Highlights

  • Reporting of most RCTs comparing anticoagulant versus antiplatelet medication for Venous Thromboembolism Prophylaxis is not yet optimal.

  • Higher impact factor journals publish RCTs with higher CONSORT compliance.

  • There was no statistically significant improvement regarding the CONSORT compliance of RCTs comparing anticoagulant versus antiplatelet medication for Venous Thromboembolism Prophylaxis over time.

  • Further improving the quality of RCTs and their reporting could assist health care providers to their clinical decisions, increase the clinical significance of RCTs, and direct more specifically future medical research.

  • Reporting quality of randomized controlled trials (RCTs) can be improved by adhering to the CONSORT guidelines.

Introduction

Many manuscripts are submitted to medical journals annually describing numerous methodologies. Among them, Randomized Controlled Trials (RCTs) are at the top level of the evidence hierarchy. Their study design prevents selection and confounding bias and permits blinding of participants and researchers [1]. The results of large RCTs have subsequently been translated into guidelines, as transparent and well-designed, conducted and reported RCTs are considered the gold standard research design in evaluating healthcare interventions [2].

However, in various fields of medicine, RCTs suffer from important methodological limitations. Poor reporting of RCTs impedes adequate understanding of the clinical indications and it is responsible for a great deal of avoidable waste in research [3, 4]. Readers and reviewers of published RCTs need complete, clearly written and transparent information on a study’s methodology and findings in order to assess the quality and results of a trial. Because biases can occur in all aspects of studies, poor reporting limits the reader’s appreciation of the finding’s validity and reliability. This situation may lead to an underestimation or overestimation of the true intervention effect [5].

In response to these concerns about the quality of reporting of RCTs, an international group of clinical researchers, statisticians, epidemiologists, journal editors and methodologists developed and published the CONsolidated Standards Of Reporting Trials (CONSORT) Statement in 1996 [6].

The CONSORT Statement was reviewed in 2001 and the most recent version was published in 2010. It contains 37 items (grouped according to the general format of a journal publication; namely, Title and Abstract, Methods, Results, Discussion, and Other Information such as funding) that allow the reader to understand the trial design, how it was conducted, its analysis and interpretation, and the validity of the study results [7].These items are recommended to be incorporated into an RCT. The statement provides guidance for reporting all RCTs with a focus on individually randomized, two groups, parallel trials. It is available as a guide with an explanation and examples for each item and a checklist [8] and it provides authors and editors with the means to write and publish RCTs that are as transparent and as complete as possible, allowing more effective sharing of knowledge and validation of research results by peers. It also consists of a flow diagram that displays the progress of all participants through the trial [7]. It was observed that the majority of previous RCTs had low adherence to the CONSORT statement [9,10,11,12].

Venous thromboembolism (VTE), consisting of deep vein thrombosis and pulmonary embolism, is a major and often unrecognized cause of morbidity and mortality in hospitalized and ambulatory patients [13]. Diagnosis of VTE is often difficult and massive pulmonary embolism may be the first clinical manifestation of the disease. Prevention is therefore of paramount importance [14, 15]. However, it is uncertain whether it is better to use anticoagulant or antiplatelet medication. Randomized controlled trials have been conducted in order to determine whether therapeutic anticoagulation is superior to antiplatelet prophylaxis in the prevention of VTE [16,17,18].

Although numerous publications have used the CONSORT statement to evaluate the quality of reports of RCTs in various subspecialties of medicine, to our knowledge no publication has evaluated the quality of RCTs focusing on the use of anticoagulant versus antiplatelet medication in primary and secondary prevention of deep vein thrombosis and pulmonary embolism. Therefore, the purpose of the present study was to evaluate the compliance of RCTs with the CONSORT Statement, covering a period of the last 19 years in order to assess the quality of reporting of data that are used to inform current treatment guidelines and, hence, influence clinical practice.

