Impactos del cambio climático

Autier et al performed a meta-analysis of 18 RCTs with a total of 57,311 male and female participants in 2007 (Autier and Gandini, 2007). The primary outcome of the majority of these RCTs was to assess the effects of Calcium and Vitamin D on fractures, bone mineral density and osteoporosis. The mean length of follow up was 5.8 years and 4,777 deaths (8.3%) of any cause occurred. Baseline mean Vitamin D was not documented but mean Vitamin D supplementation was 528IU/day. Analysis showed significant risk reduction in all cause mortality in those supplemented with Vitamin D, SRR 0.92 (0.86-0.99, 95% CI), reporting a 7% decrease in mortality in this group. There appeared to be no significant difference between lower and higher doses of Vitamin D supplementation in subgroup analysis 300-799IU Vitamin D, SRR 0.93, (0.85-1.03, 95% CI) versus >800IU Vitamin D, SRR 0.92, (0.82-1.03, 95% CI).

A Cochrane review published in 2011 (Bjelakovic et al., 2011) reviewed RCTs that compared Vitamin D supplementation with placebo or no intervention. Vitamin D supplementation included Vitamin D2, D3, alfacalcidol and calcitrol. Analysis included 50 RCTS with 94,148 participants. All trials came from high-income countries and the main primary outcomes included bone mineral density, falls, fractures and mortality. The mean age of participants was 74 years. The median duration of Vitamin D supplementation was two years. Participants in 18 trials had levels >50nmol/L and the remaining 22 trials had baseline levels <50nmol/L. The mean daily dose of Vitamin D3 (administered in 32trials) was 804IU/day. 32 trials used calcium combined with Vitamin D at a mean dose of 929mg/day. Overall mortality was lower in Vitamin D groups (11.2%) compared with placebo or no intervention groups (11.4%), RR 0.97 (0.94 to 1.00, 95% CI) p=0.03. When participants were deficient in Vitamin D, supplementation with Vitamin D3 reduced mortality, RR 0.94 0.90 to 0.99, 95% CI) p=0.02 but there was no benefit in Vitamin D3 supplementation when patients were not deficient at baseline, RR 0.92 (0.79 to 1.07, 95% CI) p=0.27. Trials using Vitamin D2 supplementation showed no significant effects on mortality overall, RR 1.02 (0.97 to 1.09, 95% CI) p=0.42. Of note, when adverse events were assessed, no significant effect on risk of hypercalcaemia was noted with Vitamin D2

significantly increased the risk of nephrolithiasis, RR 1.17 (1.00 to 1.34, 95% CI) p=0.02. Alfacalcidol and calcitrol were associated with increased the risk of hypercalcaemia, RR 3.18 (1.17-8.68, 95% CI).

Schottker et al published a systematic review and meta-analysis looking at the relationship between Vitamin D and mortality in 2012 (Schottker et al., 2013a). This analysis involved original longitudinal cohort studies conducted in the general population reporting an association of measured 25(OH)D levels and all cause mortality, only 12 studies were included, with 31,528 participants and 6,921 deaths over the follow-up time ranging from 5.9 to 14 years. The meta-analysis summarises the results of 12 cohort studies. The pooled HR demonstrated an inverse association between 25(OH)D and mortality, HR 0.92 (0.89- .095, 95% CI). For a 25nmol/L increase in 25(OH)D the HR for mortality was 0.90 (0.86- .094, 95% CI) and for a 50nmol/L increase, HR 0.81 (0.74-0.88, 95% CI).

A further meta-analysis was conducted in 2012 (Zittermann et al., 2012) and included observational studies that reported Relative Risks (RR) or crude data on Vitamin D and mortality. Ultimately 14 studies were included in non-parametric and 11 in parametric analysis. In non-parametric analysis the summary estimate for the highest versus lowest Vitamin D categories showed reduced mortality RR 0.71 (0.50-0.91, 95% CI). On parametric analysis the median reference category of Vitamin D was 27.5nmol/L, the best fitting model presented the following estimates, β1 0.085, (-0.120. -0.050, 95% CI) and β2 0.0018, (0.0008, 0.0030, 95% CI). An increase in Vitamin D concentration was associated with decreased mortality.

