Association of Vitamin D Supplementation with Cardiovascular Events: A Systematic Review and Meta-Analysis

Background: low vitamin D status has been associated with an increased incidence of cardiovascular events. However, whether vitamin D supplementation would reduce the incidence of cardiovascular events remains unclear. Purpose: To perform a systematic review and meta-analysis of the effect of vitamin D supplementation on the mortality and incidence of cardiovascular events. Data Sources: We searched Medline, Embase, and the Cochrane Central Register of Controlled Trials from their inception until 3 May 2022. Study Selection: Two authors searched for randomized clinical trials that reported vitamin D supplementation’s effect on cardiovascular events outcomes. Data Extraction: Two authors conducted independent data extraction. Data Synthesis: We identified 41,809 reports; after exclusions, 18 trials with a total of 70,278 participants were eligible for analysis. Vitamin D supplementation was not associated with the mortality of cardiovascular events (RR 0.96, 95% CI 0.88–1.06, I2 = 0%), the incidence of stroke (RR 1.05, 95% CI 0.92–1.20, I2 = 0%), myocardial infarction (RR 0.97, 95% CI 0.87–1.09, I2 = 0%), total cardiovascular events (RR 0.97, 95% CI 0.91–1.04, I2 = 27%), or cerebrovascular events (RR 1.01, 95% CI 0.87–1.18, I2 = 0%). Limitation: Cardiovascular events were the secondary outcome in most trials and thus, might be selectively reported. Conclusion: In this meta-analysis of randomized clinical trials, vitamin D supplementation was not associated with a lower risk of cardiovascular events than no supplementation. These findings do not support the routine use of vitamin D supplementation in general.


Introduction
Vitamin D insufficiency is highly prevalent among elderly persons worldwide [1]. A large number of observational studies have powerfully demonstrated that low level of vitamin D can negatively affect cardiovascular health and increase the incidence of cardiovascular events and mortality [2,3]. Thus, vitamin D supplementation may be a readily available, safe, and economical modality for preventing cardiovascular events [4]. However, evidence for causality from randomized controlled trials (RCTs) of the association between vitamin D supplementation and cardiovascular events is inconclusive or negative.
Recently, four systematic reviews and meta-analyses of RCTs evaluating vitamin D supplementation and cardiovascular outcomes were published [5][6][7][8]. Three reviews [5][6][7] showed no effect on cardiovascular mortality, stroke, and myocardial infarction. However, these meta-analyses typically included vitamin D and calcium co-administered trials. The calcium side-effects on cardiovascular events are uncommon but critically important, making those reviews challenging to interpret [9,10]. In parallel, the fourth review focuses on trials that utilized vitamin D as the monotherapy and suggested insufficient evidence to support vitamin D supplementation to reduce cardiovascular events. This review included only 11,841 participants, and its accuracy was limited. Subsequently, according to available evidence, the U.S. Preventive Services Task Force stated that vitamin D supplementation did not appear effective in preventing cardiovascular disease [11]. Since the publication of these systematic reviews, two additional large RCTs [12,13] have been published, supporting the potential for better-quality evidence from meta-analyses. The present systematic review and meta-analysis aimed to update the effect of vitamin D supplementation on preventing cardiovascular events.

Protocol and Guidance
We did the study following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [14]. The predetermined protocol was registered in the PROSPERO database (CRD42019119641). The study did not require ethical approval.

Eligibility Criteria
Eligible criteria were as follows: (1) studies involving participants >18 years; (2) vitamin D supplements alone at any dose. (Trials of vitamin D plus calcium (or other treatment) vs. calcium alone were considered vitamin D alone interventions); (3) studies in which placebo controls or no treatment were given to the other group; (4) reported at least one of the following outcomes of interest: cardiovascular mortality, myocardial infarction, stroke, total cardiovascular events, and cardiovascular mortality and cerebrovascular events; and (5) randomized controlled trials (RCT).
Exclusion Criteria: (1) observational studies, review articles, case reports, poster abstracts, and editorials; (2) trials of vitamin D analogs or hydroxylated vitamin D; (3) trials where all participants took vitamin D; (4) trials in pregnant or lactating women; and (5) trials of critically ill patients.

Information Sources and Search Strategy
The retrieval strategy was developed and implemented by a medical librarian (PX). Articles published before 3 May 2022, were included. A computerized search of Medline, Embase and the Cochrane Central Register of Controlled Trials databases was conducted. Language restrictions were not applied. Details of the retrieval strategy can be obtained in Supplement Materials Table S1. We screened references of critical articles and reviews for additional potentially relevant articles. For ongoing and unpublished studies, we searched relevant clinical trial registries.

Study Selection
After deleting duplicates, all titles and abstracts were independently screened to identify potentially relevant articles by two authors (P.Y.-Y. and L.Z.-R.). Subsequently, the two authors assessed the full text of potentially relevant studies. The final list of included trials was decided on the discussion between authors. Discrepancies were resolved by consensus or arbitrated by a third author (Z.Y.).

