Effects of Creatine Supplementation and Resistance Training on Muscle Strength Gains in Adults <50 Years of Age: A Systematic Review and Meta-Analysis

Table of Contents

Overall Summary

Overview

Creatine supplementation significantly improves both upper and lower body strength (WMD = 4.43 kg and 11.35 kg, respectively). The meta-analysis utilizes WMD, a methodological strength. Key limitations include potential publication bias (evidenced by funnel plot asymmetry) and study heterogeneity. The inclusion of forest plots and the exploration of sex differences enhance the analysis.

Key Points

Effect of Creatine on Strength (text)
The meta-analysis reports a WMD of 4.43 kg for upper body strength and 11.35 kg for lower body strength, indicating a substantial increase in strength with creatine supplementation.
Section: Results
Use of WMD (text)
The use of weighted mean difference (WMD) is appropriate for this meta-analysis as it allows for a direct and interpretable measure of the effect size in kg.
Section: Methods
Forest Plot Interpretation (non-text)
The forest plot visually represents the results of individual studies and the overall pooled effect, allowing for assessment of heterogeneity.
Section: Results (Figure 1)
Publication Bias (text)
The meta-analysis acknowledges the potential for publication bias, which could overestimate the true effect of creatine.
Section: Discussion
Funnel Plot Asymmetry (non-text)
The funnel plot asymmetry suggests the presence of publication bias, indicating that smaller studies with null or negative results might be underrepresented.
Section: Results (Figure 2)
Sex Differences (text)
The analysis explores potential sex differences in the effects of creatine, adding to the comprehensiveness of the study.
Section: Discussion

Conclusion

The meta-analysis demonstrates a statistically significant positive effect of creatine supplementation on both upper and lower body strength, with a weighted mean difference (WMD) of 4.43 kg for upper body and 11.35 kg for lower body. The use of WMD is a methodological strength, allowing for direct interpretation of the effect size in kg. However, limitations such as potential publication bias (indicated by the funnel plot asymmetry) and the heterogeneity of included studies warrant cautious interpretation. The exploration of sex differences and the inclusion of forest plots enhance the analysis, contributing valuable insights into the effects of creatine supplementation.

Section Analysis

Abstract

Key Aspects

Strengths

Suggestions for Improvement

Introduction

Key Aspects

Strengths

Suggestions for Improvement

2. Methods

Key Aspects

Strengths

Suggestions for Improvement

Non-Text Elements

Table 1. PICOS criteria for the inclusion of RCTs in which the supplementation...
Full Caption

Table 1. PICOS criteria for the inclusion of RCTs in which the supplementation of creatine was combined with a well-structured resistance training program and pre-post-training strength gain was compared with placebo supplementation with resistance training.

First Reference in Text
Last, systematic reviews and meta-analyses were excluded, as well as studies that were not available in full text, acute interventions, non-peer-reviewed articles, opinion pieces, reviews, case reports, and editorials.
Description
  • Purpose of the Table: This table details the criteria used to select studies for inclusion in a systematic review examining the effects of creatine supplementation combined with resistance training on muscle strength. The table uses the PICOS framework, which stands for Population, Intervention, Comparison, Outcome, and Study design, to organize these criteria. For each of these five categories, the table lists what characteristics a study must have to be included, as well as what characteristics would disqualify a study.
  • Explanation of PICOS: The PICOS framework is a widely used tool in evidence-based medicine to define the scope of a systematic review. The Population section defines the specific characteristics of the participants included in the eligible studies. The Intervention section specifies the treatment or intervention being studied. The Comparison section defines the control group or alternative treatment against which the intervention is compared. The Outcome section defines the specific effects or results that the review seeks to analyze. Lastly, the Study design states which study design should be included in the review.
  • Explanation of RCTs: RCTs are experiments where participants are randomly assigned to different treatment groups (e.g., creatine vs. placebo), allowing researchers to determine the effects of the treatment.
  • Explanation of Excluded Study Types: This table explains that this particular review excluded any studies that weren't available in full text, weren't randomized controlled trials, involved only a single administration of treatment, or were not peer-reviewed. They also excluded systematic reviews, meta-analyses, opinion pieces, reviews, case reports and editorials.
Scientific Validity
  • Exclusion of Non-Peer-Reviewed Material: Excluding non-peer-reviewed publications, opinion pieces, reviews, case reports, and editorials is crucial for maintaining scientific rigor, as these sources do not typically undergo the same level of scrutiny as peer-reviewed research.
  • Focus on RCTs: Focusing on RCTs is a standard practice in systematic reviews examining treatment efficacy. RCTs, due to their randomized design, are considered the gold standard for establishing causal relationships between interventions and outcomes, minimizing bias and increasing the reliability of findings.
  • Exclusion of Acute Interventions: Excluding acute interventions is justified as the research question focuses on the effects of combined creatine supplementation and resistance *training*. Acute interventions, which examine immediate effects, would not provide insight into the adaptive responses to a training program, which is the focus of the review.
Communication
  • Clarity and Structure: The table clearly lays out the inclusion and exclusion criteria based on the PICOS framework, which enhances transparency and allows readers to quickly assess the scope of the review. The criteria themselves are presented logically, categorized by PICOS components (Population, Intervention, Comparison, Outcome, Study Design).
  • Use of PICOS Framework: Using the PICOS framework is a standard and accepted practice in systematic reviews, facilitating a comprehensive and structured approach to defining eligibility criteria. This strengthens the communication by adhering to established norms in the field, making the review process more understandable and credible for other researchers.