Materials and methods

Selection of studies

An electronic structured literature search to identify possible studies for inclusion was conducted using the following databases MEDLINE/PubMed, Cochrane library and Google Scholar. The search strategy identified reports on RCTs involving patients who received either anticoagulants or antiplatelet medications for prophylaxis of deep vein thrombosis and pulmonary embolism published within the time period January 2000 to October 2019. As a search criterion the combination of the following terms were used:((antiplatelet agents OR antiplatelet drugs OR platelet aggregation inhibitors OR platelet antiaggregants OR platelet antagonists OR aspirin OR ASA OR acetylsalicylic acid OR dipyridamole OR aggrenox OR ticagrelor OR brilique OR prasugrel OR clopidogrel OR plavix OR iscover OR ticlodipine OR ticlid OR tipidogrel OR eptifibatide) OR (anticoagulants OR anticoagulation agents OR anticoagulation drugs OR blood coagulation factor inhibitors OR indirect thrombin inhibitors OR anticoagulant factor Xa protein OR new oral anticoagulant OR novel oral anticoagulant OR noac OR heparin OR lmwh OR low molecular weight heparin OR vitamin k antagonist OR antivitamins K OR acenocoumarol OR warfarin OR coumadin OR dalteparin OR fragmin OR Enoxaparin OR bemiparin OR omnadin OR fondaparinux OR arixtra OR indraparinux OR rivaroxaban OR xarelto OR dabigatran OR pradaxa OR apixaban OR eliquis OR betrixaban OR ximelagatran OR inogatran OR bendix OR coreno OR edoxaban OR darexaban OR abciximab OR lixiana OR savaysa)) AND (deep vein thrombosis OR DVT OR venous thromboembolism OR VTE OR pulmonary embolism OR PE OR pulmonary thromboembolism OR venous thrombosis)).

In PubMed, the following filters were also used the “Randomized Controlled Trial” type of article, “English” language and “Humans” for species.

In order to determine study eligibility, first, visual inspection of the study title, then the abstract, followed by the full manuscript was conducted. Trials were eligible if they had randomly assigned human subjects to at least two treatment arms and included patients who received either anticoagulant or antiplatelet medication for the prevention of deep vein thrombosis and pulmonary embolism.

A study was defined as an RCT if the participants were assigned to interventions that were described as random, randomly allocated, randomized, or if randomization was mentioned, and if a control group was included. The control group could receive a comparator.

Exclusion criteria were animal studies, reviews, and systematic reviews, meta-analyses, non-randomized studies, follow-up studies of previously published trials, studies with crossover design, economic analyses, safety analyses, dose-comparison studies, small pilot studies, ex-vivo studies, abstracts, protocols and editorials.

Reporting assessment tool

The assessment of the adequacy of reporting were evaluated according to the revised CONSORT 2010 checklist which includes a 37-item questionnaire (http://www.consort-statement.org). The CONSORT explanation and elaboration document was used as guidance [19].

As the revised CONSORT version was published in 2010, reports were grouped in two publication periods (2000–2009) and (2010–2019) so as to test for reporting differences over time. Comparison between the two time periods can detect any improvement of reporting of CONSORT items and provide an indication in improvement of validity and quality of RCTs. The revised CONSORT version was used for all extracted articles either published before or after 2010 when the revised CONSORT version was published.

Methodological evaluation

All included articles were read in-depth. During the evaluation the followed procedures were followed:

  • All items were investigated in terms of whether they were reported and not if they actually were carried out during the trial. Namely, each item can be characterized as ‘yes’ if it is clearly and adequately reported, or ‘no’ if it is partially unclear or not reported at all.

  • Alternative responses (apart of yes or no) or unclear responses to each question were coded as negative responses; Regarding items on the CONSORT checklist 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 not; then the answer of these items were analysed accordingly.

  • When an item was reported in a different section of the trial (title, abstract, methods, results, discussion) it was considered as a positive response.

  • We separated the reported articles into two publication groups: from 2000 to 2009 and from 2010to 2019.

  • Revised CONSORT version of 2010 was used for all extracted articles either or not published before 2010 when the revised CONSORT version was published.

Compliance more than 75% with the CONSORT items was regarded as an adequate cut-off in a number of studies [20,21,22]. Τhe percentage of the studies addressed at least 75% of the 37 checklist, items overall and by time period was calculated.