Data Collection
Two authors (P.Y.-Y. and Z.Y.) independently performed full-text assessments and extracted the following data from full articles: study characteristics, patient characteristics, and vitamin D supplementation. Discrepancies between the two authors were resolved by discussion.

Risk of Bias Assessment
The risk of bias in each trial was independently assessed by two authors (P.Y.-Y. and L.Z.-R.) using the Cochrane Collaboration's tool [15,16]. Discrepancies were resolved by consensus.

Quality of Evidence
Two authors (P.Y.-Y. and L.Z.-R.) independently used the Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) approach to rate the evidence quality of each outcome. Depending on the strength of evidence, it was classified as high, moderate, low, or very low [17]. The GRADE guidance used the domains of study design limitations, inconsistency, indirectness, publication bias, and imprecision in results. Discrepancies were resolved by consensus.

Data Synthesis
We analyzed data from the included trials using Review Manager (RevMan, version 5.4, the Nordic Cochrane Center, the Cochrane Collaboration) and R (version 4.1.2; R Project for Statistical Computing). Meanwhile, every analysis was based on the intention-to-treat approach. The meta-analysis was conducted using random-effects models despite the level of heterogeneity. The risk ratio (RR) with 95% CI was used as the summary measure. All statistical inference tests reflected a 2-sided of p < 0.05. I 2 values were calculated to evaluate variation among trials due to heterogeneity [18]. If there were more than 10 RCTs in a meta-analysis, publication bias was assessed by funnel plot techniques.

Trial Sequential Analysis
To explore whether cumulative data of included trials were sufficient to evaluate the results, we conducted a trial sequential analysis [19]. The purpose of the trial sequential analysis was to preserve the overall risk of type I error by 5% while retaining the power of 80% and achieving a 15% relative risk reduction from the intervention.

Sensitivity Analyses
To assess the robustness of our statistical results, we performed a series of sensitivity analyses: (1) excluding trials that were unknown or had a high risk of bias, (2) excluding the largest trial, (3) excluding quasi-randomized or cluster-randomized trials, (4) excluding trials with a high risk of bias in each domain, (5) using random-effects models.

Characteristics of Included Studies
The study selection process is shown in Supplementary Materials Figure S1. The systematic search identified 41,809 records initially, of which 18 trials [20-34] met inclusion criteria. Table 1 shows the characteristics of these trials included in the systematic review. In total, 70,278 participants were enrolled.

Quality of Evidence and Risk of Bias
GRADE summary findings for all outcomes are shown in Table 2. Risk-of-bias assessments are described in Supplementary Materials Figures S2 and S3. Of the 18 included trials, nine were at low risk of bias, eight were at unclear risk, and one was at high risk. The Funnel plots revealed no publication bias for all five outcomes (Supplementary Materials Figures S4-S8).

Cardiovascular Mortality
Of these, 9 RCTs with 63,227 participants (31,620 with vitamin D supplementation and 31,607 without vitamin D) were included in the meta-analysis for cardiovascular mortality. The supplementation of vitamin D was not associated with reductions in cardiovascular mortality compared without vitamin D (RR 0.96, 95% CI 0.88-1.06, I 2 = 0%; Figure 1).

Other Analyses
Prespecified subgroup analyses of all outcomes revealed no interactions between the number of patients, number of events, mean age, sex, baseline 25(OH)D, published year, type of vitamin D, daily dose equivalent of vitamin D, the timing of treatment, residential status, and length of follow-up ( Table 3). The results of the sensitivity analyses were broadly consistent with the primary analysis (Supplementary Materials Table S2).

Myocardial Infarction and Stroke
Stroke was reported in 12 trials with a total of 46,093 participants. Similarly, the supplementation with vitamin D was not related to reducing stroke compared with those not supplemented (RR 1.05, 95% CI 0.92-1.20, I 2 = 0%). For myocardial infarction, 14 studies reported 1118 events (552 with vitamin D supplementation as well as 566 without vitamin D). Supplementing with vitamin D did not reduce the risk of myocardial infarction compared with not supplementing vitamin D (RR 0.97, 95% CI 0.87-1.09, I 2 = 0%; Figure 1).

Total Cardiovascular Events and Cerebrovascular Events
A total of 39,046 participants were enrolled in seven trials reporting cardiovascular events. A similar incidence of cardiovascular events was observed between the vitamin D and no vitamin D groups (RR 0.97, 95% CI 0.91-1.04, I 2 = 27%; Figure 1. Total cerebrovascular events were similar between the two groups (RR 1.01, 95% CI 0.87-1.18, I 2 = 0%; Figure 1).

Other Analyses
Prespecified subgroup analyses of all outcomes revealed no interactions between the number of patients, number of events, mean age, sex, baseline 25(OH)D, published year, type of vitamin D, daily dose equivalent of vitamin D, the timing of treatment, residential status, and length of follow-up ( Table 3). The results of the sensitivity analyses were broadly consistent with the primary analysis (Supplementary Materials Table S2).
Trial sequential analyses of the effects of vitamin D supplementation on cardiovascular outcomes were shown in Supplementary Materials Figures S9-S13. In the trial sequential analysis, the pooled sample size exceeded the calculated optimum sample size for total cardiovascular events but not for cardiovascular mortality, stroke, myocardial infarction, and cerebrovascular events.