Rejnmark et al published a review in 2012 assessing the effects of Vitamin D supplementation on mortality using both Individual Participant Data (IPD) level and study level meta-analyses (Rejnmark et al., 2012). Inclusion criteria specified only RCTs with administration of Vitamin D2 or Vitamin D3 in one intervention arm and one arm without Vitamin D. Their search revealed 11 eligible trials for inclusion. IPD analysis included 70,528 participants, 86.8% female and median age 70 years. A total of 27,345 were

randomised to Calcium and Vitamin D, 7,771 to Vitamin D alone, and 35,412 received placebo or no Vitamin D. The mean follow up time was 3 years and 3,832 participants died (5.4%). In those randomised to receive Vitamin D (either alone or with calcium) 1,870 deaths occurred (5.3%) whereas 1,962 deaths (5.5%) occurred in those not receiving Vitamin D, HR 0.94 (0.88-1.00,95% CI). When adjusted, risk of death was reduced by 7%, HR 0.93 (0.88-0.99, 95% CI). When analysis of studies using stratified Cox regression models was performed, risk of death was reduced among these participants receiving Calcium and Vitamin D, HR 0.91 (0.84-0.98, 95% CI), but assessing risk of death with Vitamin D alone there was no significant difference noted HR 0.96 (0.87-1.06, 95% CI).

In 2013 Zhang et al reported a meta-analysis of the effects of long-term Vitamin D and overall mortality (Zheng et al., 2013). After exclusion criteria were applied 42 RCTS with a total of 85,466 participants were included. The mean age ranged form 37-89 years. Vitamin D levels were reported in 37 studies. Participants in 22 studies had 25(OH)D levels <50nmol/L. Prolonged treatment with Vitamin D, for longer than 3 years (13 studies, 3693 participants randomised to Vitamin D and 3,880 randomised to placebo), was found to significantly reduce mortality with a RR 0.94 (0.90-0.98, 95%CI) whereas shorter treatment duration (<3years, 29 trials, 1,175 randomised to Vitamin D and 1,118 randomised to placebo) was not found to significantly affect all-cause mortality, RR 1.04 (0.97-1.12, 95% CI). In subgroup analyses, participants with 25(OH)D levels <50nmol/L were found to have significant reduction in mortality when treated with Vitamin D, RR 0.93 (0.89-0.98, 95% CI) compared with those with levels >50nmol/L, RR 0.96 (0.89-1.03, 95% CI). Cholecalciferol was found to be more effective in reduction of mortality (RR 0.93, 0.89-.097, 95% CI) compared with ergocalciferol RR 0.98, (0.90-1.06, 95% CI).

Rush et al performed a systematic review and meta-analysis to evaluate the relationship between Vitamin D and premature mortality in a Scottish population (Rush et al., 2013). Nine studies were suitable for inclusion in the meta-analysis. In synthesis of results the overall effect size using adjusted HR for all cause mortality for lowest quartile of 25(OH)D was HR 1.19 (1.12-1.27, 95% CI). In age stratified analysis the pooled effect size was HR 1.12 (1.01-1.24, 95% CI) for those with a mean age <65 years versus HR 1.25 (1.14-1.36,

between all cause mortality and 25(OH)D level, HR 1.42 (1.30-1.55, 95% CI). In fully adjusted models, this remained significant, HR 1.19 (1.12-1.27, 95% CI).

A meta-analysis of eight observational studies was reported in 2014 (Schottker et al., 2014a) and looked at the relationship between Vitamin D levels and all-cause, CVD and cancer related mortality. Participant information for those aged between 50-79 years was included from each study. Results showed a non-significant level of heterogeneity between studies. The median Vitamin D levels varied between 24-62nmol/L with a median follow- up time between 4.2-15.8 years during which time 6,695 participants died. The meta- analysis revealed a pooled estimate effect for all cause mortality at 1.6 fold higher in the bottom quartile of Vitamin D compared with the top quartile.