3. Results

Key Aspects

Strengths

Suggestions for Improvement

Non-Text Elements

Table 2. Evaluation of the methodological quality of eligible studies (n = 23)...
Full Caption

Table 2. Evaluation of the methodological quality of eligible studies (n = 23) utilizing the Physiotherapy Evidence Database (PEDro) scale.

First Reference in Text
Using the PEDro scale, it was determined that of the 23 studies [22 44] included, 2 articles [22,41] were rated as excellent, and the other 21 articles [23 40,42 44] were categorized as good quality (Table 2).
Description
  • Purpose and Structure: Table 2 presents the methodological quality scores for each of the 23 studies included in the systematic review. The quality assessment was performed using the PEDro scale, a tool specifically designed to evaluate the methodological quality of randomized controlled trials (RCTs). The PEDro scale consists of 11 items, each scored as either yes (1) or no (0), reflecting the presence or absence of a particular methodological feature that minimizes bias. Items 2-9 contribute to the total PEDro score, with a maximum possible score of 10. Item 1 is related to external validity and is not included in the total score.
  • Components of PEDro Scale: The PEDro scale assesses various aspects of trial methodology, such as random allocation, concealed allocation, baseline comparability between groups, blinding of participants, therapists, and assessors, reporting of dropouts, intention-to-treat analysis, and between-group statistical comparisons.
  • Presentation of Data: The table lists each included study and provides a yes/no (Y/N) indication and numerical score (1/0) for each of the 11 PEDro scale items. A total PEDro score is calculated for each study (excluding Item 1) along with a quality assessment as "excellent" (9-10), "good" (6-8), "fair" (4-5), or "poor" (<4).
Scientific Validity
  • Use of PEDro Scale: Using the PEDro scale is a valid and widely accepted method for evaluating the methodological quality of RCTs in physiotherapy and related fields. Its use strengthens the review's validity by providing a standardized and transparent assessment of the included studies' methodological rigor.
  • Reporting of Methodological Quality: While the categorization of studies as "excellent" or "good" is helpful, the authors could strengthen the analysis by providing the average PEDro score and standard deviation for the included studies. This would offer a more nuanced understanding of the overall methodological quality. Additionally, explicitly stating the inter-rater reliability for the PEDro assessments (kappa statistic) would further reinforce the robustness of the quality assessment.
Communication
  • Clarity and Accessibility of PEDro Scores: Presenting the PEDro scores in a tabular format, along with a summary of excellent and good quality studies, effectively communicates the overall methodological rigor of the included studies. The table is easy to read and provides a clear overview of the quality assessment for each study.
  • Explanation of Assessment Criteria: The table's footnote explaining the scoring criteria is helpful. However, it might benefit from a more concise explanation of the criteria's meaning, particularly for readers less familiar with the PEDro scale. For instance, providing a brief explanation of what "baseline comparability" or "blinded assessors" entails would enhance understanding.
Table 3. Information on the studies that were included in the systematic review...
Full Caption

Table 3. Information on the studies that were included in the systematic review (n = 23).

