#AMSTAR2Analysis

Risk of bias assessment with AMSTAR 2.0 for: Thiazide diuretics alone or combined with potassium-sparing diuretics to treat hypertension: a systematic review and network meta-analysis of randomized controlled trials

Written by AccuScript

July 14, 2023

Based on the AMSTAR 2.0 assessment, the systematic literature review demonstrates medium quality. It scores well on several parameters such as incorporating PICO components in the research questions, having clear review methods established before the review began, and using a comprehensive literature search strategy. The selection of study designs for inclusion is explicit, and the study selection and data extraction were performed in duplicate. Excluded studies were justified and included studies were adequately detailed.

However, certain elements could be improved. The authors did not specify how the results from bias assessment were used in the synthesis. The review lacks information on the sources of funding for the included studies, and there’s no discussion on the potential impact of bias in individual studies on the meta-analysis results. The interpretation of the review’s results doesn’t account for the risk of bias in individual studies. Even though heterogeneity was statistically considered, the authors didn’t explain it adequately, and while an investigation of publication bias was conducted, its potential impact on the review’s results was not discussed.

Despite these shortcomings, the review does a good job of reporting potential conflicts of interest, enhancing its transparency. Overall, the systematic review shows a commendable effort but also highlights areas for improvement, particularly around bias and heterogeneity discussion.

Based on the AMSTAR 2.0 assessment, this systematic literature review is of medium quality due to some significant shortcomings, despite many areas of strength.

Strengths include:

1. The research questions and inclusion criteria adequately incorporated PICO components.

2. The review methods were established prior to the review, with justified deviations from the protocol.

3. The authors explicitly explained their choice of study designs for inclusion.

4. They implemented a comprehensive literature search strategy and performed study selection in duplicate.

5. Data extraction was partially performed in duplicate, although further details would have been appreciated.

6. Excluded studies were listed and justified, and included studies were described in adequate detail.

7. A satisfactory technique was used for assessing the risk of bias in individual studies, even though the report didn’t explain how these results were used in the synthesis.

8. Appropriate methods were used for statistical combination of results in the meta-analysis.

Areas of weakness include:

1. There was no report on the sources of funding for the included studies.

2. The authors didn’t discuss the potential impact of risk of bias in individual studies on the results of the meta-analysis or other evidence synthesis.

3. The risk of bias in individual studies was not accounted for when interpreting/discussing the review’s results.

4. Although heterogeneity was statistically considered, the authors didn’t provide a satisfactory explanation for it.

5. Despite performing an adequate investigation of publication bias, its likely impact on the review’s results was not discussed.

Improvements could be made by explicitly addressing how bias, both within the included studies and potential publication bias, is handled in the analysis and interpreted in the results. Detailing sources of funding for included studies and better discussing potential sources and implications of heterogeneity would also increase the review’s quality.

