,623 | 45,626 | 90 |
"‐" denotes not reported
ITT: intention‐to‐treat
Dealing with duplicate publications and companion papers
In the case of duplicate publications and companion papers of a primary study, we tried to maximise our yield of information by simultaneous evaluation of all available data.
Assessment of risk of bias in included studies
Due to the risk of overestimation of beneficial intervention effects in randomised trials of unclear or inadequate methodological quality (Kjaergard 2001; Lundh 2012; Moher 1998; Savović 2012a; Savović 2012b; Schulz 1995; Wood 2008), we assessed the influence of the risk of bias on our results. We used the following domains: allocation sequence generation, allocation concealment, blinding, incomplete outcome data reporting, selective outcome reporting, and other apparent biases (Higgins 2011). We used the following definitions:
Allocation sequence generation
-
Low risk of bias: sequence generation was achieved using computer random number generation or a random number table. Drawing lots, tossing a coin, shuffling cards, and throwing dice were adequate if performed by an independent person not otherwise involved in the trial.
-
Uncertain risk of bias: the method of sequence generation was not specified.
-
High risk of bias: the sequence generation method was not random.
Allocation concealment
-
Low risk of bias: the participant allocations could not have been foreseen in advance of, or during, enrolment. Allocation was controlled by a central and independent randomisation unit. The allocation sequence was unknown to the investigators (for example, if the allocation sequence was hidden in sequentially numbered, opaque, and sealed envelopes).
-
Uncertain risk of bias: the method used to conceal the allocation was not described so that intervention allocations may have been foreseen in advance of, or during, enrolment.
-
High risk of bias: the allocation sequence was likely to be known to the investigators who assigned the participants.
Blinding of participants, personnel, and outcome assessors
-
Low risk of bias: blinding was performed adequately, or the assessment of outcomes was not likely to be influenced by lack of blinding.
-
Uncertain risk of bias: there was insufficient information to assess whether blinding was likely to induce bias into the results.
-
High risk of bias: no blinding or incomplete blinding, and the assessment of outcomes were likely to be influenced by lack of blinding.
Incomplete outcome data
-
Low risk of bias: missing data were unlikely to make treatment effects depart from plausible values. Sufficient methods, such as multiple imputation, have been employed to handle missing data.
-
Uncertain risk of bias: there was insufficient information to assess whether missing data in combination with the method used to handle missing data were likely to induce bias into the results.
-
High risk of bias: the results were likely to be biased due to missing data.
Selective outcome reporting
-
Low risk of bias: all outcomes were predefined and reported, or all clinically relevant and reasonably expected outcomes were reported.
-
Uncertain risk of bias: it is unclear whether all predefined and clinically relevant and reasonably expected outcomes were reported.
-
High risk of bias: one or more clinically relevant and reasonably expected outcomes were not reported, and data on these outcomes were likely to have been recorded.
For a trial to be assessed with low risk of bias in the selective outcome reporting domain, the trial should have been registered either on the www.clinicaltrials.gov website or a similar register, or there should be a protocol, e.g., published in a paper journal. In the case of a trial being run and published in the years when trial registration was not required, we tried to carefully scrutinise the publication reporting on the trial to identify the trial objectives and outcomes. If usable data on all outcomes specified in the trial objectives were provided in the publications results section, then the trial was considered to be at low risk of bias in the selective outcome reporting domain.
For‐profit bias
-
Low risk of bias: the trial appeared to be free of industry sponsorship or other kind of for‐profit support that may manipulate the trial design, conduct, or results of the trial.
-
Uncertain risk of bias: the trial may or may not be free of for‐profit bias, as no information on clinical trial support or sponsorship is provided.
-
High risk of bias: the trial is sponsored by the industry or has received some other kind of for‐profit support.
Other bias
-
Low risk of bias: the trial appears to be free of other components (for example, academic bias) that could put it at risk of bias.
-
Uncertain risk of bias: the trial may or may not be free of other components that could put it at risk of bias.
-
High risk of bias: there are other factors in the trial that could put it at risk of bias (for example, authors have conducted trials on the same topic, etc).
We considered trials assessed as having 'low risk of bias' in all of the specified individual domains as being 'trials with low risk of bias'. We considered trials assessed as having 'uncertain risk of bias' or 'high risk of bias' in one or more of the specified individual domains as being trials with 'high risk of bias' (Gluud 2011).
Dealing with missing data
We tried to obtain relevant missing data from authors whenever we lacked important numerical data such as number of screened or randomised participants, or lack of data regarding the performance of intention‐to‐treat (ITT) analyses, or data on as‐treated or per‐protocol participant analyses in order to perform our analyses as rigorously as possible. We investigated attrition rates, (for example, dropouts, losses to follow‐up, and withdrawals) and critically appraised issues of missing data (for example, last‐observation‐carried‐forward (LOCF)) and imputation methods.
Regarding the primary outcomes, we included participants with incomplete or missing data in sensitivity analyses by imputing them according to the following scenarios (Hollis 1999).
-
Extreme case analysis favouring the experimental intervention ('best‐worst' case scenario: none of the dropouts/participants lost from the experimental arm, but all of the dropouts/participants lost from the control arm experienced the outcome, including all randomised participants in the denominator.
-
Extreme case analysis favouring the control ('worst‐best' case scenario): all dropouts/participants lost from the experimental arm, but none from the control arm experienced the outcome, including all randomised participants in the denominator.
Assessment of heterogeneity
We identified heterogeneity by visual inspection of the forest plots, by using a standard Chi² test and a significance level of α = 0.1, in view of the low power of such tests. We specifically examined heterogeneity with the I² statistic (Higgins 2002), where I² values of 50% or more indicated a substantial level of heterogeneity (Higgins 2003). For the heterogeneity adjustment of the required information size in the trials sequential analyses, we used diversity (D²), as I² used for this purpose consistently underestimates the required information size (Wetterslev 2009).
When we found heterogeneity, we attempted to determine the potential reasons for it by examining the individual trial and subgroup characteristics.
Assessment of reporting biases
We used funnel plots in an exploratory data analysis to assess the potential existence of bias in small trials. There are a number of explanations for the asymmetry of a funnel plot, including true heterogeneity of effect with respect to trial size, poor methodological design of small trials, and publication bias. We performed adjusted rank correlation (Begg 1994) and a regression asymmetry test (Egger 1997) for detection of bias. We considered a P value less than 0.10 significant in the latter analyses. We carefully interpreted results of funnel plots (Lau 2006).
Data synthesis
We performed statistical analyses according to the statistical guidelines referenced in the latest version of The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
For the statistical analyses, we used Review Manager 5 (RevMan 2012), Trial Sequential Analysis version 0.9 beta (TSA 2011; www.ctu.dk/tsa), STATA 8.2 (STATA Corp, College Station, Texas), and Sigma Stat 3.0 (SPSS Inc, Chicago, Ill). We analysed the data with both fixed‐effect (DeMets 1987) and random‐effects (DerSimonian 1986) model meta‐analyses. In the review text, we present the results of the random‐effects model analyses. For dichotomous outcomes, we calculated the Mantel‐Haenszel risk ratios (RRs). For all association measures, we used 95% confidence intervals (CIs). We conducted the analyses using the intention‐to‐treat principle, including all randomised participants irrespective of completeness of data. We included participants with missing data in the analyses using a carry forward of the last observed response. Accordingly, we counted participants who had been lost to follow‐up as being alive.
We calculated weighted averages for factors related to the trials, such as duration of the intervention and length of the follow‐up period.
For trials using a factorial design, we performed 'at the margins' analysis, combining all participants randomised to vitamin D (McAlister 2003). Due to the risk of interaction between different treatment regimens, we also performed 'inside the table' analysis in which we compared vitamin D only with placebo or no intervention. In the trials with a parallel group design with more than two arms and additional therapy, we compared the vitamin D arm alone with placebo or no intervention.