First Reference in Text
Detailed information on the controlled trials included in this systematic review is presented in Table 3.
Description
  • Purpose and Structure of the Table: Table 3 provides detailed information about the 23 studies included in the systematic review. Each row represents a single study, and the columns provide key characteristics of each study, including the authors, publication year, participant characteristics (training status, sex, number of participants, and age), and details about the creatine supplementation protocol (loading dose, maintenance does, and duration) and resistance training protocol (duration, sessions per week).
  • Creatine Supplementation Protocols: The "Creatine, Loading Protocol" column describes the initial high-dose creatine supplementation strategy used in some studies to rapidly increase muscle creatine stores. The "Creatine, Maintenance Protocol" column describes the lower dose of creatine used after the loading phase to maintain elevated creatine levels. Not all studies used a loading phase.
  • Muscle Group Assessed: The table also provides details about the type of muscle group assessed (upper body, lower body, or both) which is crucial as the effects of interventions can vary depending on the muscle groups examined.
Scientific Validity
  • Transparency and Reproducibility: Providing comprehensive details about each included study is essential for transparency and reproducibility in a systematic review. Table 3 effectively presents these details, allowing readers to critically assess the included studies and understand the basis for the review's conclusions. This strengthens the review's scientific validity by making the data and methodology transparent.
  • Inclusion of Participant Characteristics: The inclusion of participant characteristics like training status and age is crucial, as these factors can influence the response to creatine supplementation and resistance training. Providing these details allows for a more nuanced interpretation of the results and enables subgroup analyses based on these factors.
  • Completeness of Data: The "NR" notation for missing age data in the Hoffman et al. (2006) study slightly detracts from the completeness of the data presentation. While minor, it raises a question about the thoroughness of the data extraction process. The authors should make every effort to retrieve missing data or provide an explanation for its absence.
Communication
  • Clear Presentation of Study Characteristics: The table effectively summarizes key information from each included study, facilitating easy comparison and understanding of the different creatine supplementation and resistance training protocols used. The clear column headers and consistent formatting make it straightforward to extract relevant details about each study's methodology and participant characteristics.
  • Informative Value: While the table is comprehensive, it could be improved by adding a column indicating the main findings of each study (e.g., the magnitude of strength gains in the creatine group compared to placebo). This would enhance the table's informative value and provide readers with a quick overview of the individual study results.
Figure 1. Literature search flowchart, following the PRISMA 2020 guidelines.
First Reference in Text
The remaining 76 were discarded (Figure 1).
Description
  • Purpose and Adherence to PRISMA: Figure 1 is a flowchart depicting the step-by-step process of identifying and selecting relevant studies for inclusion in the systematic review. The flowchart follows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines, which are a set of evidence-based recommendations for reporting systematic reviews.
  • Stages of the Flowchart: The flowchart begins with the 'Identification' stage, showing the number of records identified from each database searched (PubMed, Embase, Scopus, SportDiscus). It then proceeds to the 'Screening' stage, where duplicates are removed, and records are screened based on titles and abstracts. The next stage is 'Eligibility,' where full-text articles are assessed against pre-defined inclusion and exclusion criteria. Finally, the flowchart shows the number of studies 'Included' in the systematic review after the eligibility assessment.
  • Components and Data Presentation: The flowchart uses boxes to represent each stage and arrows to indicate the flow of the selection process. The numbers in each box indicate the number of records remaining at each stage. For example, starting with a total of 1586 records identified, this number progressively decreases at each stage as studies are screened and excluded based on different criteria.
Scientific Validity
  • Adherence to PRISMA Guidelines: Adhering to the PRISMA 2020 guidelines is a standard practice in systematic reviews, strengthening the review's validity and ensuring transparency and completeness in reporting the search and selection strategy. The flowchart's clear presentation of the study selection process enhances the review's reproducibility.
  • Transparency and Completeness: Explicitly reporting the number of records identified from each database and the number of records excluded at each stage is crucial for transparency. This allows readers to understand the scope of the literature search and the reasons for excluding studies, facilitating assessment of potential biases in the selection process.
  • Detailed Reporting of Exclusion Reasons: While the flowchart provides a general overview of the reasons for excluding studies during title/abstract screening, specifying the reasons for excluding 76 studies after full-text review would further strengthen the review's scientific rigor. This detailed breakdown would provide more insights into the eligibility criteria's application and ensure transparency.
Communication
  • Clear Visualization of Study Selection Process: The flowchart clearly visualizes the study selection process, making it easy for readers to follow the steps involved in identifying and screening relevant articles. The use of standard PRISMA flowchart components (identification, screening, eligibility, included) promotes clarity and understanding.
  • Transparency of Exclusion Criteria: While the flowchart effectively presents the overall study selection process, providing specific reasons for excluding studies after full-text review would enhance transparency. This allows readers to understand the specific criteria applied during the final stage of selection and assess potential biases in the review process.
Figure 2. Effect of creatine supplementation and resistance training compared...
Full Caption