Assessment of methodological quality using Amstar 2.0

DomainAssessmentJustification
Did the research questions and inclusion criteria for the review include the components of PICO?YesThe study’s research questions and inclusion criteria incorporate the PICO components adequately.
Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol?YesThe review methods were established prior to the conduct of the review as demonstrated in the text “The protocol of this network meta-analysis was prospectively registered at the PROSPERO database (CRD42018118492) and deposited in a public repository (the Open Science Framework)”.
Did the review authors explain their selection of the study designs for inclusion in the review?YesThe authors of this systematic review have clearly explained their selection of study designs for inclusion in the review. The review was restricted to double-blind placebo or active-controlled randomized controlled trials (RCTs) with a follow-up period between 3 and 52 weeks.
Did the review authors use a comprehensive literature search strategy?YesThe authors thoroughly documented their search strategy.
Did the review authors perform study selection in duplicate?YesThe authors did perform study selection in duplicate as per the provided instructions.
Did the review authors perform data extraction in duplicate?Partial YesThe authors performed data extraction in duplicate, but does not explicitly mention the details about how the data extraction process was performed, nor does it mention the use of a Kappa score or any other specific measure of inter-rater reliability.
Did the review authors provide a list of excluded studies and justify the exclusions?YesThe authors provided a list of excluded studies and justified the exclusions, demonstrating transparency and adherence to systematic review protocols.
Did the review authors describe the included studies in adequate detail?YesThe authors provided a sufficient level of detail about the included studies and their methodology. Overall, there is sufficient detail provided to allow the reader to assess the generalizability of the results.
Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review?Yes, but it would have been helpful if they had stated how the bias assessment results were used in the synthesisThe authors used the Cochrane risk of bias tool and met most of the criteria for assessing bias. However, they did not explicitly state how the bias assessment results were used in the synthesis.
Did the review authors report on the sources of funding for the studies included in the review?NoThe authors of this systematic review did not report on the sources of funding for the studies included in the review. No mention of funding sources was found in the materials and methods section nor in the results. There is also no discussion regarding the potential impact of funding sources on the outcomes of the studies.
If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results?YesThe authors used appropriate methods for the statistical combination of results. The authors have used both frequentist and Bayesian statistical techniques. In the frequentist approach, they used random-effects meta-analysis models for direct pairwise evidence summarization.
If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis?NoThe systematic review authors did assess the risk of bias in individual studies included in the meta-analysis. However, there is no specific mention in the provided text about how the authors assessed the potential impact of risk of bias on the results of the meta-analysis or other evidence synthesis.
Did the review authors account for RoB in individual studies when interpreting/discussing the results of the review?NoThe authors did acknowledge that “unclear risk of bias was frequent,” there is no explicit discussion in the provided text on how this risk of bias might impact the findings of their review, nor a discussion on the limitations of the evidence included. This could be considered a limitation in their interpretation and discussion of the review’s results.
Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review?NoDespite having considered heterogeneity statistically, the authors haven’t adequately described how they explored and explained the potential sources of heterogeneity in their systematic review. The authors should have provided a detailed discussion of the possible sources of heterogeneity and the results of any subgroup or sensitivity analyses conducted to investigate these.
If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review?NoThe authors of the systematic review did perform an adequate investigation of publication bias, which is also commonly referred to as small-study bias. They used a common statistical tool; funnel plots, to examine this bias. The authors have not discussed the likely impact of publication bias on the results of the review. They need to include this discussion to fully address this domain.
Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review?YesThe authors reported no potential sources of conflict of interest. The document includes a conflict of interest statement. The do state that “this work was partially supported by the Coordination for the Improvement of Higher Education Personnel (CAPES), Brazil – Finance Code 001 for Graduate Program in Cardiology and Cardiovascular Sciences, School of Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil”

Detailed assessment of Amstar 2.0 domains

Domain 1: Did the research questions and inclusion criteria for the review include the components of PICO?

Based on the information provided, the research questions and inclusion criteria for this systematic review appear to include the components of the PICO framework:

1. Population: The population is clearly identified as adults with primary hypertension.

2. Intervention: The interventions of interest are thoroughly described as thiazide diuretics alone or in combination with a potassium-sparing diuretic.

3. Comparison: The study mentions both double-blind placebo and active-controlled randomized controlled trials (RCTs), indicating the existence of a control group.

4. Outcome: The primary outcome is specified as the mean difference in office systolic blood pressure (SBP), with secondary outcomes including the mean difference in various biochemical abnormalities and the incidence of non-melanoma skin cancer.

Thus, the assessment for this domain would be ‘Yes’ since all four elements of the PICO framework (Population, Intervention, Comparison, and Outcome) are well-described in the report.

Domain 2: Did the report of the review contain an explicit statement that the review methods were established prior to the conduct of the review and did the report justify any significant deviations from the protocol?

Assessment: Yes

The review methods were established prior to the conduct of the review as demonstrated in the text “The protocol of this network meta-analysis was prospectively registered at the PROSPERO database (CRD42018118492), published [25] and deposited in a public repository (the Open Science Framework)”. This explicitly shows that the methods were planned and registered in advance of the study, indicating that a structured approach was used to guide the review process. Furthermore, the use of the PROSPERO database signifies an independent verification of the review protocol, reinforcing the validity of the process. The authors have also deposited the protocol in an open repository (the Open Science Framework), which further strengthens the verification of the protocol.