Trial sequential analysis
Meta‐analyses may result in type 1 errors due to sparse data and repeated significance testing when meta‐analyses are updated with new trials (Brok 2008; Brok 2009; Thorlund 2009; Wetterslev 2008; Wetterslev 2009). In a single trial, interim analysis increases the risk of type 1 errors. To avoid type 1 errors (rejecting the null hypothesis when it is in fact true) group sequential monitoring boundaries (Lan 1983) are applied to decide whether a trial could be terminated early because of a sufficiently small P value, that is, when the cumulative Z‐curve crosses the alpha‐spending monitoring boundary. Sequential monitoring boundaries can be applied to meta‐analyses as well and are called trial sequential monitoring boundaries. In 'trial sequential analysis' (TSA), the addition of each trial in a cumulative meta‐analysis is regarded as an interim meta‐analysis and helps to decide whether additional trials are needed (Wetterslev 2008). So far, several meta‐analyses and reviews have been published, including an increasing number of trial results as new trials have been published. It therefore seems appropriate to adjust new meta‐analyses for sparse accumulating data and multiple testing to control the overall type 1 error risk in a cumulative meta‐analysis (Pogue 1997; Pogue 1998; Thorlund 2009; Wetterslev 2008). The idea behind TSA is that if the cumulative Z‐curve crosses a trial sequential monitoring boundary (TSMB), a sufficient level of evidence is reached and no further trials may be needed. However, there is insufficient evidence to reach a conclusion if the cumulative Z‐curve does not cross the TSMB or does not surpass the futility boundaries before the required information size is reached. To construct the TSMBs, a required information size is needed which is calculated as the least number of participants needed in a well‐powered single trial (Brok 2008; Pogue 1998; Wetterslev 2008). We adjusted the required information size to account for statistical between‐trial heterogeneity with a diversity adjustment factor (Wetterslev 2009).
We performed trial sequential analysis to avoid random errors due to repetitive analyses of accumulated data and to prevent premature statements of superiority of intervention or lack of effect of an intervention (TSA 2011). We used the diversity‐adjusted required information size estimated from a control event proportion of the included trials and an a priori intervention effect of 5% and 10% relative risk reduction (RRR) (Wetterslev 2009), and the diversity which was estimated in the included trials.
Subgroup analysis and investigation of heterogeneity
We mainly performed subgroup analyses if one of the primary outcomes demonstrated statistically significant differences between the intervention groups. In any other case, subgroup analyses were clearly marked as a hypothesis‐generating exercise.
We conducted the following subgroup analyses.
-
Trials with a low risk of bias compared to trials with a high risk of bias.
-
Trials without risk of for‐profit bias compared to trials with risk of for‐profit bias
-
Primary prevention trials compared to secondary prevention trials.
-
Trials including participants with vitamin D insufficiency compared to trials with vitamin D adequacy.
-
Vitamin D₃ compared to placebo or no intervention.
-
Trials that administered vitamin D₃ singly compared to trials that administered vitamin D₃ combined with calcium.
-
Vitamin D₂ compared to placebo or no intervention.
-
Calcitriol compared to placebo or no intervention.
Sensitivity analysis
We performed the following sensitivity analyses to explore the influence of these imputations on the intervention effect size.
-
Best‐worst case scenario analyses.
-
Worst‐best case scenario analyses.
We also tested the robustness of the results by repeating the analysis using different statistical models (fixed‐ and random‐effects model meta‐analyses).
Summary of findings tables
We used GRADE (tech.cochrane.org/revman/other‐resources/gradepro) to construct a 'Summary of findings' table for vitamin D for all of the review outcomes. The GRADE approach appraises the quality of a body of evidence based on the extent to which one can be confident that an estimate of effect or association reflects the item being assessed. The quality of a body of evidence considers within‐study risk of bias, the directness of the evidence, heterogeneity of the data, precision of effect estimates, and risk of publication bias (Andrews 2013a; Andrews 2013b; Balshem 2011; Brunetti 2013; Guyatt 2011a; Guyatt 2011b; Guyatt 2011d; Guyatt 2011e; Guyatt 2011f; Guyatt 2011g; Guyatt 2011h; Guyatt 2011i; ; Guyatt 2013a; Guyatt 2013b; Guyatt 2013c; Mustafa 2013). We used results obtained withTSA for the rating for imprecision. If there is insufficient evidence to reach a conclusion, i.e., if the TSA indicates that the required information size had not been reached, we downgraded the quality of the evidence by one level. In addition, we used risk of attrition bias for the rating for imprecision. If there was significant risk of attrition bias, we further downgraded the quality of the evidence by one level.
Results
Description of studies
Results of the search
We identified a total of 6211 references of possible interest through searching The Cochrane Library (n = 1208), MEDLINE (n = 1291), EMBASE (n = 1894), LILACS (n = 525), Science Citation Index Expanded (n = 82), Conference Proceedings Citation Index‐Science (n = 1201), and reference lists (n = 43). We identified an additional 10 ongoing trials through searching databases of ongoing trials. We will include data from the ongoing trials in future updates of this review. We then excluded 5190 duplicates and 960 clearly irrelevant references through reading the abstracts. Accordingly, we retrieved 103 references for further assessment. Of these, we excluded 20 references describing 18 studies because they were not randomised trials or did not fulfil the inclusion criteria of our review. We list reasons for exclusion in the table Characteristics of excluded studies. In total, 18 randomised trials described in 74 references fulfilled the inclusion criteria (Figure 1). The trials included a total of 50,623 participants.
We contacted eight authors for missing information and received answers from all of them.
Included studies
A detailed description of the characteristics of included studies is presented elsewhere (see Characteristics of included studies and appendices). The following is a succinct overview:
Trial characteristics
Out of the 18 trials reporting cancer occurrence, 14 trials used a parallel‐group design and four trials (Avenell 2012; Bolton‐Smith 2007; Gallagher 2001; Komulainen 1999) used a 2‐by‐2 factorial design (Pocock 2004). The trials were published between 1989 and 2013.
In 16 trials, vitamin D was provided free of charge by pharmaceutical companies. Two trials were not funded by industry (Trivedi 2003; Wood 2012).
The trials were conducted in Europe (n = 8) (Avenell 2012; Bolton‐Smith 2007; Janssen 2010; Komulainen 1999; Larsen 2012; Trivedi 2003; Witham 2013; Wood 2012), North America (n = 5) (Brunner 2011; Gallagher 2001; Grady 1991; Lappe 2007; Ott 1989), and Oceania (n = 5) (Daly 2008; Glendenning 2012; Murdoch 2012; Prince 2008; Sanders 2010). All 18 trials were conducted in high‐income countries.
Participants
A total of 50,623 participants were randomly assigned in the 18 trials reporting cancer occurrence (Table 1). The number of participants in each trial ranged from 70 to 36,282 (median 313). The mean age of participants was 69 years (range 47 to 97 years). The mean proportion of women was 81%.
Sixteen trials were primary prevention trials, that is, included healthy participants, or participants from the general population. Of these, 11 trials included elderly and postmenopausal women (Bolton‐Smith 2007; Brunner 2011; Gallagher 2001; Glendenning 2012; Janssen 2010; Komulainen 1999; Lappe 2007; Ott 1989; Prince 2008; Sanders 2010; Wood 2012), three trials included elderly people (Avenell 2012; Grady 1991; Trivedi 2003), and two trials included healthy volunteers (Daly 2008; Murdoch 2012).
Two trials were secondary prevention trials that included participants with cardiovascular disease (arterial hypertension) (Larsen 2012; Witham 2013).
Of the 18 trials reporting cancer occurrence, 16 (89%) reported the baseline vitamin D status of participants based on serum 25‐hydroxyvitamin D levels. Participants in nine trials (Bolton‐Smith 2007; Daly 2008; Gallagher 2001; Glendenning 2012; Grady 1991; Larsen 2012; Murdoch 2012; Ott 1989; Trivedi 2003) had baseline 25‐hydroxyvitamin D levels at or above vitamin D adequacy (20 ng/ml). Participants in the other seven trials had baseline 25‐hydroxyvitamin D levels considered vitamin D insufficient (< 20 ng/ml) (Avenell 2012; Brunner 2011; Janssen 2010; Prince 2008; Sanders 2010; Witham 2013; Wood 2012). Two trials did not report the baseline vitamin D status of participants (Komulainen 1999; Lappe 2007).
The main outcomes in the trials were cancer occurrence, all‐cause mortality, bone mineral density, and number of falls and fractures.
Experimental interventions
Vitamin D₃ ‐ cholecalciferol
Vitamin D was administered as vitamin D₃ (cholecalciferol) in 14 trials (49,891 participants; 78% women; mean age 67 years). Vitamin D₃ was tested singly in seven trials (Glendenning 2012; Murdoch 2012; Sanders 2010; Trivedi 2003; Witham 2013; Wood 2012) and combined with calcium in six trials. One trial tested vitamin D₃ singly and combined with calcium (Avenell 2012). Vitamin D₃ was administered orally in all trials. Vitamin D₃ was given daily in nine trials and at intervals in five trials (monthly (Murdoch 2012); three‐monthly (Glendenning 2012; Witham 2013); four‐monthly (Trivedi 2003); and yearly (Sanders 2010)). The daily dose of the vitamin D₃ was 300 IU to 3333 IU (mean daily dose 1146 IU; median daily dose 810 IU). The duration of supplementation in trials using vitamin D₃ was five months to seven years (weighted mean 6.0 years), and the duration of the follow‐up period was five months to seven years (weighted mean 6.3 years).