Comparison among different time periods was made by using the Pearson chi-square test for trend. It was also calculated the percentage of the items per group reported in at least 75% of the articles for the 19-year period and in each time period. We ranked the included articles according to the ISI (Institute for Scientific Information) impact factor (IF) list for 2018 and we searched if high impact factor medical journals presented high compliance with the CONSORT statement. We divided articles into two groups to compare the adherence to the CONSORT statement of the articles in major IF medical journals (IF ≥ 10) with the remaining eligible papers (IF < 10). The selection of IF = 10 as the cut-off point was arbitrary. The average compliance score for high ranked versus low ranked medical journals has been calculated and a comparison has been performed using the Pearson chi-square test. The statistical analysis was made on the IBM SPSS v.21 package. The cutoff point for statistical significance was set at the two-sided 0.05 level.

Results

The process was made in five steps as can be seen in the flow diagram (Fig. 1).

Fig. 1
figure 1

Flow diagram

Full size image

The initial search retrieved 76967 titles. After duplications removed, 50891 titles were screened. 50869 studies were considered irrelevant after title and abstract evaluation. The remaining 22 articles were retrieved in full text, nine of which were found ineligible and finally 13 articles were included in qualitative analysis, requiring complete full-text evaluation. Out of the total 13 eligible trials 3articles were published before 2010 and 10 after 2010.

All CONSORT items and the frequency of adherence to the individual criterion for each of these two periods and for the combined period are shown in Table 1.

Table 1 Proportion of reporting of 37 data items in a total of 13 randomized controlled trials grouped by publication period
Full size table

In all of the time periods, only 13 of the 37 items of the checklist (35.1%) were reported by 75% or more of the studies, whereas nine items of the 37 items of the checklist (24.3%) were reported in all (100%) of the articles. These items with the respective number of articles reported and the percentage are shown in Table 2.

Table 2 CONSORT items reported by 75% or more of the articles sorted by decreasing value and percentage
Full size table

In contrast, a number of items were reported only in a small percentage of the articles in the two periods. In particular, only 31% (4 of 13) of the articles provided the identification that it is a randomized trial in the title and only one article reported where the full trial protocol can be accessed. Furthermore, the reporting of any changes to trial outcomes after the trial commenced, with reasons, was not reported at all.

The numbers and the percentages of CONSORT items reported by each article are presented in Table 3. The average CONSORT compliance score was 59.69% (38% − 83%). Only one RCT covered more than 75% of the CONSORT items (83%) [23], whereas there were 9 articles with a CONSORT compliance more than 50%.

Table 3 List of the included RCTs along with their CONSORT score
Full size table

Period effect

Improvements over time were seen for 14 of 37 CONSORT items as seen in Table 4. However, the comparison of each item showed no statistically significant differences between the two groups (p-value > 0.05). Moreover, the reporting of the type of randomization as well as the description of the mechanism used to implement the random allocation sequence was found to be inadequate in both periods.

Table 4 List of the CONSORT items which show improvement during time
Full size table

Impact of CONSORT in high-ranked journals

Of the 13 study reports that were used in the analysis, six were published in high-ranked medical journals (IF ≥ 10) and 7 in lower ranked medical journals (IF < 10). It was observed that the RCTs of high-ranked journals performed better compliance with the CONSORT Statement items, as seen in Table 5. The average compliance of the articles published in low (IF < 10) and high (IF > 10) Impact Factor journals was 45.6% and 70% respectively. There is a statistical significant difference between the two comparing groups (p-value < 0.05).

Table 5 Impact Factors 2018 of medical journals and Compliance score of relevant RCTs.
Full size table

Discussion

In this study, CONSORT 2010 was used to evaluate the reporting quality of 13 RCTs with respect to their adherence to all the 37 items of the checklist. To our knowledge, this is the first application of CONSORT analysis to RCTs testing anticoagulant versus antiplatelet medication in venous thromboembolism prophylaxis, i.e. deep vein thrombosis and pulmonary embolism prophylaxis, covering a period of the last 19 years.