Figure 2. Effect of creatine supplementation and resistance training compared with a placebo and resistance training on upper-body strength. The red circle represents the pooled weighted mean difference following a random effect meta-analysis, expressed in kg.

First Reference in Text
Creatine supplementation combined with resistance training resulted in greater increases in upper-body strength compared with a placebo, with a very high probability (WMD = 4.43 kg, 95% CI [3.12,5.75], p < 0.001) (Figure 2).
Description
  • Purpose of the figure: meta-analysis results: Figure 2 presents the results of a meta-analysis examining the effects of creatine supplementation combined with resistance training on upper-body strength. The figure uses a forest plot to visually display the effect sizes and confidence intervals from each of the included studies.
  • Explanation of forest plot elements: A forest plot is a graphical representation commonly used in meta-analyses. Each horizontal line in the plot represents a single study. The square box on each line represents the study's effect size, with the size of the box indicating the study's weight in the meta-analysis (larger boxes mean more weight). The horizontal lines extending from each box represent the 95% confidence intervals for each study's effect size. If a confidence interval crosses the vertical line at zero, it indicates that the effect in that study is not statistically significant. The diamond at the bottom represents the pooled or overall effect size from all the studies combined, which takes into account the variability of results between individual studies. The width of the diamond represents the confidence interval for the pooled effect size.
  • Explanation of WMD and confidence interval: The weighted mean difference (WMD) is a statistical measure of the difference between the mean outcomes in two groups. In this case, it represents the difference in upper-body strength gains between the creatine group and the placebo group in kilograms. The caption says that the pooled weighted mean difference is 4.43kg and is statistically significant (p < 0.001). The confidence interval for this difference is 3.12 kg to 5.75 kg. In layman's terms, this means that people who took creatine gained, on average, 4.43kg more upper body strength compared to those who took the placebo.
  • Explanation of random-effects model: A random-effects model is a statistical approach used in meta-analysis that assumes the true effect size varies between studies. This approach is often used when there is heterogeneity between the results of individual studies. This model provides a more conservative estimate of the overall effect size.
Scientific Validity
  • Appropriate statistical methods: Using a forest plot and reporting the WMD, confidence intervals, and p-value are standard and appropriate methods for presenting meta-analysis results. The use of a random-effects model is justified, especially if there is substantial heterogeneity in the effect sizes of the individual studies.
  • Assessment of heterogeneity: The forest plot allows for a visual inspection of the between-study heterogeneity. The figure itself does not display the heterogeneity statistically, so it's difficult to fully assess the validity of choosing a random-effects model without knowing the extent of the heterogeneity (e.g., I-squared value). Reporting the heterogeneity statistic would strengthen the analysis.
  • Justification of random-effects model: The caption should clearly justify why a random-effects model is necessary, e.g., by stating that substantial heterogeneity was detected among included studies. If the authors are certain that there is no heterogeneity, a fixed effects model could be used, though random-effects are almost always preferred as they are more conservative.
Communication
  • Clear presentation of meta-analysis results: The forest plot clearly visualizes the results of the meta-analysis, displaying the effect size (WMD) for each included study, along with the corresponding confidence intervals. This allows readers to assess the variability between studies and the precision of the overall effect estimate. The red circle representing the pooled WMD is visually prominent, effectively highlighting the overall effect of creatine supplementation.
  • Use of WMD in kg: Expressing the effect size in kilograms (WMD) makes the results more interpretable and meaningful for readers, especially those interested in practical applications. This contrasts with other effect size metrics like standardized mean differences (SMD), which are less intuitive for understanding the magnitude of the effect in real-world terms.
  • Clarity of caption: The figure's caption clearly explains the key elements of the forest plot, including the meaning of the red circle and the use of a random-effects model. However, it could be improved by briefly mentioning why a random-effects model was chosen. Was there substantial heterogeneity between studies? This would enhance the figure's scientific rigor.
Figure 3. Subgroup analyses for creatine supplementation combined with...
Full Caption

