Mainstream Meta-Analysis of Clinical Trials produces strongly Inconsistent Estimators

REX B. KLINE Beyond Significance Testing: Reforming Data Analysis Methods in Behavioral Research Washington, DC: American Psychological Association, 2004, 336 pages (ISBN 1-59147-118-4, US$49.95 Hardcover) In 1999, a blue-ribbon task force assembled by the American Psychological Association published their findings with regards to the long-standing controversy pertaining to null hypothesis testing (NHST). The task force dictated effect sizes and confidence intervals be reported, and p values and dichotomous accept-reject decisions be given less weight. Editorial policies in a number of journals came to reflect the views of the task force as did a subsequent revision to the American Psychological Association Publication Manual. Rex B. Kline wrote Beyond Significance Testing. Reforming Data Analysis Methods in Behavioral Research as a follow-up to both the task force recommendations and the revision to the publication manual. Kline's 1998 book Principles and Practice of Structural Equation Modeling (Guilford Press) was well received and a second edition is being published this fall. In Beyond Significance Testing, Kline reviews the controversy regarding testing, offers methods for effect size and confidence interval estimation, and suggests some methodologies. There is an accompanying website that includes resources for instructors and students. Part I of the book is a review of fundamental concepts and the debate regarding testing. Part II provides statistics for effect size and confidence interval estimation for parametric and nonparametric two-group, oneway, and factorial designs. Part III examines metaanalysis, resampling, and Bayesian estimation procedures. In the first chapter, Kline provides a scholarly summary of the null hypothesis testing debate concluding with the APA task force findings and what Kline regards as ambiguous recommendations in the publication manual. Kline predicts the future will see a smaller role for traditional statistical testing (p values) in psychology. This change will take time and may not occur until the next generation of researchers are trained, but Kline anticipates the social sciences will then become more like the natural sciences in that we will report the directions and magnitudes of our effects, determine whether they replicate, and evaluate them for their theoretical, clinical, or practical significance (p. 15). Chapter 2 is a review of fundamental concepts of research design, including sampling and estimation, the logic of statistical testing, and t, F, and chi-square tests. The problems with statistical tests are revisited in Chapter 3. What follows is a long list of errors in interpretation of p values and conclusions made after null hypothesis testing. The emphasis on null hypothesis testing in psychology is also argued to inhibit advancement of the discipline. To be fair, Kline recognizes there is yet no magical alternative to statistical tests and that such tests are appropriate in some circumstances when applied correctly. Nonetheless, Kline envisions a future where effect sizes and confidence intervals are reported, substantive rather than statistical predominates, and NHST-Centric thinking has diminished. Part II covers effect size and confidence interval calculations. Chapter 4 is a presentation of parametric effect size indexes. Independent and dependent sample statistics are covered separately. The textbook's website has a supplementary chapter on twogroup multivariate designs. Group difference indexes such as d are distinguished from measures of association such as r. Case level analyses of group differences are also reviewed. Sections not relevant to a reader's needs can be skipped without loss of continuity. Interpretive guidelines for effect size magnitude and how one might be fooled by effect size estimation are sections that should not be passed over. …

BackgroundEstimating a realistic effect size is an important issue in the planning of clinical studies of complementary and alternative medicine therapies. When a minimally important difference is not available, researchers may estimate effect size using the published literature. This evidence-based effect size estimation may be used to produce a range of empirically-informed effect size and consequent sample size estimates. We provide an illustration of deriving plausible effect size ranges for a study of acupuncture in the relief of post-chemotherapy fatigue in breast cancer patients.MethodsA PubMed search identified three uncontrolled studies reporting the effect of acupuncture in relieving fatigue. A separate search identified five randomized controlled trials (RCTs) with a wait-list control of breast cancer patients receiving standard care that reported data on fatigue. We use these published data to produce best, average, and worst-case effect size estimates and related sample size estimates for a trial of acupuncture in the relief of cancer-related fatigue relative to a wait-list control receiving standard care.ResultsUse of evidence-based effect size estimation to calculate sample size requirements for a study of acupuncture in relieving fatigue in breast cancer survivors relative to a wait-list control receiving standard care suggests that an adequately-powered phase III randomized controlled trial comprised of two arms would require at least 101 subjects (52 per arm) if a strong effect is assumed for acupuncture and 235 (118 per arm) if a moderate effect is assumed.ConclusionEvidence-based effect size estimation helps justify assumptions in light of empirical evidence and can lead to more realistic sample size calculations, an outcome that would be of great benefit for the field of complementary and alternative medicine.