Vitamin D₂ ‐ ergocalciferol
Vitamin D was administered as vitamin D₂ (ergocalciferol) in one trial (302 participants; 100% women; mean age 77.2 years). Vitamin D₂ was tested in a dose of 1000 IU, combined with 1000 mg of calcium, orally, and daily for a one‐year period.
Calcitriol ‐ 1,25‐dihydroxyvitamin D
Vitamin D was administered as calcitriol in three trials (430 participants; 85% women; aged 50 to 97 years). Calcitriol was tested singly in two trials and combined with calcium in one trial (Ott 1989). Calcitriol was tested orally and daily in all trials. The dose of calcitriol was 0.5 μg in two trials (Gallagher 2001; Grady 1991); while two doses of calcitriol (0.5 μg and 2 μg) were tested in another trial (Ott 1989). The duration of supplementation in trials using calcitriol was two to five years (weighted mean 2.5 years) and the duration of the follow‐up period was two to five years (weighted mean 4.0 years).
Comparator interventions
Seventeen trials used placebo, and one trial used no intervention in the control group (Daly 2008).
Co‐interventions
Seven trials used calcium combined with vitamin D in the experimental intervention groups. Two trials tested calcium separately in one of the intervention groups (Avenell 2012; Lappe 2007). Calcium was administered orally and daily in all trials. The dose of calcium was 500 mg to 1500 mg (mean 883 mg; median 1000 mg).
Three trials used calcium in the control group, combined with vitamin D placebo, in a dose of 500 mg to 1000 mg (mean 833 mg; median 1000 mg). These trials used an equal dose of calcium in the experimental intervention groups. One trial with a 2‐by‐2 factorial design tested a combination of vitamin D₃, vitamin K₁, and calcium in one group (Bolton‐Smith 2007). The factorial design of this trial allowed us to compare only the vitamin D₃ plus calcium group with the placebo group of this trial. Two trials with a 2‐by‐2 factorial design tested vitamin D and hormone replacement (Gallagher 2001; Komulainen 1999). We have compared only the vitamin D group with the placebo group of these trials.
Excluded studies
A detailed description of the characteristics of included studies is presented elsewhere (see Characteristics of excluded studies and appendices).
Risk of bias in included studies
Two trials (11%) were considered to be at low risk of bias. The remaining 16 trials had unclear bias control in one or more of the components assessed (Figure 2; Figure 3). Inspection of the funnel plot does not suggest potential bias (asymmetry) (Figure 4). The adjusted‐rank correlation test (P = 1.00) found no significant evidence of bias, while the regression asymmetry test found significant evidence of bias (P = 0.007).
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Funnel plot of comparison: 1 Vitamin D versus placebo or no intervention, outcome: 1.1 Cancer occurrence in trials with a low or high risk of bias.
Allocation
The generation of the allocation sequence was adequately described in 16 trials. The remaining two trials (Grady 1991; Ott 1989) were described as randomised, but the method used for sequence generation was not described.
The method used to conceal allocation was adequately described in 15 trials. We judged the method used for allocation concealment to be unclear in two trials (Grady 1991; Ott 1989) and inadequate in one trial (Daly 2008).
Blinding
The method of blinding was adequately described in 14 trials. The method of blinding was unclear in three trials (Grady 1991; Lappe 2007; Ott 1989). One trial was not blinded (Daly 2008).
Incomplete outcome data
Incomplete data were addressed adequately in 17 trials. In one trial, information is insufficient to allow assessment of whether the missing data mechanism in combination with the method used to handle missing data is likely to introduce bias into the estimate of effect (Lappe 2007).
Selective reporting
Predefined primary and secondary outcomes were reported in all trials.
For‐profit bias
Two trials were not funded by industry (Trivedi 2003; Wood 2012). Fifteen trials were funded by industry (Avenell 2012; Bolton‐Smith 2007; Brunner 2011; Daly 2008; Gallagher 2001; Grady 1991; Komulainen 1999; Janssen 2010; Lappe 2007; Larsen 2012; Murdoch 2012; Ott 1989; Prince 2008; Sanders 2010; Witham 2013). The source of funding is not clear for one trial (Glendenning 2012).
Other potential sources of bias
All included trials appear to be free of other components that could put them at risk of bias.
Effects of interventions
Primary outcomes
Cancer occurrence
Overall, vitamin D had no statistically significant effect on cancer occurrence (risk ratio (RR) 1.00 (95% CI 0.94 to 1.06); P = 0.88; I² = 0%; 18 trials; 50,623 participants; Analysis 1.1). A total of 1927 of 25,275 participants (7.6%) randomised to the vitamin D group and 1943 of 25,348 participants (7.7%) randomised to the placebo or no intervention group had cancer at the end of follow‐up. Trial sequential analysis of all vitamin D trials suggests that we reached the futility area after the 10th trial, allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction (Figure 5).
Trial sequential analysis on cancer occurrence in the 18 vitamin D trials was performed based on cancer occurrence of 10% in the control group, a relative risk reduction of 5% with vitamin D supplementation, a type I error of 5%, and a type II error of 20% (80% power). There was no diversity. This resulted in a required information size of 110,505 participants. Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. The blue line represents the cumulative Z‐score of the meta‐analysis. The green lines represent the conventional statistical boundaries. The red inward sloping lines represent the trial sequential monitoring boundaries.
Intervention effects according to bias risk of trials
Vitamin D had no statistically significant effect on cancer occurrence in trials with a low risk of bias (RR 1.08 (95% CI 0.89 to 1.31); P = 0.41; I² = 0%; 2 trials; 2991 participants), nor in trials with a high risk of bias (RR 0.99 (95% CI 0.93 to 1.05); P = 0.67; I² = 0%; 16 trials; 47,632 participants; Analysis 1.1). The difference between the effect estimate of vitamin D on cancer in trials with low risk of bias and trials with a high risk of bias was not statistically significant (P = 0.36); (Analysis 1.1).
Trials without risk of for‐profit bias compared to trials with risk of for‐profit bias
Vitamin D had no significant effect on cancer occurrence in the trials without risk of for‐profit bias (RR 1.08 (95% CI 0.89 to 1.31); P = 0.41; I² = 0%; 2991 participants; 2 trials; Analysis 1.2). Vitamin D had no significant effect on cancer occurrence in the trials with risk of for‐profit bias (RR 0.99 (95% CI 0.93 to 1.05); P = 0.67; I² = 0%; 47,632 participants; 16 trials; Analysis 1.2). The difference between the estimate of the effect of vitamin D on cancer occurrence in the trials without risk of for‐profit bias and the trials with risk of for‐profit bias was not statistically significant by the test of interaction (P = 0.36; Analysis 1.2).
Primary prevention compared to secondary prevention
Vitamin D had no significant effect on cancer occurrence in the primary prevention trials (RR 1.00 (95% CI 0.94 to 1.06); P = 0.87; I² = 0%; 50,334 participants; 16 trials; Analysis 1.3). Vitamin D had no statistically significant effect on cancer occurrence in the secondary prevention trials (RR 1.33 (95% CI 0.26 to 6.96); P = 0.73; I² = 0%; 289 participants; 2 trials; Analysis 1.3). The difference between the estimates of the effect of vitamin D on cancer occurrence in the primary prevention and the secondary prevention trials was not statistically significant by the test of interaction (P = 0.73; Analysis 1.3).
Intervention effects according to vitamin D status
Vitamin D had no statistically significant effect on cancer occurrence in participants with vitamin D insufficiency (RR 0.99 (95% CI 0.93 to 1.05); P = 0.68; I² = 0%; 7 trials; 44,668 participants), nor in participants with vitamin D adequacy (RR 1.12 (95% CI 0.94 to 1.34); P = 0.21; I² = 0%; 9 trials; 4544 participants; Analysis 1.4). The difference between the estimates of vitamin D on cancer in trials including participants with vitamin D adequacy and trials including participants with vitamin D insufficiency was not statistically significant (P = 0.19; Analysis 1.4).
Sensitivity analyses taking attrition into consideration
Of the 18 trials reporting cancer occurrence, 17 trials reported the exact numbers of participants with missing outcomes in the intervention and control groups. One trial did not report losses to follow‐up for the intervention groups separately (Lappe 2007). A total of 849 of 24,829 participants (3.4%) had missing outcomes in the vitamin D group versus 791 of 24,615 participants (3.2%) in the control group.
'Best‐worst case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group had no cancer and all those with missing outcomes in the control intervention group developed cancer, vitamin D supplementation significantly decreased cancer occurrence (RR 0.41 (95% CI 0.31 to 0.54); P < 0.00001; I² = 82%; 49,444 participants; 17 trials; Analysis 1.5).