The document does not indicate any significant deviations from the protocol, which implies that the review was conducted as planned. Any changes that may have occurred were not explicitly justified in the text provided.

In summary, the review was carried out with a pre-established protocol that was registered in a recognized database and a public repository, meeting the requirements of having both an explicit statement that the review methods were established prior to the conduct of the review, and independent verification of the protocol.

Domain 3: Did the review authors explain their selection of the study designs for inclusion in the review?

The authors of this systematic review have clearly explained their selection of study designs for inclusion in the review. The review was restricted to double-blind placebo or active-controlled randomized controlled trials (RCTs) with a follow-up period between 3 and 52 weeks. Crossover studies were included, given that there was at least a 2-week washout period between study phases. The systematic review was also conducted and reported following the PRISMA-NMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Network Meta-analysis) guidelines.

While the authors did not include non-randomized studies, they justified this decision implicitly by adhering to the rigorous methodological standards of the PRISMA guidelines, which emphasize the inclusion of RCTs due to their high level of internal validity.

The explicit statement of inclusion criteria, protocol registration, and adherence to a reputable guideline like PRISMA all suggest a strong strategy for selecting study types. Therefore, my assessment of this domain is ‘Yes’. The authors have clearly explained and justified their selection of study designs.

Domain 4: Did the review authors use a comprehensive literature search strategy?

Based on the provided text from the original article, the authors of this systematic review used a comprehensive literature search strategy:

1. They searched at least two bibliographic databases: the authors searched in the Cochrane Central Register of Controlled Trials, PubMed/MEDLINE, Embase, Web of Science, Scopus, LILACS, and ClinicalTrials.gov. They also sought results in non-indexed journals or other reporting forms via the Educational Resources Information Center (ERIC [ProQuest]).

2. They provided information on the years and databases examined: the authors mentioned that their search covered the period from inception to 15 September 2021, with no language restrictions.

3. They listed the key words and/or MESH terms: the authors used specific keywords and Medline subject heading (MeSH) terms to improve their search strategy’s sensitivity. Terms included the names of the specific drugs of interest and related concepts like ‘thiazide diuretics’, ‘inhibitor of the epithelial sodium channel’, ‘potassium sparing diuretic’, and ‘hypertension’.

4. The full search strategy is available on request: the authors state that comprehensive search strategies are provided in a supplement to the article.

5. They supplemented their searches by checking published reviews, specialized registers, experts in the field, and reviewing reference lists: the authors looked at the reference lists of other studies to find additional relevant works. They also contacted authors to seek any potential unpublished outcomes.

6. Publications in all relevant languages were sought: the authors did not restrict their search based on language.

7. They searched the grey literature where necessary: the authors searched ClinicalTrials.gov for possible results in unpublished studies and the Educational Resources Information Center (ERIC [ProQuest]) for results in non-indexed journals or other reporting forms.

Given these factors, I would assess this domain as “Yes”, the authors did use a comprehensive literature search strategy.

Domain 5: Did the review authors perform study selection in duplicate?

Based on the text provided, the authors of the systematic review did perform study selection in duplicate. This is demonstrated in the section “Study selection”, which mentions that “Titles and abstracts were independently selected by pairs of independent reviewers using the liberal accelerated approach.” Furthermore, the text confirms that when disagreements arose between the paired reviewers, these were resolved by consensus or by the intervention of a third reviewer.

Hence, my assessment is ‘Yes’, the authors did perform study selection in duplicate as per the provided instructions. The relevant justification from the article is as follows:

“Titles and abstracts were independently selected by pairs of independent reviewers using the liberal accelerated approach. Disagreements were resolved by consensus or by a third reviewer.”

Domain 6: Did the review authors perform data extraction in duplicate?

Based on the provided text, the authors of the systematic review performed data extraction in duplicate. Specifically, the text states that “Titles and abstracts were independently selected by pairs of independent reviewers using the liberal accelerated approach [27]. Disagreements were resolved by consensus or by a third reviewer.” While this quote primarily speaks to the process of study selection, the following description: “Two reviewers independently evaluated the following items for each study: selection bias due to random sequence generation, selection bias due to allocation concealment, performance bias, detection bias, attrition bias, reporting bias and other bias (e.g. industry sponsorship)” suggests that a similar approach was likely applied to data extraction, given that these biases are typically assessed as part of the data extraction process.