Preface. Introduction. Data Sets. Tests of Statistical Significance of Combined Results. Vote-Counting Methods. Estimation of a Single Effect Size: Parametric and Nonparametric Methods. Parametric Estimation of Effect Size from a Series of Experiments. Fitting Parametric Fixed Effect Models to Effect Sizes: Categorical Methods. Fitting Parametric Fixed Effect Models to Effect Sizes: General Linear Models. Random Effects Models for Effect Sizes. Multivariate Models for Effect Sizes. Combining Estimates of Correlation Coefficients. Diagnostic Procedures for Research Synthesis Models. Clustering Estimates of Effect Magnitude. Estimation of Effect Size When Not All Study Outcomes Are Observed. Meta-Analysis in the Physical and Biological Sciences. Appendix. References. Index.

To conduct a meta-analytic review of interventions to reduce HIV individuals' sexual risk. Studies were included if they examined a deliberate sexual risk-reduction strategy in a sample that included HIV participants, used a randomized controlled trial design, measured condom use or number of sexual partners after the intervention, and provided sufficient information to calculate effect size (ES) estimates. Reports were gathered from computerized databases, by contacting individual researchers, by searching relevant journals and conference proceedings, and by reviewing reference sections of obtained papers. Data from 15 studies (N = 3234 participants) available as of November 30, 2004 were included. ES estimates were standardized mean differences. Across the studies, intervention participants exhibited lowered sexual risk relative to control participants on condom use (mean ES = 0.16, 95% confidence interval [CI]: 0.08 to 0.25) but not for number of sexual partners (mean ES = -0.01, 95% CI: -0.16 to 0.14). Interventions were more successful at increasing condom use if the sample included fewer men who have sex with men (MSM) or younger participants and when interventions included motivational and skills components. Behavioral interventions reduced sexual risk especially if they included motivational and skills components. Such interventions have been less effective for older samples, suggesting the need for further refinement to enhance their efficacy. Motivation-and skill-based interventions have not yet been tested with HIV MSM who, in general, seem to have benefited less from extant risk-reduction interventions.

Several studies have investigated the effect of Co-Q10 on diabetes biomarkers, but findings are inconsistent. This systematic review and meta-analysis of clinical trials was conducted to summarize the effect of Co-Q10 supplementation on diabetes biomarkers. We searched Pubmed, EMBASE, Science direct, ISI web of science, and Google Scholar for randomized controlled trials from 1989 until March 2016. We included randomized controlled trials reporting fasting blood glucose (FBG), fasting insulin and HbA1c. In total, we found 16 articles that examined the effect of Co-Q10 on fasting blood glucose, fasting insulin and HbA1c. Mean ± standard deviations (SD) of defined outcomes were used for calculating standardized mean differences (SMD) and its 95% confidence interval (95%CI) between intervention and control treatments based on Hedges' estimator. Our preliminary meta-analysis on 14 eligible studies regarding the effect of Co-Q10 supplementation on FBG indicated a slightly significant decrement (SMD:-0.28 mg/d; 95% CI: -0.12, 0.04), with a substantial between-study heterogeneity (Cochrane Q test, I2 = 93.9%, P < 0.0001). After removal of three studies, heterogeneity was eliminated and a slightly significant decrease was found in FBG (SMD: -0.20 mg/dL, 95% CI: -0.38, -0.02). The effect of Co-Q10 on HbA1c and fasting insulin was not significant. SMDs for the effect of Co-Q10 on HbA1c and fasting insulin were -0.05% (95% CI: -0.22, 0.12) and 0.12 pmol/L (95% CI: -0.21, 0.44), respectively. Co-Q10 supplementation slightly but significantly reduced fasting blood glucose, but not fasting insulin and HbA1c. More long-term studies are necessary to examine the association between Co-Q10 supplementation and diabetes biomarkers. This study was funded by the School of Nutrition and Food Science, Isfahan University of Medical Sciences.