Figure 3. Subgroup analyses for creatine supplementation combined with resistance training on upper-body muscle strength compared with placebo supplementation combined with resistance training. MD, mean difference (kg); P1, p-value for the within-subgroup comparison (i.e., pre-post-intervention changes within each subgroup); P2, p-value for the between-subgroup comparison (i.e., comparison of the pre-post-intervention changes between subgroups); NA, Not Available.

First Reference in Text
Six subgroup meta-analyses were performed (Figure 3) according to the characteristics of the creatine supplementation protocol (the existence of creatine loading and creatine dose), the characteristics of the resistance training protocol (duration and frequency of training), and the characteristics of the participants (training status and sex).
Description
  • Purpose of subgroup analyses: Figure 3 presents the results of subgroup analyses conducted as part of the meta-analysis. Subgroup analyses involve splitting the overall data into smaller groups (subgroups) based on specific characteristics (e.g., creatine loading, dose, training status, sex) to investigate whether the effects of creatine supplementation differ between these groups.
  • Explanation of forest plot elements and statistical measures: The figure uses a forest plot format, similar to Figure 2, but each section now represents a different subgroup analysis. The mean difference (MD) is shown as a square, with the size corresponding to the weight of each study. The horizontal lines are confidence intervals, and the diamond represents the pooled MD. P1 is the p-value for comparing pre- and post-training within each subgroup (does creatine have an effect in this group?). P2 is the p-value comparing the creatine effect *between* subgroups (do men and women respond differently?).
  • Interpretation of subgroup analyses: Subgroup analyses help explore potential moderators of the treatment effect. For example, the analysis stratified by sex would reveal if creatine's effect on muscle strength is similar in males and females. Similarly, analyzing subgroups based on creatine dose (low vs. high) can help determine if there's a dose-response relationship.
Scientific Validity
  • Rationale for subgroup analyses: Conducting subgroup analyses is a scientifically sound approach to exploring potential sources of heterogeneity in treatment effects. The selected subgrouping variables (creatine loading, dose, training characteristics, participant characteristics) are relevant to the research question and reflect factors that could plausibly influence the effects of creatine supplementation.
  • Reporting of heterogeneity within subgroups: The authors should explicitly report the heterogeneity statistic (I-squared) for *each* subgroup analysis to provide a quantitative measure of variability within each subgroup. This allows readers to assess whether the effect of creatine is consistent across studies within a given subgroup or if there are substantial differences.
  • Statistical methods for between-subgroup comparisons: While the figure shows p-values (P2) for comparisons between subgroups, the authors should clearly state the statistical method used for these comparisons. Commonly used methods include interaction tests or subgroup differences tests within the meta-analysis framework. Specifying the method used increases transparency and allows for critical evaluation of the subgroup findings.
Communication
  • Clear visual representation of subgroup analyses: Presenting the subgroup analyses as a forest plot allows for visual comparison of the effects of creatine within different subgroups. The use of distinct visual elements (e.g., square size, diamond) to represent different statistical measures (MD, p-values) facilitates quick interpretation of the results. The overall layout is clear, making it easy to identify subgroups and corresponding effect sizes.
  • Clarity of abbreviations and statistical tests: The figure caption clearly defines the abbreviations used (MD, P1, P2, NA), ensuring clarity and avoiding ambiguity. Explaining P1 and P2 in terms of within- and between-group comparisons further enhances understanding of the statistical analyses performed for each subgroup.
  • Heterogeneity within subgroups: While the visual presentation is generally clear, adding numerical values for the I-squared statistic for each subgroup analysis directly on the forest plot would enhance the interpretation of heterogeneity within subgroups. This would complement the p-values (P2) for between-subgroup heterogeneity.
Figure 4. Effect of creatine supplementation and resistance training compared...
Full Caption