'Worst‐best case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group developed cancer and all those with missing outcomes in the control intervention group had no cancer, vitamin D supplementation significantly increased cancer occurrence (RR 2.76 (95% CI 1.97 to 3.86); P < 0.00001; I² = 88%; 49,444 participants; 17 trials; Analysis 1.5).
Intervention effects according to administered form of vitamin D
Vitamin D₃ (cholecalciferol)
Vitamin D₃ was tested in 14 trials (49,891 participants). Inspection of the funnel plot does not suggest potential bias (asymmetry). The adjusted‐rank correlation test found no significant evidence of bias (P = 0.59), while a regression asymmetry test found significant evidence of bias (P = 0.01). Overall, vitamin D₃ had no statistically significant effect on cancer (RR 1.00 (95% CI 0.94 to 1.06); P = 0.88; I² = 0%; Analysis 1.6). Vitamin D₃ had no statistically significant effect on cancer in trials with a low risk of bias (RR 1.08 (95% CI 0.89 to 1.31); P = 0.41; I² = 0%; 2 trials; 2991 participants), nor in trials with a high risk of bias (RR 0.99 (95% CI 0.93 to 1.05); P = 0.67; I² = 0%; 12 trials; 46,900 participants; Analysis 1.6). The difference between the estimates of vitamin D₃ on cancer in trials with low risk of bias versus trials with a high risk of bias was not statistically significant (P = 0.36).
Vitamin D₃ and calcium
Vitamin D₃ administered singly versus placebo had no statistically significant effect on cancer (RR 1.03 (95% CI 0.90 to 1.17); P = 0.69; I² = 0%; 9200 participants; 8 trials; Analysis 1.7). Vitamin D₃ combined with calcium versus placebo or no intervention had no statistically significant effect on cancer (RR 0.97 (95% CI 0.91 to 1.04); P = 0.36; I² = 0%; 40,670 participants; 7 trials; Analysis 1.7).
Cancer site occurrence in trials using vitamin D₃
Vitamin D₃ had no significant effect on lung cancer (RR 0.86 (95% CI 0.69 to 1.07); P = 0.17; I² = 0%; 5 trials; 45,509 participants; Analysis 1.8); breast cancer (RR 0.97 (95% CI 0.86 to 1.09); P = 0.61; I² = 0%; 7 trials; 43,669 participants; Analysis 1.9); colorectal cancer (RR 1.11 (95% CI 0.92 to 1.34); P = 0.26; I² = 0%; 5 trials; 45,598 participants; Analysis 1.10); or pancreatic cancer (RR 0.91 (95% CI 0.57 to 1.46); P = 0.69; I² = 0%; 2 trials; 36,405 participants; Analysis 1.11). Vitamin D₃ had no significant effect on prostate, uterine, ovarian, oesophageal, stomach, or liver cancer (one trial each) (Analysis 1.12; Analysis 1.13; Analysis 1.14; Analysis 1.15, Analysis 1.16; Analysis 1.17).
Vitamin D₂ (ergocalciferol)
Vitamin D₂ was tested in one trial (302 participants). Vitamin D₂ had no statistically significant effect on cancer (RR 0.20 (95% CI 0.02 to 1.69); P = 0.14; Analysis 1.18).
Calcitriol (1,25‐dihydroxyvitamin D)
Calcitriol was tested in three trials (430 participants). Inspection of the funnel plot does not suggest potential bias (asymmetry). Overall, calcitriol had no statistically significant effect on cancer (RR 1.45 (95% CI 0.52 to 4.06); P = 0.48; I² = 0%; Analysis 1.19).
Cancer site occurrence in trials using calcitriol
Calcitriol had no significant effect on breast, uterine, or stomach cancer (one trial each) (Analysis 1.20; Analysis 1.21; Analysis 1.22)
All‐cause mortality
Overall, vitamin D significantly decreased all‐cause mortality (RR 0.93 (95% CI 0.88 to 0.98); P = 0.009; I² = 0%; 15 trials; 49,866 participants; Analysis 1.23). A total of 1854 of the 24,846 participants (7.5%) randomised to the vitamin D group and 2007 of the 25,020 participants (8.0%) randomised to the placebo or no intervention group died. In a trial with low risk of bias, mortality was not significantly changed (RR 0.90 (95% CI 0.77 to 1.07); P = 0.23; 1 trial; 2686 participants). In trials with a high risk of bias, mortality was significantly decreased in the vitamin D group (RR 0.93 (95% CI 0.88 to 0.99); P = 0.02; I² = 0%; 14 trials; 47,180 participants; Analysis 1.23). The difference between the effect estimate of vitamin D on mortality in trials with low risk of bias and in trials with a high risk of bias was not statistically significant (P = 0.75; Analysis 1.23). Trial sequential analysis on mortality in the 15 vitamin D trials was performed based on a mortality rate in the control group of 10%, a relative risk reduction (RRR) of 5% in the experimental group, a type I error of 5%, and type II error of 20% (80% power). There was no diversity. The required information size was 110,505 participants. The cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit after the 15th trial.
Sensitivity analyses taking attrition into consideration
Of the 15 trials reporting mortality, 14 trials reported the exact numbers of participants with missing outcomes in the intervention and control groups. One trial did not report losses to follow‐up for the intervention groups separately (Lappe 2007). A total of 797 of 24,400 participants (3.3%) had missing outcomes in the vitamin D group versus 757 of 24,287 participants (3.1%) in the control group.
'Best‐worst case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group survived and all those with missing outcomes in the control intervention group died, vitamin D supplementation significantly decreased mortality (RR 0.43 (95% CI 0.31 to 0.60); P < 0.00001; I² = 89%; 48,687 participants; 14 trials; Analysis 1.24).
'Worst‐best case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group died and all those with missing outcomes in the control intervention group survived, vitamin D supplementation significantly increased mortality (RR 2.03 (95% CI 1.47 to 2.80); P < 0.0001; I² = 89%; 48,687 participants; 14 trials; Analysis 1.24).
Cancer mortality
Vitamin D₃ may decrease cancer mortality (RR 0.88 (95% CI 0.78 to 0.98); P = 0.02; I² = 0%; 4 trials; 44,492 participants; Analysis 1.25). However, we lack firm evidence for even a 10% RRR since the required information size of 110,505 participants has not yet been reached for such an effect, and the cumulative Z‐curve did not cross the monitoring boundaries (Figure 6).
Trial sequential analysis on cancer mortality in the four vitamin D trials was performed based on cancer mortality of 3% in the control group, a relative risk reduction of 10% with vitamin D₃ supplementation, a type I error of 5%, and a type II error of 20% (80% power). There was no diversity. The required information size was 110,505 participants. The cumulative Z‐curve (blue line) did not cross the trial sequential monitoring boundary (red line) after the fourth trial. The blue line represents the cumulative Z‐score of the meta‐analysis. The green lines represent the conventional statistical boundaries. The red inward sloping lines represent the trial sequential monitoring boundaries.
Sensitivity analyses taking attrition into consideration
All four trials reporting cancer mortality reported the exact numbers of participants with missing outcomes in the intervention and control groups. A total of 613 of 22,286 participants (2.8%) had missing outcomes in the vitamin D group versus 600 of 22,206 participants (2.7%) in the control group.
'Best‐worst case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group survived and all those with missing outcomes in the control intervention group died from cancer, vitamin D supplementation significantly decreased mortality (RR 0.48 (95% CI 0.33 to 0.70); P < 0.0001; I² = 88%; 44,492 participants; 4 trials; Analysis 1.26).
'Worst‐best case' scenario
If we assume that all participants lost to follow‐up in the experimental intervention group died from cancer and all those with missing outcomes in the control intervention group survived, vitamin D supplementation significantly increased mortality (RR 1.69 (95% CI 1.04 to 2.75); P < 0.03; I² = 93%; 44,492 participants; 4 trials; Analysis 1.26).
Secondary outcomes
Adverse events
Several adverse events were reported (for example, hypercalcaemia, nephrolithiasis, hypercalciuria, renal insufficiency, cardiovascular disorders, gastrointestinal disorders, and psychiatric disorders). The supplemental forms of vitamin D (D₃ and D₂) (RR 1.41 (95% CI 0.64 to 3.09); P = 0.39; I² = 0%; 4 trials; 5879 participants), and active form of vitamin D (calcitriol) (RR 4.03 (95% CI 0.56 to 29.22); P = 0.17; I² = 53%; 2 trials; 332 participants) had no statistically significant effect on the risk of hypercalcaemia (Analysis 1.27).