A systematic review and meta-analysis from 2014 evaluated the effect of Vitamin D deficiency on infection rates, sepsis and mortality in critically ill patients (de Haan et al., 2014). A total of 14 studies including 9,715 participants were included. The mean 25(OH)D level was 45nmol/L, mean age 62 years and mean 30-day mortality rate was 17.5% and mean in-hospital mortality rate was 18.4%. In pooled analysis for 30-day mortality associated with Vitamin D deficiency in a subgroup of 2,572 participants was RR 1.76 (1.37-2.26, 95% CI) and the pooled in-hospital mortality, RR 1.79 (1.49-2.16, 95% CI) for 3,606 participants.

A meta-analysis reported by Garland et al analysed the strength of the inverse association between Vitamin D and age-adjusted mortality (Garland et al., 2014). Studies reporting a measure of association according to two or more categories of 25(OH)D were included with 32 studies considered eligible. Of these, 25 showed a significant inverse relationship between 25(OH)D concentration and age adjusted mortality rates. In a further two studies an inverse relationship existed but was not statistically significant. The overall age-adjusted mortality HR was 1.9 (1.6-2.2, 95% CI) p<0.001 when comparing lowest 25(OH)D levels (<22.5nmol/L) to highest levels (>125nmol/L). When 25(OH)D levels >75nmol/L were compared with levels <75nmol/L, lower levels were once again found to be associated with

higher all cause mortality, p<0.01. A pooled dose response curve showed a plateau at 25(OH)D concentrations beyond 125nmol/L.

Bjelakovc et al completed a Cochrane review in 2014 looking at Vitamin D supplementation for the prevention of mortality in adults (Bjelakovic et al., 2014). 56 RCTs with 95,286 participants, with Vitamin D compared to placebo or no intervention were included in this analysis. The age ranged from 18 to 107 years. Vitamin D was administered for a weighted mean of 4.5 years. 80% of trials reported Vitamin D levels at baseline and of these, participants in 19 trials had levels >50nmol/L and those participants in the remaining 26 trials had insufficient levels of Vitamin D (<50nmol/L). More than 8% of participants dropped out. When all 56 trials were analysed together it was found that Vitamin D appeared to be associated with reduced mortality, RR 0.97 (0.94-0.99, 95% CI). Only Vitamin D3 was found to decrease mortality when different forms of Vitamin D were analysed, RR 0.94 (0.91-0.98, 95% CI). There was no mortality benefit found with Vitamin D2, or alfacalcidol or calcitrol. Also combining Vitamin D3 with calcium increased the risk of nephrolithiasis.

Putzu et al performed a meta-analysis to evaluate the effects of Vitamin D on outcomes in critically ill adults including mortality (Putzu et al., 2017). A total of seven RCTs were included. Six studies administered Vitamin D3 (cholecalciferol) and one study administered Vitamin D2 (ergocalciferol). All studies used placebo as control. Analysis based on the trials with low risk of bias (n=5) showed that administering Vitamin D in critically ill patients was associated with reduced mortality, (31.6% in the Vitamin D group compared with 40.1% in the control group), OR 0.70 (0.50-0.98, 95% CI) p=0.04. The effect on mortality remained when all seven studies were included in the analysis.

A recent independent participant data (IPD) meta-analysis of eight prospective European studies with a total of 26,916 participants found an association with lower Vitamin D levels and all-cause, and cardiovascular mortality but no association with cancer related mortality (Gaksch et al., 2017). The mean age of participants was 61.6 years with a mean 25(OH)D concentration of 53.8nmol/L. There were 6,802 deaths in the mean follow-up period of

defined as <30nmol/L, inadequacy; 30-39.9nmol/L and 40-49.9nmol/L and sufficiency; 75- 99.9nmol/L. All cause mortality was significantly increased in those with serum 25(OH)D levels from 30-39.9nmol/L, HR 1.24 (1.07-1.42, 95% CI) and also in those with levels below 30nmol/L, HR 1.50 (1.28-1.71, 95%CI). Heterogeneity was reported as low.