We examined the degree of autocorrelation among single‐case design data with six measures used to estimate treatment effect size. The most commonly used measures of effect size for single‐case data over the last 5 years published in peer‐reviewed journals were selected for comparison. The overall mean degree of autocorrelation was 0.46 (SD = 0.33) across the 304 data paths, which represents a moderate degree of autocorrelation. Overall, it appears that non‐parametric measures of effect size (i.e., percent of non‐overlapping data [PND], non‐overlap of all pairs [NAP], and improvement rate difference [IRD] values) were substantially and significantly more influenced by the degree of autocorrelation. Tau‐U effect size estimate was the non‐parametric exception as it was not significantly influenced by the degree of autocorrelation. Parametric measures of effect sizes such as standardized mean difference (SMD) and log response ratio (LRR) values did not appear to be significantly influenced by the degree of autocorrelation. For SMD, LRR, and Tau‐U values, the correlation between the effect size value and the degree of autocorrelation was minimal. For NAP, IRD, and PND values, the correlation between the effect size value and the degree of autocorrelation was moderate, indicating that these estimates of effect size should be avoided as the degree of autocorrelation between data points increases.

Many empirical studies estimating effective population size apply the temporal method that provides an estimate of the variance effective size through the amount of temporal allele frequency change under the assumption that the study population is completely isolated. This assumption is frequently violated, and the magnitude of the resulting bias is generally unknown. We studied how gene flow affects estimates of effective size obtained by the temporal method when sampling from a population system and provide analytical expressions for the expected estimate under an island model of migration. We show that the temporal method tends to systematically underestimate both local and global effective size when populations are connected by gene flow, and the bias is sometimes dramatic. The problem is particularly likely to occur when sampling from a subdivided population where high levels of gene flow obscure identification of subpopulation boundaries. In such situations, sampling in a manner that prevents biased estimates can be difficult. This phenomenon might partially explain the frequently reported unexpectedly low effective population sizes of marine populations that have raised concern regarding the genetic vulnerability of even exceptionally large populations.

The Publication Manual of the American Psychological Association (American Psychological Association, 2001, American Psychological Association, 2010) calls for the reporting of effect sizes and their confidence intervals. Estimates of effect size are useful for determining the practical or theoretical importance of an effect, the relative contributions of factors, and the power of an analysis. We surveyed articles published in 2009 and 2010 in the Journal of Experimental Psychology: General, noting the statistical analyses reported and the associated reporting of effect size estimates. Effect sizes were reported for fewer than half of the analyses; no article reported a confidence interval for an effect size. The most often reported analysis was analysis of variance, and almost half of these reports were not accompanied by effect sizes. Partial η2 was the most commonly reported effect size estimate for analysis of variance. For t tests, 2/3 of the articles did not report an associated effect size estimate; Cohen's d was the most often reported. We provide a straightforward guide to understanding, selecting, calculating, and interpreting effect sizes for many types of data and to methods for calculating effect size confidence intervals and power analysis.

Many primary research studies in ecology are underpowered, providing very imprecise estimates of effect size. Meta-analyses partially mitigate this imprecision by combining data from different studies. But meta-analytic estimates of mean effect size may still remain imprecise, particularly if the meta-analysis includes a small number of studies. Imprecise, large-magnitude estimates of mean effect size from small meta-analyses likely would shrink if additional studies were conducted (regression towards the mean). Here, I propose a way to estimate and correct this regression to the mean, using meta-meta-analysis (meta-analysis of meta-analyses). Hierarchical random effects meta-meta-analysis shrinks estimated mean effect sizes from different meta-analyses towards the grand mean, bringing those estimated means closer on average to their unknown true values. The intuition is that, if a meta-analysis reports a mean effect size much larger in magnitude than that reported by other meta-analyses, that large mean effect size likely is an overestimate. This intuition holds even if different meta-analyses of different topics have different true mean effect sizes. Drawing on a compilation of data from hundreds of ecological meta-analyses, I find that the typical (median) ecological meta-analysis overestimates the absolute magnitude of the true mean effect size by ~10%. Some small ecological meta-analyses overestimate the magnitude of the true mean effect size by >50%. Meta-meta-analysis is a promising tool for improving the accuracy of meta-analytic estimates of mean effect size, particularly estimates based on just a few studies.