Combined vitamin D₃ and calcium supplements significantly increased nephrolithiasis (RR 1.17 (95% CI 1.03 to 1.34); P = 0.02; I² = 0%; 3 trials; 42,753 participants; Analysis 1.27). Calcitriol had no statistically significant effect on nephrolithiasis (RR 0.33 (95% CI 0.01 to 8.10); P = 0.50; 1 trial; 246 participants; Analysis 1.27). The effect of vitamin D on the other adverse events was not statistically significant: hypercalciuria (RR 12.49 (95% CI 0.72 to 215.84); P = 0.08; 1 trial; 98 participants); renal insufficiency (RR 0.65 (95% CI 0.23 to 1.82); P = 0.41; I² = 4%; 3 trials; 5549 participants); cardiovascular disorders (RR 0.95 (95% CI 0.86 to 1.05); P = 0.28; I² = 0%; 8 trials; 4938 participants); gastrointestinal disorders (RR 1.19 (95% CI 0.88 to 1.59); P = 0.26; I² = 3%; 7 trials; 1624 participants); and psychiatric disorders (RR 1.42 (95% CI 0.46 to 4.38); P = 0.54; I² = 0%; 2 trials; 332 participants; Analysis 1.27).
Health‐related quality of life and health economics
We did not find any data on health‐related quality of life or health economics in the randomised trials included in this review.
Discussion
Summary of main results
Our systematic review contains a number of important findings. We have found evidence that vitamin D supplements in the form of vitamin D₃, vitamin D₂, or calcitriol have no clear effect on occurrence of cancer in mainly elderly community‐dwelling women. This effect does not seem to be due to random errors ('play of chance'). Vitamin D supplements have no clear effect on any specific cancer type. Vitamin D supplements decreased cancer mortality and all‐cause mortality, but these estimates are at risk of type I errors due to the fact that too few participants were examined and to substantial attrition bias. Combined vitamin D₃ and calcium supplements increased nephrolithiasis, but it remains unclear from the included trials whether vitamin D₃, calcium, or both were responsible for this effect. Vitamin D supplements have no clear effects on other adverse events.
Overall completeness and applicability of evidence
Our published protocol described our plan to analyse the effect of vitamin D on cancer in primary and secondary prevention randomised trials in adults. We included all eligible randomised trials up to February 2014. All trials were conducted in high‐income countries. Participants of both genders were included. Most of the participants were elderly community‐dwelling women. The vast majority of the participants came from primary prevention trials, and we assume that they were apparently healthy when included in the trials. Few trials of secondary prevention with low participation rates were included, so our ability to say anything about such patients is severely limited. We included randomised trials with both vitamin D–deficient participants and those who seemed to have adequate vitamin D levels at entry. We were unable to detect substantial differences regarding these variables on the estimated intervention effect on cancer. We found surprisingly little statistical heterogeneity in any of our analyses. Most trials assessed vitamin D₃, and our major conclusions relate to this intervention. Most of the trials were considered to be at high risk of bias, mainly for‐profit bias. Our analyses revealed that outcome reporting was missing on more than 3% of participants. This number is too high when cancer occurrence is about 7% in the placebo or no intervention/placebo group. Accordingly, our 'best‐worst case' and 'worst‐best case' analyses revealed that our results were compatible with both a very large beneficial effect and a very large detrimental effect of vitamin D on cancer. Although these extreme sensitivity analyses are unlikely, they reveal how few missing participants would need to have developed cancer to substantially change our findings of modest benefit into a null effect or maybe even harm. We therefore warn against uncritical application of our findings.
Quality of the evidence
Our review follows the overall plan of a published, peer‐reviewed Cochrane protocol (Bjelakovic 2008). It represents a comprehensive review of the topic, including 18 randomised trials with more than 50,000 participants. This increases the precision and power of our analyses (Higgins 2011; Turner 2013). We were not able to find in the literature earlier meta‐analyses of preventive trials of vitamin D on cancer occurrence. We conducted a thorough review in accordance with The Cochrane Collaboration methodology (Higgins 2011) and implementing findings of methodological studies (Kjaergard 2001; Lundh 2012; Moher 1998; Savović 2012a; Savović 2012b; Schulz 1995; Wood 2008). Between‐trial statistical heterogeneity is almost absent from our meta‐analyses, which enhances the consistency of our findings.
We conducted a number of subgroup analyses. We observed no statistically significantly different effects of the intervention of vitamin D supplementation on cancer in subgroup analyses of trials at low risk of bias compared to trials at high risk of bias; of trials at no risk of for‐profit bias compared to trials at risk of for‐profit bias; of trials assessing primary prevention compared to trials assessing secondary prevention; of trials including participants with vitamin D level below 20 mg/mL at entry compared to trials including participants with normal vitamin D levels at entry; and of vitamin D₃ trials using concomitant calcium supplementation compared to vitamin D₃ trials without calcium.
We also performed trial sequential analyses to control for the risk of random errors in a cumulative meta‐analysis and to prevent premature statements of superiority of vitamin D, based on estimation of the diversity‐adjusted required information size (Brok 2008; Brok 2009; Thorlund 2009; Thorlund 2011a; Thorlund 2011b; Wetterslev 2008; Wetterslev 2009).
A major obstacle in many of the included trials is the relatively large proportion of about 3% of participants with missing outcomes in both experimental and control groups. This opens the way for attrition bias, and our 'best‐worst' and 'worst‐best' intention‐to‐treat analyses demonstrate that the intervention effect of vitamin D may be either beneficial or harmful. Although both of the two extreme scenarios are unlikely, they demonstrate that we cannot depend fully on the estimates we arrive at. Our 'best‐worst case' and 'worst‐best case' scenario analyses revealed much more extreme confidence limits for cancer occurrence (95% CI 0.31 to 3.76) compared to our 'complete‐case' scenario analysis (95% CI 0.94 to 1.06); for all‐cause mortality (95% CI 0.31 to 2.80) compared to our 'complete‐case' scenario analysis (95% CI 0.88 to 0.98), and for cancer mortality (0.33 to 2.75) compared to our 'complete‐case' scenario analysis (95% CI 0.78 to 0.98). Those analyses convey a message of a noticeable degree of uncertainty regarding our results. This observation calls for more comprehensive meta‐analyses of individual participant data.
We used GRADE to construct a 'Summary of findings' table for vitamin D for the review outcome measures. Results obtained by use of TSA were applied for the rating for imprecision. If there was insufficient evidence to reach a conclusion, i.e., if the TSA indicates that the required information size had not been reached, we downgraded the quality of the evidence by one level. We also used risk of attrition bias for the rating for imprecision. As the cumulative Z‐curve for the intervention effect on cancer occurrence has reached the futility area for a 5% relative risk reduction (RRR) even though the required information size for such an effect has not been reached, there is precision enough to refute a 5% RRR. However, if, for example, a 2.5% RRR is still a clinically relevant effect, there may still be imprecision for detecting or rejecting such an effect of 2.5% RRR. If we consider a 2.5% RRR, which is equivalent to a number needed to treat for an additional beneficial outcome (NNTB) of 400, a worthwhile effect, then the downgrading for imprecision seems appropriate for the ability to refute a 5% RRR of cancer occurrence.
Potential biases in the review process
Most of the included trials are at high risk of bias, which undermines the validity of our results (Kjaergard 2001; Lundh 2012; Moher 1998; Savović 2012a; Savović 2012b; Schulz 1995; Wood 2008). Our meta‐analyses show a lack of statistical heterogeneity, which may emphasise the consistency of our findings but should also raise concern (Ioannidis 2006). Statistical homogeneity may be due to inappropriate inferences of the asymptotic Q test with sparse data (Ioannidis 2006). We have also performed trial sequential analyses, based on the estimation of the diversity‐adjusted required information size in order to avoid an undue risk of random errors in a cumulative meta‐analysis and to prevent premature statements of superiority of vitamin D or of lack of effect (Brok 2008; Brok 2009; Thorlund 2009; Wetterslev 2008; Wetterslev 2009).
Certain potential limitations of this review warrant consideration. As with all systematic reviews, our findings and interpretations are limited by the quality and quantity of available evidence on the effects of vitamin D on cancer. Despite extensive speculations in the literature and a large number of epidemiological studies that claim cancer preventive effects of vitamin D, few randomised trials have been conducted assessing cancer occurrence. The duration of supplementation and duration of follow‐up was short in some of the included trials compared to the long process of carcinogenesis. This may make it difficult to detect any effects, beneficial or harmful. Among the 18 included randomised trials, cancer occurrence was the primary prespecified outcome in few of them. Different forms of vitamin D were used for supplementation. The majority of included trials used vitamin D₃, one trial tested vitamin D₂, and three trials tested calcitriol. Our subgroup analyses found that the effect of vitamin D on cancer was neutral, irrespective of the form of vitamin D used for supplementation. Most trials investigated the effects of vitamin D administered at lower doses than those recently suggested as cancer‐preventive (Garland 2011). All the included trials came from high‐income countries, and the majority of them included participants without overt deficiencies of vitamin D. Accordingly, we are unable to assess how vitamin D affects cancer in populations with specific needs.