Figure 4. Effect of creatine supplementation and resistance training compared with a placebo and resistance training on lower-body muscle strength. The red circle represents the pooled weighted mean difference following a random effect meta-analysis, expressed in kg.

First Reference in Text
Creatine supplementation combined with resistance training produced greater increases in maximal lower-body strength compared with a placebo, with a very high probability (WMD = 11.35 kg, 95% CI [8.44,14.25], p < 0.001; Figure 4).
Description
  • Purpose and format: meta-analysis results in forest plot: Figure 4, a forest plot, presents the meta-analysis results for the effect of creatine supplementation with resistance training on lower-body strength. Each horizontal line corresponds to a study, the square represents the study's effect size (WMD in kg), and the line extending from the square shows the 95% confidence interval.
  • Explanation of pooled WMD and confidence interval: The red circle represents the overall pooled weighted mean difference (WMD), summarizing the average effect of creatine across all studies. Its position on the x-axis indicates the magnitude of the effect in kg, and the diamond at the bottom represents the pooled effect size with its confidence interval.
  • Random-effects model and interpretation of WMD: A random-effects model assumes variation in true effect sizes between studies and provides a more conservative pooled estimate. The large positive WMD (11.35 kg) suggests that creatine supplementation substantially improves lower-body strength compared to placebo.
Scientific Validity
  • Appropriate statistical methods: Using a forest plot, WMD, confidence intervals, and p-value is appropriate for presenting meta-analysis results. The random-effects model is justified, given potential heterogeneity among studies examining lower-body strength outcomes.
  • Quantifying heterogeneity: Reporting the I-squared value is essential to quantify heterogeneity. While visually assessing the spread of effect sizes in the forest plot gives some indication, a numerical measure strengthens the interpretation and justifies the use of a random-effects model.
  • Justification of random-effects model: While not shown directly, the authors mention heterogeneity was not significant (p=0.897, I2=0%). Given that random effects models are generally preferred, this information should be provided in the caption to strengthen transparency.
Communication
  • Clear visual presentation and interpretability: The forest plot effectively communicates the results of the meta-analysis on lower-body strength, visually presenting the effect size (WMD) and confidence intervals for each study. The red circle clearly highlights the pooled WMD, emphasizing the overall effect of creatine. Expressing the effect size in kg enhances interpretability.
  • Justification of random-effects model: While the caption explains the key elements, briefly stating the reason for choosing a random-effects model (e.g., presence of heterogeneity) would improve transparency and scientific rigor.
  • Clarity of x-axis labels: Labeling the x-axis as "Favours Placebo" and "Favours Creatine" clarifies the direction of the effect, making it readily apparent that positive values indicate greater strength gains with creatine.
Figure 5. Subgroup analyses for creatine supplementation and resistance...
Full Caption

Figure 5. Subgroup analyses for creatine supplementation and resistance training compared with a placebo and resistance training on lower-body strength. MD, mean difference(kg); P1, p-value for the within-subgroup comparison (i.e., pre-post-intervention changes within each subgroup); P2, p-value for the between-subgroup comparison (i.e., comparison of the pre-post intervention changes between subgroups); NA, Not Available.

