This systematic review and meta-analysis protocol is prepared in accordance with the PRISMA-P (Preferred Reporting Items for Systematic review and Meta-Analysis Protocols) 2015 checklist [25] and the population, intervention, comparison, and outcome (PICO) guideline [26]. The protocol for this systematic review and meta-analysis is registered with the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY202260022) and is reported according to the PRISMA-P (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols) checklist (Multimedia Appendix 1).

This proposed study aimed to evaluate the effect of vitamin D on markers of inflammation and lipid profile among adult patients with T2D.

The research question is “Can supplementation with vitamin D improve inflammation and ameliorate dyslipidemia in T2D?” This protocol will subscribe to PICO guidelines [26]. Participants are adult patients with T2D. The intervention will be vitamin D supplementations of any dose, and the comparators are placebo or healthy patients without treatment.

The outcomes are inflammation and dyslipidemia, and the study design includes RCTs.

The inclusion and exclusion criteria will be carried out following our PICO guidelines. The review will include RCTs that have focused on adult patients with T2D. There will be no language limitations, and all studies published from inception until December 30, 2022, will be considered. The reviews, books, letters, and animal studies experimenting with diabetes will be excluded. Furthermore, cross-sectional studies, cohort studies, case-control studies, and case reports will all be excluded.

Reviewer 1 (RM) and reviewer 2 (KM) will independently conduct a comprehensive literature search through different databases (eg, Embase, Google Scholar, Web of Science, PubMed, and Science Direct). The following keywords (Medical Subject Headings terms), including their synonyms, will be applied in all databases: “Vitamin D,” “calciferol,” “Dihydroxycholecalciferol,” and “type 2 diabetes mellitus.” Furthermore, the evidence will be sought from RCTs published from inception until December 30, 2022, with no restriction regarding publication language. The preliminary search strategy is presented in Table 1 and Table 2.

The search will be conducted independently by 2 reviewers (RM and KM) to ensure accuracy and consistency based on the eligibility criteria. Firstly, keywords, titles, and abstracts will be screened, followed by retrieval of the full study. Mendeley desktop reference manager (version 1.19.4) will be used to save studies retrieved through web-based database search or manual reference screening. All studies that meet the inclusion criteria will be subjected to a data collection process and quality assessment. For any discrepancy, the 2 reviewers will conclude by discussing and reevaluating the study in dispute. The Shiny app [27] will be used to create a PRISMA flow diagram showing study selection.

The data will be extracted following the eligibility criteria. The 2 reviewers RM and KM will conduct data extraction independently and then compare the spreadsheet; in case of disagreement, they will discuss and review the study again. The information sought from each study for extraction will include the author and year of publication, the country where the study was conducted, sample size, participants’ gender and age, form of vitamin D and its dosage, duration of treatment, lipid profiles, and pro-inflammatory markers. Data will be confirmed by contacting investigators if insufficient information has been provided in the published article.

The risk of bias will be assessed using the methods recommended in the Cochrane risk of bias assessment tool [28]; assessment includes 6 domains; selection, attention, reporting, performance, attrition, and other bias. Judgments will be used to determine whether the bias is a high-risk or low-risk bias, is unclear due to insufficient information, or is a potential bias. The Jadad guideline will be used to assess the quality of individual studies [29]. The author (RM) and supervisor (KM) will discuss and resolve any bias and come to a consensus should there be any disagreement.

Following initial data analysis, we will evaluate the level of heterogeneity based on I² [30] and Q tests. If the I² is moderate (I²≥ 50%), we will explore the source of heterogeneity by performing subgroup analysis [31]. This will be based on the form of vitamin D, its dosage, BMI, study quality, the origin of the study, age, and the duration of the intervention.

Publication bias among the included studies will be assessed and visualized graphically through funnel plots [32] if there are more than 10 studies included in the meta-analysis. In case of no publication bias, the funnel plot will form a symmetrical (ie, inverted) funnel. In contrast, if bias is present due to smaller studies without statistically significant effects, the funnel plot will show an asymmetrical shape.

Sensitivity analyses will be performed to evaluate the robustness and stability [33] of our effect size by using one study exclusion approach at a time based on the study with a high risk of bias, small sample size, or the largest sample size.

We will evaluate the certainty of our evidence by using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) approach [34] using the web-based GRADEprofiler. This approach takes into consideration the risk of bias, imprecision, inconsistency, indirectness, publication bias, and effect size. The evidence will be rated as low, moderate, or high based on 2 independent reviewers’ ratings. The overall findings will be presented in a summary of the findings table.

Statistical analysis will be conducted using the Review Manager (RevMan; version 5.4; Cochrane Collaboration). Each study reviewer (RM and KM) will independently extract data as mean, SD, and sample size in each group. Where mean and SD values are not reported, the estimation will be made using a web-based calculator by Hozo et al [35]. The forest plots will be used to determine the pooled estimates and classify the level of heterogeneity based on I² and P values [36]. The estimated pool will be reported as either mean difference or standardized mean difference and 95% CIs.

Furthermore, a chi-square test (χ²) and I² will assess the heterogeneity among the included studies [37,38]. An I² of 0% to 50% will be considered low to minimal heterogeneity, while an I² above 50% will be considered moderate heterogeneity.

Consistently, sources of heterogeneity will be explored through subgroup analysis, and sensitivity analysis will be used to assess the stability of the re-analyzed pool estimates. P<.05 will be considered statistically significant. In case meta-analysis cannot be performed due to insufficient data or the number of studies, qualitative synthesis will be performed [39].

There will be no involvement from patients and the public in this study. This study synthesizes data from studies that were already peer-reviewed and published. Thus, no ethical approval is required.