Agreements and disagreements with other studies or reviews
We found that vitamin D had neutral effect on cancer occurrence. This finding is in accord with the result of our previous review (Bjelakovic 2014), and contradicts the results of epidemiological studies (Bikle 2014; Garland 2006; Giovannucci 2005; Ordóñez‐Mena 2013; Redaniel 2014; Woloszynska‐Read 2011; Yin 2013). Our results are in accord with the conclusions of the recently published International Agency for Research on Cancer and Institute of Medicine reports stating that vitamin D status is not correlated with cancer occurrence (IARC 2008; IOM 2011). Recently, an updated meta‐analysis prepared for the US Preventive Services Task Force found inconclusive evidence regarding vitamin D supplementation for the prevention of cancer (Chung 2011).
We found no substantive differences regarding the effect of vitamin D on cancer in trials including participants with vitamin D insufficiency (25‐hydroxyvitamin D level less than 20 ng/mL) compared to trials including participants with optimal vitamin D status. Optimal vitamin D status has been linked to decreased incidence of several types of cancer (Garland 2007; Gorham 2007). However, a number of observational studies have also suggested that high vitamin D status might be connected with increased oesophageal (Chen 2007), pancreatic (Stolzenberg‐Solomon 2006), breast (Goodwin 2009), and prostate cancer risks (Ahn 2008). One should consider the possibility of a U‐shaped relationship between vitamin D status and cancer risk (Toner 2010; Tuohimaa 2012). Once again, we may be witnessing the flawed inferences that are sometimes drawn from observational epidemiological data (Jakobsen 2013).
We have examined the influence of different forms of vitamin D on cancer occurrence. We have found neutral effect irrespective of the form of vitamin D used. Our previous systematic review on vitamin D and mortality has found weak evidence that only vitamin D₃ may be of benefit for survival (Bjelakovic 2014). We found no evidence for vitamin D₂, alfacalcidol, and calcitriol affecting mortality, but these estimates are at risk of type II errors (the chance of not rejecting the null hypothesis when it is in fact false) due to the fact that small groups of participants were examined. A number of recently published clinical trials (Armas 2004; Heaney 2011; Lehmann 2013; Leventis 2009; Logan 2013; Romagnoli 2008; Trang 1998), and the systematic review (Tripkovic 2012) found evidence that vitamin D₃ increases serum 25‐hydroxyvitamin D more efficiently than vitamin D₂. Several clinical trials have been conducted to examine the effects of alfacalcidol and calcitriol on different health outcomes. These trials have mostly included women with osteoporosis (Ott 1989; Shiraki 2004) and people with chronic kidney disease (Palmer 2009). Due to the lack of a demonstrable effect, and observed adverse events (hypercalcaemia), the active forms of vitamin D are increasingly being replaced with the supplemental forms of vitamin D (cholecalciferol and ergocalciferol). Three trials included in our present systematic review tested calcitriol. The evidence for an effect on cancer is inconclusive, but the sample size was too small to exclude possible effects. We were not able to identify trials testing alfacalcidol that reported cancer occurrence at the end of the follow‐up period.
The effect of vitamin D₃ on cancer occurrence was not statistically significant in trials using vitamin D₃ singly or trials using vitamin D₃ combined with calcium. Because of the small number of included trials assessing vitamin D₃ alone or combined with calcium, the findings could be due to a type II error. Our finding seems consistent with the result obtained by Bristow 2013, who found that calcium supplements did not affect cancer, but contradict the results of recent meta‐analyses examining the influence of vitamin D on mortality (Rejnmark 2012) or bone health (DIPART 2010). These meta‐analyses concluded that vitamin D is effective in preventing mortality (Rejnmark 2012) and hip fractures (DIPART 2010) only when combined with calcium. A recent meta‐analysis observed that calcium supplementation (with or without co‐administration of vitamin D) is associated with increased risk of cardiovascular events, especially myocardial infarction (Bolland 2010; Bolland 2011). Another review of prospective studies and randomised clinical trials found no evidence for an effect of calcium (Patel 2012). A US Preventive Services Task Force recently recommended against daily supplementation with 400 IU or less of vitamin D₃ and 1000 mg or less of calcium for the primary prevention of fractures in non‐institutionalised postmenopausal women (Moyer 2013).The complex interactions between vitamin D and calcium make it difficult to separate their effects (Lips 2012). More clinical research seems needed.
We have examined the influence of vitamin D supplementation on different cancer types. Vitamin D showed no substantive effect on any specific cancer type. Our results are in accord with the results of the recently published systematic review and meta‐analysis investigating the role of vitamin D supplementation on breast cancer (Sperati 2013). Our results contradict earlier speculations in the literature about the preventive effects of vitamin D on certain cancer types (Fleet 2012; Woloszynska‐Read 2011).
We have found no evidence for an effect of vitamin D supplements on cancer occurrence. However, vitamin D supplements seemed to decrease cancer mortality. Epidemiological studies that examined the relationship between vitamin D status and cancer mortality have shown mixed results. Pilz and coworkers found that optimal vitamin D status seems to be related to decreased cancer mortality in some studies (Pilz 2009; Pilz 2013). On the other hand, Freedman 2010 found no overall relationship between vitamin D status and cancer mortality in the general population. Our finding of no evidence that vitamin D has an effect on cancer occurrence seems to be at odds with the statistically significant beneficial effect of vitamin D on cancer mortality. However, due to low participation rates in the included trials, the decreased cancer mortality could be a random error as indicated by our trial sequential analysis. One should also consider that as yet unknown preventive effects of vitamin D supplementation on premature death, which may not necessarily influence cancer incidence, might play a role (Bjelakovic 2014).
Vitamin D supplements decreased all‐cause mortality. This is in accord with the results of our previous review of the role of vitamin D in mortality prevention (Bjelakovic 2014), in which we disregarded the risks of random error and attrition bias. However, if these risks are considered, we do not yet know whether vitamin D affects mortality.
Combined vitamin D₃ and calcium supplements increased nephrolithiasis. This is in accord with the results of our previous review (Bjelakovic 2014). Other adverse events, including elevated urinary calcium excretion; renal insufficiency; cardiovascular, gastrointestinal, or psychiatric disorders, were not substantively influenced by vitamin D supplementation.
We lack sufficient evidence for the effect of vitamin D supplementation on health‐related quality of life or the cost effectiveness of vitamin D supplementation. However, vitamin D₃ products and calcium are affordable, with multiple producers across the world, so these interventions may be cost‐effective.
Despite biological plausibility for the role of vitamin D in the prevention of cancer, the available evidence does not seem to support this possibility. Several recently published evidence reports have assessed the influence of vitamin D and calcium on different health outcomes (Bolland 2014; Chung 2011; Cranney 2007; Moyer 2013; Rosen 2012). The majority of the findings on different health outcomes including cancer were equivocal. The US Institute of Medicine reported that available evidence supports a role for vitamin D and calcium in skeletal health (IOM 2011). The evidence was, however, considered insufficient and inconclusive for extraskeletal outcomes including cancer (IOM 2011). Most recent systematic review and meta‐analyses concluded that vitamin D supplementation for osteoporosis prevention in free‐living adults without specific risk factors for vitamin D deficiency seems to be inappropriate (Reid 2014). In the same vein, the recently updated Systematic Evidence Review for the US Preventive Service Task Force found no evidence of benefit from vitamin supplementation for the prevention of cancer and cardiovascular disease (Fortmann 2013). Accordingly, these comprehensive reports are compatible with our findings.
It seems that health claims are again ahead of the evidence. It is very likely that low vitamin D status is not the cause, but rather the consequence of disease (Autier 2014; Guallar 2010; Guallar 2013; Harvey 2012; Kupferschmidt 2012). Results of several large ongoing randomised trials that assess vitamin D supplementation for different health outcomes will likely help us to elucidate the role of vitamin D.
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.
Funnel plot of comparison: 1 Vitamin D versus placebo or no intervention, outcome: 1.1 Cancer occurrence in trials with a low or high risk of bias.