Only 13 items of the 37 items of the Checklist were addressed in 75% or more of the studies. Important methodological information was also underreported like item 3b (Important changes to methods after trial commencement with reasons) which was reported in 15% of the articles, 7b (explanation of any interim analyses and stopping guidelines) which was reported in 23% of the articles, 8b (type of randomization; details of any restriction) which was reported in 31% of the articles, 9 (Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned) which was reported in 15% of the articles, 10 (who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions) which was reported in 8% of the articles, 11b (if relevant, description of the similarity of interventions) which was reported in 23% of the articles and 12b (methods for additional analyses, such as subgroup analyses and adjusted analyses) which was reported in 31% of the articles. Item 6b (any changes to trial outcomes after the trial commenced, with reasons) was the most underreported item: it was not reported in any article. However, it is quite encouraging that important items that of reporting scientific background and explanation of rationale and that of description of trial design including allocation ratio were reported in 100% of the articles.

However, it was noticed that the RCTs of high-ranked medical journals have adhered better to the CONSORT statement and this can be explained by the fact that these journals usually receive and select RCTs with the utmost quality. In our study, six from thirteen study reports were published in high-ranked medical journals (IF > 10) and had in average higher CONSORT compliance rate as their CONSORT compliance score was more than 60%.

The present study shows that articles describing RCTs involving patients receiving anticoagulant or antiplatelet medication as preventive treatment in deep vein thrombosis and pulmonary embolism do not conform entirely to the guidelines of CONSORT. It was found that essential aspects of RCTs are underreported and no article satisfied all criteria evaluated in the study.

The present study suggests that authors should be recommended to follow the CONSORT statement during the writing as the journal endorsement of the CONSORT statement might politely influence the completeness of reporting of RCTs. Pandis et al. [24] in their study concluded that the articles published in the period after the implementation of CONSORT reported more items. The CONSORT statement provides authors and editors with the means to write and publish RCTs that are as transparent and as complete as possible, allowing more effective sharing of knowledge and validation of research results by peers. Improving transparency is particularly important in the context of the replication crisis in science.

The study had several strengths. First and foremost among the strengths of the current study is that it included articles published in medical journals that clinicians can find in the PubMed database, in the Cochrane Library and in the ScienceDirect database. Besides, as the above mentioned search engines are open databases and the CONSORT statement is free, the methodology of this study is easily reproducible.

The present study, however, has some limitations. First, since the research had restricting criteria (specific period of time, written in the English language) the list of research articles found may not be exclusive, contributing to overall bias. Besides, the number of studies is low because RCTs comparing the efficacy and safety of anticoagulants with antiplatelet agents for the prevention of recurrent venous thromboembolism are still being carried out. Hence, generalizability of the findings may be limited. An imbalance, also, occurred in the amount of articles compared in the two periods.

Furthermore, the limitations of CONSORT must be considered. We used the revised CONSORT 2010 checklist for all the RCTs despite they were published before or after its publication. Besides, it was designed so as to assess the quality of reporting RCTs overall and not the actual performance of the trial procedures. Thus, a method of a trial that is not reported does not mean that it has not been performed. In addition, each item in the CONSORT Checklist carries a significant weight. For example, randomization, blinding, sample size determination, flow diagram, and registration number are important methodological items that can weigh differently in different studies; hence, reporting the whole score may not show the overall quality of the reported RCTs.

It is strongly recommended that investigators should be encouraged to adhere to the CONSORT Statement when reporting their RCTs, or even better, to emphasize the need to consider important aspects of interval validity during the planning stage of a trial. Editors of journals should also follow the recommendations. Thereby, RCTs reports based on CONSORT statement criteria can be improved specifically in the areas of methodology, results and discussions of papers and they will provide more reliable estimates about the effects of interventions and, hence, they will contribute in the development of evidence-based guidelines.

Conclusion

In conclusion, this study shows that the quality of reporting according to the CONSORT statement of most RCTs comparing anticoagulant versus antiplatelet medication for Venous Thromboembolism Prophylaxis is low.