However, the text does not explicitly mention the details about how the data extraction process was performed, nor does it mention the use of a Kappa score or any other specific measure of inter-rater reliability. Therefore, while it appears likely that the data extraction was performed in duplicate, there is not enough information in the provided text to conclusively state this.

Assessment: Partial Yes(More information needed to confirm whether the data extraction was performed in duplicate).

Domain 7: Did the review authors provide a list of excluded studies and justify the exclusions?

The authors of the systematic review provided a list of excluded studies and justified the exclusions. In the results section, it is mentioned that an initial search identified 21,161 titles and abstracts. Of these, 15,819 were excluded during the initial screening process based on titles, abstracts, and registers. The text also mentions that “910 potentially eligible studies were read in full, of which 634 were excluded” and that the reasons for these exclusions are presented in Table S2, which can be accessed through the supplement.

Based on the provided text, my evaluation is a “Yes” because the authors have mentioned a list of excluded studies and referred to a table containing the reasons for exclusion.

Domain 8: Did the review authors describe the included studies in adequate detail?

Yes, the authors of the systematic review have described the included studies in adequate detail.

In terms of study design, the authors describe the types of studies included (double-blind placebo or active-controlled randomized controlled trials, parallel, crossover or factorial) and the follow-up period between 3 and 52 weeks. The report also provides information about the types of participants included (adults with primary hypertension) and the interventions and comparators involved (various thiazide diuretics alone or combined with a potassium-sparing diuretic). Outcomes were also clearly defined, with the primary outcome being the mean difference in office SBP, and safety outcomes including mean difference in biochemical abnormalities and the incidence of non-melanoma skin cancer.

The locations where the research was conducted were not specified directly in the provided text, however, the authors detail that multiple databases were searched without language restrictions, implying global inclusivity. This also indirectly provides an idea about the settings where the research was conducted.

In terms of the results, the report provides a high-level summary of the number of studies included and the characteristics of the participants (mean age, proportion of women), as well as the interventions used most often and the risk of bias within studies. Further information, including raw data and other specifics, is provided in supplementary materials, tables, and through a public repository link.

The data extraction process, risk of bias assessment, and statistical analysis methods are also thoroughly described, allowing the reader to assess the validity and reliability of the results. The full glossary of extracted variables with their definitions is available at the specified link.

Overall, there is sufficient detail provided to allow the reader to assess the generalizability of the results.

Domain 9: Did the review authors use a satisfactory technique for assessing the risk of bias (RoB) in individual studies that were included in the review?

Based on the provided text, the authors of the systematic review have indeed used a satisfactory technique for assessing the risk of bias in the individual studies that were included in the review.

The text mentions that the risk of bias was assessed using the Risk of Bias for Interventions tool v. 1.0 from the Cochrane Collaboration. This is a standard tool and is appropriate for the study design (double-blind placebo or active-controlled RCTs) as they used.

It is also mentioned that two reviewers independently evaluated the risk of bias for each study, which fulfills the criterion of having at least two independent assessors. Moreover, the elements evaluated by these reviewers cover a wide range of bias types including selection bias, performance bias, detection bias, attrition bias, reporting bias and other biases (e.g., industry sponsorship). This indicates a comprehensive approach to bias assessment.

In case of disagreement, it is mentioned that the final decision was made through a consensus process or by a third reviewer.

The results of the risk of bias assessment are available publicly at a specified URL, which indicates that the results are reported for each study.

From the text, it is not explicitly mentioned how the results of the risk of bias assessment were used in any subsequent synthesis. However, given the use of a comprehensive and standard tool, the rigorous independent dual assessment, and the transparency in making their assessments publicly available, it is likely that the risk of bias was taken into account appropriately in the analysis.

Therefore, my assessment is “Yes”, the authors used a satisfactory technique for assessing the risk of bias in the individual studies that were included in the review.

