Introduction

The prevalence of Autoimmune thyroid diseases (AITDs) is about five percent which rendersthem the most common amongautoimmune disorders with a continuing rise in incidence. Thefemale populationis at a greater riskof developing thyroid autoimmunity than men [1]. The most common AITDs are Hashimoto thyroiditis (HT), Graves’ disease (GD) among the general population and post-partum thyroiditis (PPT) in pregnant women.AITDare caused by multiple factors, involving both environmentaland genetic factors [2,3,4].

Vitamin D is a secosteroidal hormone precursor and has been identified as a key hormone inthe musculoskeletal, nervous system and insulin sensitivity [5,6,7]. Several studies have reported a low vitamin D status in AITD, indicating an association between vitamin D deficiency and thyroid autoimmunity [8,9,10,11,12,13]. On the other hand, a small number of studies,showed no significant association between AITDs and vitamin D deficiency [14,15,16,17]. These pieces of evidence led several researchers to examine the effectiveness of vitamin D supplementation in the prevention/treatment of this group of conditions [18, 19]. The results are conflicting, so the potential of vitamin D in thyroid diseases treatmentneeds to be clarified.

Double-blind RCTs are considered to be the highest ranked mean of evidence-based medicine and their results are crucial in the formulation of the therapeutic guidelines [20]. RCTs represent better the whole strategy and philosophy of the research [21].

Readers have access to a plethora of articles, so there is a need for a tool to assess the guidance of RCTs [22].

In 1996, an international group of experts created the CONSORT (Consolidated Standards of Reporting Trials) Statement [23]. Two revisions followed in 2001 and 2010with detailed explanation and elaboration documents [24, 25]. This statement is an evidence-based set of advice, including a checklist of 37 items and a flow diagram whose reporting ensures the avoidance of failing to include important information [25]. For that reason, an increasing number of journals endorse compliance with the CONSORT statement to improve reporting standards [26].

The quality of RCTshas beeninvestigated in a variety of specialties [27,28,29,30,31]. Our team, in a previous study concerning anticoagulant versus antiplatelet medication for venous thromboembolism prophylaxis, the average CONSORT compliance score was found to be 59.69% (38–83%). Only one RCT achievedmore than 75% of the CONSORT items (83%) [32].

To our knowledge, no published study has evaluated the quality of RCTs for vitamin D supplement in thyroidautoimmunity based on the CONSORT statement. The most recent study published in December 2021 was a meta-analysis focusing on cases of Hashimoto disease and the evaluation was conducted using the Cochrane CollaborationRisk of Bias tool Statistical analysis [19].

The purpose of this study is to evaluate the reporting quality of RCTsfor vitamin D supplementation in autoimmune thyroid disorders according to Consortstatement covering a period from January 2011, onwards following the release of the updated CONSORT 2010 guidelines in March 2010, until December 31st, 2021.

Methods

Data sources and search strategies

An electronic structured literature search was organized using the following databases MEDLINE/PubMed, Cochrane library and Google Scholar. We attempted to identify relevant RCTs published within the time period from January 2011 onwards following the release of the updated CONSORT 2010 guidelines in March 2010, until December 31st, 2021.

The implemented combination of the following terms is reproduced:

(((((“Vitamin D”[Mesh] OR “Ergocalciferols”[Mesh] OR “Vitamin D Response Element”[Mesh] OR “Vitamin D-Binding Protein”[Mesh] OR “Vitamin D Deficiency”[Mesh] OR “Receptors, Calcitriol”[Mesh] OR “Vitamin D3 24-Hydroxylase”[Mesh] OR “vitamin D-binding protein-macrophage activating factor” [Supplementary Concept] OR “Cholecalciferol”[Mesh] OR “MED4 protein, human” [Supplementary Concept] OR “vitamin D binding protein 2, primate” [Supplementary Concept] OR “vitamin D binding protein 1, primate” [Supplementary Concept] OR “vitamin D response element-binding protein 2” [Supplementary Concept] OR “vitamin D 1-alpha hydroxylase” [Supplementary Concept] OR “vitamin D3 glucosiduronate” [Supplementary Concept]) OR (“Calcitriol”[Mesh] OR “25-O-ethyl-calcitriol” [Supplementary Concept] OR “22-dehydro-1,25-dihydroxy-24-dihomovitamin D3” [Supplementary Concept] OR “24,24-difluoro-1,25-dihydroxy-26,27-dimethylvitamin D3” [Supplementary Concept] OR “1,25-dihydroxyvitamin D3-23,26-lactol” [Supplementary Concept] OR “Vitamin D supplementation”)) AND (“Hashimoto Disease”[Mesh] OR “Hypothyroidism, Autoimmune” [Supplementary Concept])) OR (“Thyroiditis”[Mesh] OR “Postpartum Thyroiditis”[Mesh] OR “Thyroiditis, Autoimmune”[Mesh] OR “Thyroiditis, Chronic” [Supplementary Concept])) OR “Hypothyroidism”[Mesh]) OR (“anti-thyroid autoantibodies” [Supplementary Concept] OR “Autoantibodies”[Mesh] OR Graves’ disease OR Hyperthyroidism OR postpartum thyroiditis).