Trial sequential analysis on cancer occurrence in the 18 vitamin D trials was performed based on cancer occurrence of 10% in the control group, a relative risk reduction of 5% with vitamin D supplementation, a type I error of 5%, and a type II error of 20% (80% power). There was no diversity. This resulted in a required information size of 110,505 participants. Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. The blue line represents the cumulative Z‐score of the meta‐analysis. The green lines represent the conventional statistical boundaries. The red inward sloping lines represent the trial sequential monitoring boundaries.
Trial sequential analysis on cancer mortality in the four vitamin D trials was performed based on cancer mortality of 3% in the control group, a relative risk reduction of 10% with vitamin D₃ supplementation, a type I error of 5%, and a type II error of 20% (80% power). There was no diversity. The required information size was 110,505 participants. The cumulative Z‐curve (blue line) did not cross the trial sequential monitoring boundary (red line) after the fourth trial. The blue line represents the cumulative Z‐score of the meta‐analysis. The green lines represent the conventional statistical boundaries. The red inward sloping lines represent the trial sequential monitoring boundaries.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 1 Cancer occurrence in trials with a low or high risk of bias.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 2 Cancer occurrence and risk of for‐profit bias.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 3 Cancer occurrence in primary and secondary prevention trials.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 4 Cancer occurrence and vitamin D status.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 5 Cancer occurrence ('best‐worst case' and 'worst‐best case' scenario).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 6 Cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 7 Cancer occurrence in trials using vitamin D₃ singly or combined with calcium.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 8 Lung cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 9 Breast cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 10 Colorectal cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 11 Pancreatic cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 12 Prostate cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 13 Uterine cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 14 Ovarian cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 15 Oesophageal cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 16 Stomach cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 17 Liver cancer occurrence in trials using vitamin D₃ (cholecalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 18 Cancer occurrence in trials using vitamin D₂ (ergocalciferol).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 19 Cancer occurrence in trials using calcitriol.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 20 Breast cancer occurrence in trials using calcitriol.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 21 Uterine cancer occurrence in trials using calcitriol.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 22 Stomach cancer occurrence in trials using calcitriol.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 23 All‐cause mortality in trials with a low or high risk of bias.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 24 All‐cause mortality ('best‐worst case' and 'worst‐best case' scenario).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 25 Cancer mortality.
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 26 Cancer mortality ('best‐worst case' and 'worst‐best case' scenario).
Comparison 1 Vitamin D versus placebo or no intervention, Outcome 27 Adverse events.
| Vitamin D versus placebo or no intervention for prevention of cancer in adults | ||||||
| Patient or population: healthy participants or recruited among the general population; individuals diagnosed with a specific disease in a stable phase or with vitamin D deficiency Settings: outpatients | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| Control | Vitamin D versus placebo or no intervention | |||||
| Cancer occurrence Follow‐up: 0.5 to 7 years | Study population | RR 1.00 | 50623 | ⊕⊕⊕⊝ moderatea | Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. | |
| 77 per 1000 | 77 per 1000 | |||||
| Moderate | ||||||
| 28 per 1000 | 28 per 1000 | |||||
| Cancer occurrence in trials using vitamin D₃ (cholecalciferol) Follow‐up: 0.5 to 7 years | Study population | RR 1.00 | 49891 | ⊕⊕⊕⊝ moderatea | Trial sequential analysis of all vitamin D trials suggests that the futility area is reached after the 10th trial allowing us to conclude that any possible intervention effect, if any, is lower than a 5% relative risk reduction. | |
| 77 per 1000 | 77 per 1000 | |||||
| Moderate | ||||||
| 28 per 1000 | 28 per 1000 | |||||
| All‐cause mortality Follow‐up: 0.5 to 7 years | Study population | RR 0.93 | 49866 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 80 per 1000 | 75 per 1000 | |||||
| Moderate | ||||||
| 16 per 1000 | 15 per 1000 | |||||
| Cancer mortality in trials using vitamin D₃(cholecalciferol) Follow‐up: 5 to 7 years | Study population | RR 0.88 | 44492 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 29 per 1000 | 25 per 1000 | |||||
| Moderate | ||||||
| 37 per 1000 | 33 per 1000 | |||||
| Adverse events: nephrolithiasis in trials using vitamin D₃(cholecalciferol) combined with calcium Follow‐up: 0.5 to 7 years | Study population | RR 1.17 | 42753 | ⊕⊕⊝⊝ lowb | Trial sequential analysis of all trials irrespective of bias risks showed that the required information size had not yet been reached and that the cumulative Z‐curve did not cross the trial sequential monitoring boundary for benefit. | |
| 18 per 1000 | 21 per 1000 | |||||
| Moderate | ||||||
| 1 per 1000 | 1 per 1000 | |||||
| Health‐related quality of life | See comment | Not investigated. | ||||
| Health economics | See comment | Not investigated. | ||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded by one level because of risk of attrition bias bDowngraded by two levels because of risk of attrition bias and imprecision | ||||||
| Characteristic | Intervention(s) and comparator(s) | Screened/eligible | Randomised | ITT | Finishing study | Randomised finishing study |
| (1) Avenell 2012
| I1: vitamin D₃ | 15,024 | 1343 | 1343 | 1813 | 68 |
| I2: vitamin D₃ plus calcium | 1306 | 1306 | ||||
| C1: calcium | 1311 | 1311 | 1762 | 67 | ||
| C2: matched placebo tablets | 1332 | 1332 | ||||
| total: | 5292 | 5292 | 3575 | 68 | ||
| (2) Bolton‐Smith 2007
| I1: vitamin D₃ plus calcium | ‐ | 62 | 62 | 50 | 81 |
| C1: matched placebo | 61 | 61 | 56 | 92 | ||
| total: | 123 | 123 | 106 | 86 | ||
| (3) Brunner 2011
| I1: vitamin D₃ plus calcium | 68,132 | 18,176 | 18,176 | 16,936 | 93 |
| C1: matched placebo | 18,106 | 18,106 | 16,815 | 93 | ||
| total: | 36,282 | 36,282 | 33,751 | 93 | ||
| (4) Daly 2008
| I1: calcium‐vitamin D₃‐fortified milk plus calcium | 422 | 85 | 85 | 76 | 89 |
| C1: usual diet | 82 | 82 | 73 | 89 | ||
| total: | 167 | 167 | 149 | 89 | ||
| (5) Gallagher 2001
| I1: calcitriol | 1905 | 123 | 123 | 101 | 82 |
| C1: matched placebo | 123 | 123 | 112 | 91 | ||
| total: | 246 | 246 | 213 | 87 | ||
| (6) Glendenning 2012
| I1: cholecalciferol | 2110 | 353 | 353 | 331 | 94 |
| C1: placebo vitamin D | 333 | 333 | 307 | 92 | ||
| total: | 686 | 686 | 638 | 93 | ||
| (7) Grady 1991
| I1: calcitriol | 98 | 50 | 50 | 49 | 98 |
| C1: placebo vitamin D | 48 | 48 | 48 | 100 | ||
| total: | 98 | 50 | 97 | 99 | ||
| (8) Janssen 2010
| I1: vitamin D₃ plus calcium | 91 | 36 | 36 | 18 | 50 |
| C1:placebo vitamin D₃ plus calcium | 34 | 34 | 31 | 91 | ||
| total: | 70 | 70 | 49 | 70 | ||
| (9) Komulainen 1999
| I1: vitamin D₃ plus calcium | 13,100 | 116 | 116 | 112 | 97 |
| C1: placebo | 116 | 116 | 115 | 99 | ||
| total: | 232 | 232 | 227 | 98 | ||
| (10) Lappe 2007
| I1: vitamin D₃ plus calcium | 1180 | 446 | 446 | 403 | 90 |
| C1: vitamin D₃ placebo plus calcium | 445 | 445 | 416 | 93 | ||