Domain 10: Did the review authors report on the sources of funding for the studies included in the review?

Based on the provided text, the authors of the systematic review did not report on the sources of funding for the studies included in their review. They thoroughly described their methods, data extraction and analysis, as well as risk of bias, but no explicit mention was made of funding sources for the studies they reviewed. Moreover, they did not acknowledge or discuss whether the source of funding for a study might be associated with the direction and strength of the study results. Therefore, the assessment for this domain is ‘No’. There is no information regarding the source of funding for the individual studies or the potential influence of funding on study outcomes in the provided text.

Domain 11: If meta-analysis was performed did the review authors use appropriate methods for statistical combination of results?

Based on the provided information, it appears that the authors of the systematic review have employed appropriate methods for the statistical combination of results. Their approach appears to follow current best practices in several respects:

1. **Statistical Methods**: The authors have used both frequentist and Bayesian statistical techniques. In the frequentist approach, they used random-effects meta-analysis models for direct pairwise evidence summarization. In the Bayesian framework, they conducted a multiple treatment comparison model combining all direct and indirect evidence to compare all treatments. This is in line with current practices for network meta-analyses, which require the combination of both direct and indirect evidence.

2. **Heterogeneity and Asymmetry Assessment**: The authors considered heterogeneity in their analyses by estimating I² statistics. They evaluated the presence of publication bias through the inspection of funnel plots.

3. **Model Selection**: They employed both fixed and random effects models and used the deviance information criterion (DIC) to decide between them. This is a common method to compare models and make choices based on model fit.

4. **Assumption Testing**: The authors checked the validity of their models by employing the split-node method with Bonferroni correction for the inconsistency assumption and by checking the characteristics of individual studies for the transitivity assumption.

5. **Handling of Missing Data**: Although the authors don’t directly discuss how they handled missing data, they have mentioned contacting authors for potential unpublished outcomes, which indicates their effort to minimize missing information.

6. **Combining Different Study Designs**: The authors included different types of randomized controlled trials (RCTs) – parallel, crossover, and factorial – in their analyses. They used a washout period for crossover studies, which is important to prevent carryover effects from influencing results.

So, in summary, the methods used for statistical combination of results in the meta-analysis are appropriate and robust based on the provided information. Therefore, my assessment is “Yes”.

Domain 12: If meta-analysis was performed, did the review authors assess the potential impact of RoB in individual studies on the results of the meta-analysis or other evidence synthesis?

The systematic review authors did assess the risk of bias in individual studies included in the meta-analysis. As per the provided text, “The risk of bias was assessed using the Risk of Bias for Interventions tool v. 1.0 from the Cochrane Collaboration (i.e. low, unclear or high risk of bias).” The authors also independently evaluated the following items for each study: selection bias due to random sequence generation, selection bias due to allocation concealment, performance bias, detection bias, attrition bias, reporting bias, and other bias (e.g., industry sponsorship).

However, there is no specific mention in the provided text about how the authors assessed the potential impact of risk of bias on the results of the meta-analysis or other evidence synthesis. The authors do not appear to have conducted sensitivity analyses excluding studies at high risk of bias, or subgroup analyses based on the risk of bias, or any other appropriate methods to assess the potential impact of bias on their findings.

So, while the authors do assess the risk of bias within individual studies, they do not seem to evaluate the potential impact of that bias on their overall results. Hence, the evaluation for this domain should be ‘No’, due to the missing or inadequately described assessment of the potential impact of risk of bias in individual studies on the results of the meta-analysis.

Domain 13: Did the review authors account for RoB in individual studies when interpreting/ discussing the results of the review?

The authors of the systematic review did account for the risk of bias in individual studies when interpreting and discussing the results of the review, as indicated in the section titled “Risk of bias within individual studies.” They used the Risk of Bias for Interventions tool v. 1.0 from the Cochrane Collaboration to assess risk, which allowed them to evaluate potential issues such as selection bias, performance bias, detection bias, attrition bias, reporting bias, and other sources of bias, including industry sponsorship. They evaluated these items for each individual study, with two reviewers independently conducting the assessment to ensure reliability.