In order to restrict the search in PubMed, the “Randomized Controlled Trial”filter for study type, the “English” filter for language and lastly the “Humans”species filterwere used.

Eligibility of studies

Inclusioncriteria:

  • Published from January 1st 2011 until December 31st, 2021

  • Parallel groupRCTs

  • One group was randomized to receive calcitriol or other Vitamin D analogs

  • They recruit patients with autoimmune thyroid disease

Exclusion criteria:

  • Non-randomized studies

  • Reviews

  • Pilot studies

  • Non-human studies

  • Studies with crossover design

  • Economic analyses

  • Small pilot studies

  • Study protocols

  • Articles not in English

Reporting assessment tool

The revised CONSORT checklist was used, which includes a 37-item questionnaire [25]. The CONSORT elaboration and explanationstatementguided the process [33]. CONSORT offersrecommendations for eachpart of an RCT, such as title, introduction, methods, results, discussion or other information, coveringall aspects of an optimal clinical trial [34].

The immediate period (until December 31st2010) following the publication of the latest revision of CONSORT statement (Mar 2010) was not included in the assessment. This decision was made to provide authors with enough time to abide by the revisedrecommendations.

Methodological evaluation

During the evaluation process, the selected articles were reviewed one by one according to the revised CONSORT version of 2010.Each item was appraised one of the following scores: ‘yes’ 1 point when adequately reported, ‘no’ or ‘unclear’ 0 points when inadequately reported or absent. When an item was reported in a different section of the trial, it was considered as a positive response.Regarding items on the CONSORT check- list with statements such as “When applicable” (7b), “If done” (11a)or “If relevant” (11b) they were checked as “non-applicable” if the answer was definite yes or no; then the answer of these items wasanalyzed accordingly.This resulted in a score range from 0 to 37.

Additional information included publication year,journal ranking [5-year Impact Factor (IF)published in 2020 by Clarivate Analytics via Journal Citation Reports], reporting of funding sources, number of authors, continent of first author, sample size, number of citations.

Outcome measures and Statistical analysis

The period from January 2011until December 31st, 2021 was assessed.It was decided that the remaining part of 2020 would not be evaluated in order to provide a sufficient time for authors to conform with the newest recommendations.The primary outcome measure was the mean CONSORT adherence of the included RCTs.Compliance above 75% with the CONSORT items was regarded as cut-off [31, 32]. We investigated the adherence of each itemseparately and the existence of possible determinant factors were also investigated.

All parameters were analyzed as categorical variables:IF (<2.86, ≥2.86 based on the median of our sample), sample size (≥82, <82 based on the median of our sample), citations (≥ 5, <5 based on the median of our sample), number of authors(≥7, <7 based on the median of our sample), funding source(yes/no), Covid-19 pandemic(earlier/in the course of). Pearson’s chi squared test (or Fisher’s exact test) was used for univariate analysis.A relaxed p-value of 0.20 was established arbitrary as a cut-off value in order to enter the binary logistic regression. A strict P value of 0.05 was set to be important for the multivariate analysis. Odds ratios (ORs), 95% confidence intervals (95% Cls) and P value are presented. An additional analysis was performed in order to examine a possible linear correlation between IF and reporting quality.SPSS v.26 package was used for statistical analysis.

Ethical view

No approval from any Ethical committee was sought, since this study analyzed existing data from publicly available sources.

Results

Initially, 8196 studies were obtained through the selected databases(Pubmed, Cochrane library and Google scholar). After removal of duplicated items, 6467 records were remained. Following evaluation of title and abstracts, 20 potentially eligible articles were identified. Finally, the full-text of these studies were examined and 13 studies were included in further assessment. Fig. 1 describes the five steps of the search strategy in a PRISMA flow diagram.

Fig. 1
figure 1

Flow diagram

Full size image

CONSORTadherence

The mean compliance to the CONSORT statement for RCTs was calculated at 61.15% with SD = 14.86% (Median = 62%, minimum & maximum adherence were 38% and 86% respectively). Among the studies, only 3 (23%) achieved a good reporting quality (≥75% of the items), while 10 (77%) presented with inadequate reporting (< 75% of the items). The mean proportion of adherence to the CONSORT statement for each study are presented in Table 1 and Fig. 2.

Table 1 List of randomized controlled trials along with the CONSORT (Consolidated Standards of Reporting Trials) score
Full size table
Fig. 2
figure 2

Distribution of the total CONSORT (Consolidated Standards of Reporting Trials) scores of the 13 studies

Full size image

Adherence per CONSORT item was estimated (Table 2, Fig. 3). Specially, 5 of the 37 items of the checklist (13.5%) were reported in all (100%) of the articles and only 16 of the 37 items of the checklist (43.2%) were reported by 75% or more of the studies. Among methodological items, randomization process (items 8a and 8b) and blinding (items 10 and 11a) were mainly inadequately reported. In contrast, a structured abstract (item 1b) was reported adequately (77%) among the studies and is considered of crucial importance, taking into account that most readers base their decision to acquire or not a full text on its abstract.