| C2: vitamin D₃ placebo plus calcium placebo | 288 | 288 | 266 | 92 | ||
| total: | 1179 | 1179 | 1085 | 92 | ||
| (11) Larsen 2012 | I1: vitamin D₃ | 136 | 65 | 65 | 55 | 85 |
| C1: vitamin D placebo | 65 | 65 | 57 | 88 | ||
| total: | 130 | 130 | 112 | 86 | ||
| (12) Murdoch 2012 | I1: vitamin D₃ | 351 | 161 | 161 | 148 | 92 |
| C1: vitamin D placebo | 161 | 161 | 146 | 91 | ||
| total: | 322 | 322 | 294 | 91 | ||
| (13) Ott 1989
| I1: calcitriol plus calcium | ‐ | 43 | 43 | 39 | 91 |
| C1: placebo vitamin D plus calcium | 43 | 43 | 37 | 86 | ||
| total: | 86 | 86 | 76 | 88 | ||
| (14) Prince 2008
| I1: vitamin D₂ plus calcium | 827 | 151 | 151 | 144 | 95 |
| C1: placebo vitamin D plus calcium | 151 | 151 | 145 | 96 | ||
| total: | 302 | 302 | 289 | 95 | ||
| (15) Sanders 2010
| I1: vitamin D₃ | 7204 | 1131 | 1131 | 1015 | 90 |
| C1: vitamin D placebo | 1127 | 1127 | 1017 | 90 | ||
| total: | 2258 | 2258 | 2032 | 90 | ||
| (16) Trivedi 2003
| I1: vitamin D₃ | ‐ | 1345 | 1345 | 1262 | 94 |
| C1:placebo vitamin D | 1341 | 1341 | 1264 | 94 | ||
| total: | 2686 | 2686 | 2526 | 94 | ||
| (17) Witham 2013 | I1: vitamin D₃ | 341 | 80 | 80 | 73 | 91 |
| C1: placebo vitamin D | 79 | 79 | 69 | 87 | ||
| total: | 159 | 159 | 142 | 89 | ||
| (18) Wood 2012 | I1: vitamin D₃ | 424 | 102 | 102 | 84 | 82 |
| I2: vitamin D₃ | 101 | 101 | 90 | 89 | ||
| C1: placebo vitamin D | 102 | 102 | 91 | 89 | ||
| total: | 305 | 305 | 265 | 87 | ||
| Grand total | All interventions | 25,275 | 22,799 | 90 | ||
| All controls | 25,348 | 22,827 | 90 | |||
| All interventions and controls | 50,623 | 45,626 | 90 | |||
| "‐" denotes not reported | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Cancer occurrence in trials with a low or high risk of bias Show forest plot | 18 | 50623 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 1.1 Trials with low risk of bias | 2 | 2991 | Risk Ratio (M‐H, Random, 95% CI) | 1.08 [0.89, 1.31] |
| 1.2 Trials with high risk of bias | 16 | 47632 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.93, 1.05] |
| 2 Cancer occurrence and risk of for‐profit bias Show forest plot | 18 | 50623 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 2.1 Trials without risk of for‐profit bias | 2 | 2991 | Risk Ratio (M‐H, Random, 95% CI) | 1.08 [0.89, 1.31] |
| 2.2 Trials with risk of for‐profit bias | 16 | 47632 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.93, 1.05] |
| 3 Cancer occurrence in primary and secondary prevention trials Show forest plot | 18 | 50623 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 3.1 Primary prevention trials | 16 | 50334 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 3.2 Secondary prevention trials | 2 | 289 | Risk Ratio (M‐H, Random, 95% CI) | 1.33 [0.26, 6.96] |
| 4 Cancer occurrence and vitamin D status Show forest plot | 18 | 50623 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 4.1 Vitamin D insufficiency | 7 | 44668 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.93, 1.05] |
| 4.2 Vitamin D adequacy | 9 | 4544 | Risk Ratio (M‐H, Random, 95% CI) | 1.12 [0.94, 1.34] |
| 4.3 Unknown vitamin status | 2 | 1411 | Risk Ratio (M‐H, Random, 95% CI) | 0.59 [0.33, 1.05] |
| 5 Cancer occurrence ('best‐worst case' and 'worst‐best case' scenario) Show forest plot | 17 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 5.1 'Best‐worst' case scenario | 17 | 49444 | Risk Ratio (M‐H, Random, 95% CI) | 0.41 [0.31, 0.54] |
| 5.2 'Worst‐best' case scenario | 17 | 49444 | Risk Ratio (M‐H, Random, 95% CI) | 2.76 [1.97, 3.86] |
| 6 Cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 14 | 49891 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.06] |
| 6.1 Vitamin D₃ trials with low risk of bias | 2 | 2991 | Risk Ratio (M‐H, Random, 95% CI) | 1.08 [0.89, 1.31] |
| 6.2 Vitamin D₃ trials with high risk of bias | 12 | 46900 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.93, 1.05] |
| 7 Cancer occurrence in trials using vitamin D₃ singly or combined with calcium Show forest plot | 14 | 49870 | Risk Ratio (M‐H, Random, 95% CI) | 0.98 [0.92, 1.04] |
| 7.1 Vitamin D₃ singly | 8 | 9200 | Risk Ratio (M‐H, Random, 95% CI) | 1.03 [0.90, 1.17] |
| 7.2 Vitamin D₃ combined with calcium | 7 | 40670 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.91, 1.04] |
| 8 Lung cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 5 | 45509 | Risk Ratio (M‐H, Random, 95% CI) | 0.86 [0.69, 1.07] |
| 9 Breast cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 7 | 43669 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.86, 1.09] |
| 10 Colorectal cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 5 | 45598 | Risk Ratio (M‐H, Random, 95% CI) | 1.11 [0.92, 1.34] |
| 11 Pancreatic cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 2 | 36405 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.57, 1.46] |
| 12 Prostate cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 13 Uterine cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 14 Ovarian cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 15 Oesophageal cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 16 Stomach cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 17 Liver cancer occurrence in trials using vitamin D₃ (cholecalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 18 Cancer occurrence in trials using vitamin D₂ (ergocalciferol) Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 19 Cancer occurrence in trials using calcitriol Show forest plot | 3 | 430 | Risk Ratio (M‐H, Random, 95% CI) | 1.45 [0.52, 4.06] |
| 20 Breast cancer occurrence in trials using calcitriol Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 21 Uterine cancer occurrence in trials using calcitriol Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 22 Stomach cancer occurrence in trials using calcitriol Show forest plot | 1 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 23 All‐cause mortality in trials with a low or high risk of bias Show forest plot | 15 | 49866 | Risk Ratio (M‐H, Random, 95% CI) | 0.93 [0.88, 0.98] |
| 23.1 Trials with low risk of bias | 1 | 2686 | Risk Ratio (M‐H, Random, 95% CI) | 0.90 [0.77, 1.07] |
| 23.2 Trials with high risk of bias | 14 | 47180 | Risk Ratio (M‐H, Random, 95% CI) | 0.93 [0.88, 0.99] |
| 24 All‐cause mortality ('best‐worst case' and 'worst‐best case' scenario) Show forest plot | 14 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 24.1 'Best‐worst' case scenario | 14 | 48687 | Risk Ratio (M‐H, Random, 95% CI) | 0.43 [0.31, 0.60] |
| 24.2 'Worst‐best' case scenario | 14 | 48687 | Risk Ratio (M‐H, Random, 95% CI) | 2.03 [1.47, 2.80] |
| 25 Cancer mortality Show forest plot | 4 | 44492 | Risk Ratio (M‐H, Random, 95% CI) | 0.88 [0.78, 0.98] |
| 26 Cancer mortality ('best‐worst case' and 'worst‐best case' scenario) Show forest plot | 4 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 26.1 'Best‐worst' case scenario | 4 | 44492 | Risk Ratio (M‐H, Random, 95% CI) | 0.48 [0.33, 0.70] |
| 26.2 'Worst‐best' case scenario | 4 | 44492 | Risk Ratio (M‐H, Random, 95% CI) | 1.69 [1.04, 2.75] |
| 27 Adverse events Show forest plot | 15 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 27.1 Hypercalcaemia in trials using supplemental forms of vitamin D | 4 | 5879 | Risk Ratio (M‐H, Random, 95% CI) | 1.41 [0.64, 3.09] |
| 27.2 Hypercalcaemia in trials using active forms of vitamin D | 2 | 332 | Risk Ratio (M‐H, Random, 95% CI) | 4.03 [0.56, 29.22] |
| 27.3 Nephrolithiasis in trials using vitamin D₃ combined with calcium | 3 | 42753 | Risk Ratio (M‐H, Random, 95% CI) | 1.17 [1.03, 1.34] |
| 27.4 Nephrolithiasis in trials using calcitriol | 1 | 246 | Risk Ratio (M‐H, Random, 95% CI) | 0.33 [0.01, 8.10] |
| 27.5 Hypercalciuria | 1 | 98 | Risk Ratio (M‐H, Random, 95% CI) | 12.49 [0.72, 215.84] |
| 27.6 Renal insufficiency | 3 | 5549 | Risk Ratio (M‐H, Random, 95% CI) | 0.65 [0.23, 1.82] |
| 27.7 Cardiovascular disorders | 8 | 4938 | Risk Ratio (M‐H, Random, 95% CI) | 0.95 [0.86, 1.05] |
| 27.8 Gastrointestinal disorders | 7 | 1624 | Risk Ratio (M‐H, Random, 95% CI) | 1.19 [0.88, 1.59] |
| 27.9 Psychiatric disorders | 2 | 332 | Risk Ratio (M‐H, Random, 95% CI) | 1.42 [0.46, 4.38] |