Discussion
In this meta-analysis of 18 RCTs with a total of 70,278, supplementation of vitamin D was not associated with reducing the risk of cardiovascular mortality, myocardial infarction, stroke, total cardiovascular events, or cerebrovascular events. No significant difference between subgroup analyses was found. Evidence of trial sequential analyses indicated that vitamin D supplementation did not decrease the relative risk of myocardial infarction and total cardiovascular events by 15%. It is unlikely that further similar trials will alter the conclusions of these trial sequential analyses. However, the evidence was not enough on the outcomes of cardiovascular mortality, stroke, myocardial infarction, and cerebrovascular events.

Comparison with Other Studies
Our findings were consistent with previous systematic reviews on this topic [5,6,8]. Low-quality evidence of a Cochrane review suggested no significant difference in cardiovascular mortality (RR 0.98, 95% CI 0.90-1.07). A meta-analysis by Bolland et al. showed negative findings with stroke (RR 1.00, 95% CI 0.88-1.13) and myocardial infarction (RR 1.04, 95% CI 0.91-1.17) [6]. The two systematic reviews included trials with mixed interventions of vitamin D supplementation combined with calcium supplementation, which is problematic because calcium has been proven to be associated with an increased risk of cardiovascular events [38][39][40]. In the meta-analysis by Ford et al., evidence of supplementation with vitamin D alone does not appear to protect against myocardial infarction (RR 0.95, 95% CI 0.82-1.10) or stroke (RR 1.08, 95% CI 0.91-1.29) [8].
In this study, we provided a different method compared with recent meta-analyses. Our study additionally included two large RCTs, which take up 34% of the participants in this study, providing more robust results of these relationships. Moreover, our study used different selection criteria: unlike these meta-analyses, we did not include more than ten trials of vitamin D co-administered with calcium, six trials [41][42][43][44][45][46][47] of vitamin D analogs, or hydroxylated vitamin D, and one trial by Sato et al. [48] which was retracted in 2017. Furthermore, we quantified a new outcome: total cardiovascular events.

Implications for Future Research
Though the available evidence did not show meaningful effects of vitamin D supplementation on cardiovascular outcomes, there have been several points for future research. First, the baseline 25(OH)D concentrations of included trials may be too high. Cardiovascular events appear when the serum 25(OH)D levels are below 37 nmol/L [4]. However, only three trials [30,32,34] involving 28,520 participants reported a mean baseline 25(OH)D level of less than 37 nmol/L. The future RCTs enrolling populations with lower baseline 25(OH)D might produce different results. Second, cardiovascular disease is a chronic disease. The follow-up time of patients included in the study is only a few years. Through the different blood sampling times after administration, we can see that the vitamin D level of blood samples will decrease with the extension of sampling time. Positive results may be obtained if the drug is continuously administered for a long time and the follow-up time is extended [28]. Third, eligible trials may be carried out in the wrong populations. Like RCTs of aspirin used to prevent cardiovascular events, the cardiovascular benefits are mainly observed in participants with a high risk of cardiovascular events, such as patients with a prior history of atherosclerotic cardiovascular disease [49]. In this study, however, most participants came from primary prevention trials. Further vitamin D trials aiming to prevent cardiovascular events should design through risk factor optimization (e.g., secondary prevention).

Strengths and Limitations
We conducted the present systematic review based on a protocol published in the PROSPERO database, which followed a rigorous methodological approach derived from the Cochrane Handbook. Additionally, we assessed the quality of evidence with GRADE and the minimum information size by using trial sequential analysis. Compared with previous reviews [5,6,8], the strengths of this study included a lower degree of heterogeneity, higher quality of evidence, and large sample size, which was adequate to assess rare outcomes, as confirmed by the trial sequential analysis. Some articles [9,10] have shown that the side effects of calcium on cardiovascular events are extremely important, Previous meta-analyses usually include trials of vitamin D and calcium combinations, so it is difficult to explain the role of vitamin D. We aim to give a clearer answer of vitamin D in the occurrence of cardiovascular disease.
We are, however, aware of several limitations. First, our study focused on published trials reporting cardiovascular events. In contrast, these results were not reported in many vitamin D supplementation studies, resulting in selective reporting bias.
Second, the relatively short duration of vitamin D supplementation and follow-up left it challenging to detect benefits in preventing cardiovascular events.
Third, the 15 included trials were heterogeneous in terms of vitamin D dose, duration of treatment, comorbid conditions, population characteristics, and most important, baseline 25(OH) D level. Outcomes definitions among trials differed depending on contemporary consensus definitions, contributing to clinical heterogeneity.

Conclusions
This meta-analysis of randomized clinical trials suggests that vitamin D supplementation is not associated with a decrease in cardiovascular events. There is no evidence that the routine use of these supplements is effective in preventing cardiovascular events.