The summary with individual assessments for each study and raw data with commentary from the reviewers are also openly available to the public, further demonstrating transparency and accountability in their review process.

However, while the authors did acknowledge that “unclear risk of bias was frequent,” there is no explicit discussion in the provided text on how this risk of bias might impact the findings of their review, nor a discussion on the limitations of the evidence included. This could be considered a limitation in their interpretation and discussion of the review’s results.

Therefore, the answer to the assessment of this domain is “No.” The authors did not provide a clear discussion on how the potential risk of bias in the individual studies could impact the overall findings of their review, or a discussion on the limitations of the evidence.

Domain 14: Did the review authors provide a satisfactory explanation for, and discussion of, any heterogeneity observed in the results of the review?

Based on the provided text, the authors have conducted the systematic review using a robust methodology, following PRISMA-NMA guidelines, and a thorough approach towards assessing potential sources of bias. They performed various statistical analyses and addressed heterogeneity using I2 statistics and used Bayesian multiple treatment comparison models for exploring the differences between treatments. However, the text doesn’t explicitly mention the exploration of the reasons for heterogeneity through subgroup analyses, sensitivity analyses, or meta-regression.

The authors assessed the validity of their models using the split-node method with Bonferroni correction for the inconsistency assumption and characteristics of the individual studies for the transitivity assumption. They also used the DerSimonian & Laird estimator for the variability among studies, indicating they considered heterogeneity in their meta-analysis.

However, it’s not clear from the text provided how the authors specifically dealt with the observed heterogeneity in their results. No explicit discussion or explanation for the presence of heterogeneity among studies’ results, nor any detailed subgroup or sensitivity analyses to explore potential sources of heterogeneity, were mentioned.

Based on this, the assessment for this domain would be “No”. Despite having considered heterogeneity statistically, the authors haven’t adequately described how they explored and explained the potential sources of heterogeneity in their systematic review. The authors should have provided a detailed discussion of the possible sources of heterogeneity and the results of any subgroup or sensitivity analyses conducted to investigate these.

Domain 15: If they performed quantitative synthesis did the review authors carry out an adequate investigation of publication bias (small study bias) and discuss its likely impact on the results of the review?

The authors of the systematic review did perform an adequate investigation of publication bias, which is also commonly referred to as small-study bias. They used a common statistical tool, funnel plots, to examine this bias. This method is a scatter plot of the effect estimate from each study against the standard error, and an asymmetrical plot suggests possible publication bias. The text specifically mentions: “Asymmetry was evaluated using funnel plots.”

However, it is important to note that the authors do not provide a discussion regarding the likely impact of publication bias on the results of the review in the provided text. Without this discussion, we cannot conclude if the authors believe that their results were impacted by publication bias, and if so, how significantly this bias affected their results. Therefore, while they have used an appropriate tool to investigate the possibility of such bias, their review is incomplete without this discussion.

The assessment for this domain would be a ‘No’, as the authors have not discussed the likely impact of publication bias on the results of the review. They need to include this discussion to fully address this domain.

Domain 16: Did the review authors report any potential sources of conflict of interest, including any funding they received for conducting the review?

Based on the provided text, the authors of this systematic review reported no potential sources of

conflict of interest. The document includes a conflict of interest statement and funding statement

which are necessary elements to evaluate this domain. Therefore, the assessment of this domain is

Yes.

Related Articles

A Sweet Dilemma: Does Diabetes Raise the Stakes for Breast Cancer?

A Sweet Dilemma: Does Diabetes Raise the Stakes for Breast Cancer?

Diabetes in Breast Cancer: Diabetes mellitus (DM) and breast cancer (BCa) are two significant health concerns affecting women worldwide. To shed light on the relationship between these two conditions, a study conducted a systematic review and meta-analysis following...

Empowering Health: The Patient Engagement Revolution

Empowering Health: The Patient Engagement Revolution

Welcome to a new era in healthcare where patients are active contributors to their own well-being. The traditional healthcare landscape is being revolutionised by patient engagement, which has become a driving force. You can learn the fundamentals of patient...

Stay Up to Date With The Latest News & Updates

Join Our Newsletter

Follow US