Negative Studies

A meta-analysis was published by Chowdhury et al looking at the relationship between Vitamin D and mortality in observational cohort studies, along with quantifying the effects of Vitamin D supplementation on mortality when given alone and compared with placebo/no treatment in RCTs (Chowdhury et al., 2014). 95 studies were included comprising a total of 880,128 participants and 71,625 mortality outcomes. In 73 observational cohort studies, the median age of participants was 63 years and the follow-up period ranged from 0-29 years and the median baseline Vitamin D level was 51.7nmol/L. For primary prevention cohorts the pooled RR for all cause mortality for those in the lowest versus highest Vitamin D concentrations were 1.35 (1.22-1.49, 95%CI). In further analysis, for each 25nmol/L decline in Vitamin D concentration there was a 16% increase in all cause mortality. 22 RCTs were included in the analysis of Vitamin D supplementation and all-cause mortality with a follow-up period ranging from 0.38 to 6.8 years. Fourteen studies reported the effects of Vitamin D3 and the remaining eight reported the effects of Vitamin D2. The mean age of participants ranged from 56-85 years. There were a total of 2,527 all cause mortality events in intervention group versus 2,587 in the control group. The total RR was 0.8 (0.94-1.02, 95% CI), a RR of 0.89 (0.80-0.99, 95%CI) for Vitamin D3 supplementation and a RR of 1.04 (0.97-1.11, 95% CI) for Vitamin D2 supplementation.

Zheng et al performed a further meta-analysis in 2015 looking at the effect of high dose intermittent Vitamin D on fracture and overall mortality prevention in adults (Zheng et al., 2015). RCTs referring to high dose intermittent Vitamin D supplementation were included, daily dosing was excluded, the dose of Vitamin D was >100,000IU (with or without calcium supplementation) and treatment times must be greater than one month. Controls

received no treatment or calcium alone. Of the nine trials included, mortality data was available in seven, hip fracture data in four, falls data in eight and vertebral fracture data in five. Follow-up ranged from 6 months to 5 years. Mean age ranged from 77-85 years. This study showed no significant benefit from intermittent high dose Vitamin D in the risk of all-cause mortality falls and fracture risk.

6.10 Discussion

In recent years there have been many studies assessing the effects of Vitamin D on mortality. Cross-sectional and prospective studies suggest an association between Vitamin D and mortality risk, although there are inconsistencies in study findings.

Overall RCTs to date do not show a benefit in Vitamin D supplementation in mortality, although these studies were often carried out in specific populations such as ICU or post- operative patients, with the exception of the recent 2018 trail, and so results are not applicable to a more general population. Better management of patients in ICU may compensate for Vitamin D associated risk though this is obviously speculative. Meta- analyses also show mixed results without conclusive evidence to support the benefit of Vitamin D in all-cause mortality.

The variability in study results are likely due to multiple factors including limitations in individual study design; such as young study populations, participants who were not deficient in Vitamin D at baseline, short follow-up study periods for mortality data and incomplete or inadequate adjustment for confounding.

Further studies, including large prospective studies with prolonged follow-up periods, but ideally larger RCT of older adults who are deficient in Vitamin D are needed to fully evaluate the association between Vitamin D deficiency and mortality in older adults.

7.1 Introduction

There have been a number of studies published in regarding Vitamin D status and its relationship to cognitive impairment, ranging from small case control studies to cross- sectional and longitudinal studies, with some conflicting results and findings. The effects of Vitamin D on the brain and on cognition are potentially mediated through a number of mechanisms, including effects on neurotransmitters, neuro-inflammation, the Vitamin D Receptor and genetic effects.