Table 2 Adherence per CONSORT (Consolidated Standards of Reporting Trials) item
Full size table
Fig. 3
figure 3

Correlation (scatter-plot) between reporting quality and IF (Impact Factor)

Full size image

Determinants of reporting quality

According to univariate analysis high IF of journal was the only with superior statistical significance (p < 0.05). Large sample size, great number of authors, existence of funding source was all associated with an adequate p value (p < 0.20) in order to enter binary logistic regression. Results are summarized at Table 3.

Table 3 Univariate analysis of possible determinants of reporting quality
Full size table

The four predictors of the univariate analysis were entered into a multivariable model. None of these was associated significantly with adequate reporting. Particularly, the journal impact factor (p = 0.150) failed to demonstrate significant effect, whereas the effect of number of citations (p = 0.650), sample size (p = 0.161) and number of authors (p = 0.892) persisted inadequately. Results of binary logistic regression are illustrated at Table 4.

Table 4 Multivariate analysis of possible determinants of reporting quality
Full size table

Finally, an additional analysis (Fig. 3) discovered the occurrence of satisfactory positive linear corellation between reporting quality and IF

[Pearson’s correlation (r = 0.740, p = 0.004)].

Discussion

CONSORT adherence

The present study evaluated the reporting quality of RCTsthat examined the effect of vitamin D supplement in thyroid autoimmunity according to 2010 CONSORT statement. The conclusion is that the overall CONSORT adherence is far from optimal, with the mean compliance equal to 61.15%. The number and sample size of RCTs based on our subjectis smaller than that of other endocrinological diseases probably due to rising interest of researchers in the last decade [35,36,37]. We collected and analyzed 13 articles referring to 1174 randomized participants. Only three of them showed compliance above 75%.

Furthermore, 16 of 37 checklist items (43.2%) were addressed by 75% or more. The report of crucial methodological characteristics like randomization (item 9: allocation concealment method—38%; item 10: implementation—7.7%) and blinding (item 11a: who was blinded—38%) was found to be suboptimal. Unclear or absent description of randomization and blinding degrades RCTs due to complicated risk of bias [38]. Also, inadequate explanation of adverse effects in their articles (item 19: harms or unintended effects—23%) will probably misguide the medical approach of the physicians and may even give wrong advice to their patients. Item 14b (Why the trial ended or was stopped—0%) was the least reported item.On the contrary, it is hopeful that significant items such as trial design (item 3a – 92%) and report of the interventions for each group (item 5–85%) achieved a strong representation.

Determinants of reporting quality

Univariate analysis suggested thatlarger sample size,higher number of authors,the presence of funding wereall associated with abetter reporting quality but not statistically significant. Only RCTs of high-ranked medical journals showed superior adherence to the CONSORT statement giving statistically significant results (p < 0.05) and additionally a strong linear correlation (r = 0.740). IF was previously studied and a number of studies demonstrated an important association between IF and reporting quality [28, 29, 32, 34]. This is because journals with a higher IF have more strictrules for the publication of studies.

Despite the indications of univariate analysis, logistic regressionof possible determining factors canceled the previous effectof impact factor in the reporting quality of RCTs.In any case, we have to make referenceto commercial funding. It is crucial that our study comes in harmony with previous showing non-significant impact in scientific information [28, 38,39,40].

In one hand, the reporting quality of RCTs for Vitamin D supplementation in autoimmune thyroid disorders appeared not to be affected by Covid-19 pandemic. On the other hand, several fields of research are being lured away from their main area of interest to the pandemic, including the possibility that other health topics are ignored or not done properly [41]. It is important to highlight that literature search involved three databases: PubMed/MEDLINE, Cochrane Library and Google scholar creating a source of 8196 studies and increasing the overall efficacy of search strategy. As is well known, CONSORT statement is free and the methodology of current study is easily accessible.

However, our results must be interpreted with skepticism and some points need to be addressed. Vitamin Dsupplementation in autoimmune thyroid disorders is not a field well studied by the research community. As a result, the number of RCTs we analyzed, is quiet low. Moreover, articles not published in English orreleased beyond the time limit were excluded. The researcher was not blinded to journal and all items were rated as equal. So, the methodological analysis becomes more susceptible to subjectivity as certain items like flow diagram, randomization and blinding are more important than others.

Considering the increasing number of publications, investigators are recommended to report their RCTs according to the CONSORT statement and the CONSORT statement should be implemented in the editorial process. The improvement of the quality of RCTs could assist to reach more conclusive results, to minimize biased conclusions, to elucidate better the clinical significance of RCTs, and to direct more specifically future medical research.

Conclusion

To the best of our knowledge, the present study is the first to evaluate the reporting quality of RCTs for Vitamin D supplementation in autoimmune thyroid disorders according to 2010 CONSORT statement. The results we obtained were discouraging. It is our feeling that our subject is generally badly reported. Taking into account the controversial role of VitD supplementation on the prevention and/or treatment of AITD and the increasing number of publications, we concluded that the compliance with CONSORT guidelines becomes essential in order to provide more reliable and consistent answers to scientific question.