First Reference in Text
Similar to upper-body strength, six subgroup meta-analyses were analyzed for lower-body strength, including the same categories (Figure 5).
Description
  • Purpose of subgroup analyses: Figure 5 presents subgroup analyses exploring how the effect of creatine on lower-body strength varies across different subgroups based on factors like creatine loading, dose, training characteristics, and participant characteristics (similar to Figure 3).
  • Explanation of forest plot elements and statistical measures: The figure uses a forest plot where each section represents a subgroup. The square indicates the mean difference (MD) in kg, the horizontal lines depict confidence intervals, and the diamond shows the pooled MD. P1 represents the p-value for changes within each subgroup, and P2 represents the p-value for differences between subgroups.
  • Interpretation of subgroup analyses: Subgroup analyses help identify potential moderators. For instance, the "Gender" subgroup analysis investigates whether the effect of creatine differs between males and females. The "Creatine dose" subgroup analysis explores if there's a dose-response relationship.
Scientific Validity
  • Rationale for subgroup analyses: Conducting subgroup analyses is a valid approach to investigate potential effect modifiers. The chosen subgrouping variables are relevant and align with factors that could influence the response to creatine supplementation and resistance training.
  • Reporting heterogeneity within subgroups: Reporting I-squared values for each subgroup analysis is crucial for quantifying heterogeneity within subgroups and would enhance the interpretation of results. This would help understand the consistency of creatine's effect within each subgroup.
  • Clarity of statistical methods for between-subgroup comparisons: While between-subgroup p-values (P2) are shown, specifying the statistical method used for these comparisons (e.g., interaction tests, subgroup differences tests) is important for transparency and allows critical assessment of the statistical rigor.
Communication
  • Clear visual presentation of subgroup analyses: The forest plot effectively presents the subgroup analyses for lower-body strength, allowing visual comparison of effect sizes (MD) and confidence intervals across subgroups. The labeling of the x-axis ("Placebo Better", "Creatine Better") clearly indicates the direction of effect. Consistent use of visual elements (squares, diamonds) maintains clarity and coherence with Figure 3.
  • Clarity of abbreviations and statistical tests: The caption provides clear definitions of abbreviations (MD, P1, P2), enhancing understanding. Explaining P1 and P2 in terms of within- and between-group comparisons is helpful, ensuring readers grasp the statistical analyses performed.
  • Visual representation of between-subgroup differences: While the figure presents p-values (P2) for between-subgroup differences, visually representing the magnitude of these differences (e.g., using different colors or shading) would further enhance understanding and allow quicker identification of statistically significant differences.
Figure 6. Observed and imputed funnel plot for upper-body muscle strength. The...
Full Caption

Figure 6. Observed and imputed funnel plot for upper-body muscle strength. The funnel plot displays the distribution of studies included in this meta-analysis, with white circles representing the original observed studies and black circles indicating the imputed studies added to account for potential publication bias using the trim and fill method. At the bottom of the funnel plot, the white and black diamonds represent the combined effect sizes, with the white diamond indicating the overall effect from the observed studies, and the black diamond showing the adjusted effect after imputation.

First Reference in Text
The funnel plot for the gains obtained with creatine over the placebo on upper-body muscle strength was asymmetric, suggesting possible publication bias. This was confirmed by Egger's linear regression test (t = 2.102, p = 0.048).
Description
  • Purpose of funnel plot: Figure 6 presents a funnel plot, which is a scatterplot used in meta-analyses to visually assess publication bias. Publication bias arises when studies with statistically significant or positive results are more likely to be published than those with non-significant or negative results.
  • Elements of a funnel plot: In a funnel plot, each dot represents a study. The x-axis represents the study's effect size (strength of the effect of creatine in this case), and the y-axis represents a measure of the study's precision, usually the standard error. Studies with high precision will cluster near the top and the spread should widen as precision decreases. In the absence of publication bias, the points form a roughly symmetrical inverted funnel shape. Asymmetry, as seen in Figure 6, can suggest publication bias.
  • Trim and fill method: The trim and fill method is a technique used to address publication bias. It imputes "missing" studies (represented by black circles) that would make the funnel plot symmetrical if publication bias were the sole source of asymmetry. The white diamond at the bottom shows the original overall effect, and the black diamond shows the adjusted effect after accounting for the imputed studies.
Scientific Validity
  • Appropriate methods for assessing publication bias: Using a funnel plot and Egger's regression test are appropriate methods for assessing publication bias in meta-analyses. Egger's test provides statistical confirmation of the visual asymmetry observed in the funnel plot.
  • Limitations of trim and fill: The trim and fill method is a widely used technique for adjusting for publication bias, but it is important to note that it relies on assumptions that may not always hold true. The results of the trim and fill analysis should be interpreted cautiously.
  • Reporting adjusted effect size: Providing the adjusted effect size (WMD after imputation) and its confidence interval would be valuable. Quantifying the impact of the imputed studies helps readers understand the extent to which publication bias may have influenced the overall effect estimate.
Communication
  • Visual representation of publication bias: The funnel plot visually represents the potential publication bias, with the asymmetry suggesting a bias towards publishing studies with larger effects. The use of white and black circles to differentiate observed and imputed studies, respectively, is effective, and the diamonds clearly represent the combined effect sizes before and after imputation.
  • Clarity of explanation: The caption clearly explains the purpose of the funnel plot, the meaning of the different symbols (circles and diamonds), and the trim and fill method used to adjust for publication bias. This makes the figure accessible to readers who may not be familiar with these concepts.
  • Informative value: Adding the effect size (WMD or SMD) values corresponding to the white and black diamonds in the caption would make the figure more informative. It would allow readers to quickly quantify the impact of the imputed studies on the overall effect estimate.
Figure 7. Observed and imputed funnel plot for lower-body muscle strength. The...
Full Caption

Figure 7. Observed and imputed funnel plot for lower-body muscle strength. The funnel plot displays the distribution of studies included in this meta-analysis, with white circles representing the original observed studies. At the bottom of the funnel plot, the white and black diamonds represent the combined effect sizes, with the white diamond indicating the overall effect from the observed studies and the black diamond showing the adjusted effect after imputation.

First Reference in Text
The funnel plot for the gains obtained with creatine over the placebo on lower-body muscle strength was symmetrical (Figure 7), indicating no significant publication bias, which was confirmed by Egger's linear regression test (t = 0.122, p = 0.90).
Description
  • Purpose of funnel plot: Figure 7 displays a funnel plot to visually assess publication bias in the meta-analysis of lower-body muscle strength outcomes.
  • Elements of funnel plot and interpretation of symmetry: Each white circle in the plot represents an individual study, with the x-axis showing the study's effect size and the y-axis indicating the standard error (a measure of study precision). A symmetrical, inverted funnel shape indicates a lack of substantial publication bias.
  • Combined effect sizes and trim and fill method: The white diamond represents the combined effect size of the observed studies. The black diamond represents the adjusted effect size after using the trim and fill method, a technique to account for potential publication bias by imputing "missing" studies. The near-overlap of the white and black diamonds suggests minimal impact from imputation and reinforces the lack of publication bias.
Scientific Validity
  • Appropriate methods for assessing publication bias: Using a funnel plot and Egger's regression test are valid approaches for assessing publication bias. The symmetrical funnel plot and non-significant Egger's test (p = 0.90) appropriately support the conclusion of no significant publication bias.
  • Justification of trim and fill: Applying the trim and fill method, even in the absence of substantial asymmetry, is a conservative approach. It allows for estimating the potential impact of imputed studies on the overall effect size, further confirming the robustness of the findings against publication bias.
  • Reporting effect sizes with and without imputation: Including the effect size values (with confidence intervals) corresponding to the white and black diamonds in the figure or caption would strengthen the interpretation. This would quantitatively demonstrate the minimal impact of imputation on the overall effect size.
Communication
  • Visual representation of lack of publication bias: The funnel plot effectively communicates the lack of significant publication bias for lower-body strength. The symmetrical distribution of studies around the center line visually supports this conclusion. The white circles clearly represent the observed studies, and the diamonds indicate the combined effect sizes before and after imputation.
  • Clarity of explanation: The caption clearly explains the elements of the funnel plot, such as the white circles representing observed studies and the diamonds representing combined effect sizes. The reference to the trim and fill method helps clarify the imputation process for addressing potential publication bias.
  • Emphasis on minimal imputation impact: While the figure demonstrates symmetry, visually highlighting the overlap or near-overlap of the white and black diamonds would emphasize the minimal impact of imputation, further reinforcing the absence of substantial publication bias.

4. Discussion

Key Aspects

Strengths

Suggestions for Improvement

5. Conclusions

Key Aspects

Strengths

Suggestions for Improvement

↑ Back to Top