The Crohn's Disease Endoscopic Index of Severity (CDEIS) and the Simple Endoscopic Score for Crohn's Disease (SES-CD) are commonly used to assess Crohn's disease (CD) activity; however neither instrument is fully validated. We evaluated the responsiveness to change of the SES-CD and CDEIS using data from a trial of adalimumab, a drug therapy of known efficacy. Paired video recordings (N=112) of colonoscopies (baseline and week 8-12) obtained from patients with CD who participated in a trial of adalimumab therapy were reviewed in random order, in duplicate, by four central readers (56 pairs of videos by 2 groups of readers). Responsiveness of the SES-CD and the CDEIS was evaluated by comparing correlations between the observed and pre-specified predictions of change scores for these endoscopic indices with a global endoscopic evaluation of severity (GELS), a patient reported outcome (PRO2), and the Crohn's disease activity index (CDAI), and by calculation of the standardized effect size, and Guyatt's Responsiveness statistic (GRS) using 2 definitions of change; (1) treatment assignment and (2) an absolute change in total PRO2 of 50. The potential application of effect size estimates was demonstrated by calculating hypothetical sample sizes for comparing two independent groups. The impact of removing stenosis as an index item and adjusting for the number of segments observed was also assessed. Changes in both endoscopic instruments and the GELS were highly correlated. The SES-CD displayed numerically higher effect sizes for both definitions of change. The standardized effect size and GRS estimates (95% confidence interval) for the SES-CD based on treatment assignment were 0.84 (0.53, 1.15) and 0.79 (0.48, 1.09). Corresponding values for the CDEIS were 0.72 (0.42, 1.02) and 0.75 (0.45, 1.06). The standardized effect size and GRS estimates for the SES-CD based on an absolute change in total PRO2 of 50 points or greater were 0.76 (0.49, 1.02) and 0.93 (0.64, 1.21). Corresponding values for CDEIS were 0.70 (0.44, 0.97), 0.83 (0.55, 1.10). Removal of stenosis as an index item and adjusting for observed segments did not improve responsiveness estimates. Although both the SES-CD and CDEIS are valid measures of endoscopic disease activity that are moderately responsive to changes in endoscopic disease activity, the SES-CD displayed numerically greater responsiveness in this data set.

Replication is often viewed as the demarcation between science and nonscience. However, contrary to the commonly held view, we show that in the current (selective) publication system replications may increase bias in effect size estimates. Specifically, we examine the effect of replication on bias in estimated population effect size as a function of publication bias and the studies’ sample size or power. We analytically show that incorporating the results of published replication studies will in general not lead to less bias in the estimated population effect size. We therefore conclude that mere replication will not solve the problem of overestimation of effect sizes. We will discuss the implications of our findings for interpreting results of published and unpublished studies, and for conducting and interpreting results of meta-analyses. We also discuss solutions for the problem of overestimation of effect sizes, such as discarding and not publishing small studies with low power, and implementing practices that completely eliminate publication bias (e.g., study registration).

Background: Clinical trials have shown remote symptom monitoring (RSM) using electronic patient-reported outcomes (ePROs) reduces healthcare utilization among patients with different cancer types. However, the impact of RSM is not well understood in real-world, standard-of-care patients diagnosed with early stage or advanced breast cancer. Methods: This analysis of a hybrid, type 2 implementation-effectiveness trial evaluated the impact of RSM on healthcare utilization (hospitalizations &amp; emergency room [ER] visits) among patients with breast cancer receiving chemotherapy, immunotherapy, or targeted therapy at the University of Alabama at Birmingham. We compared patients receiving RSM with historical controls who met RSM eligibility criteria yet received care prior to RSM implementation. Healthcare utilization was assessed 3 and 6 months post-RSM enrollment or date of first RSM qualifying treatment for controls. Unadjusted analyses included crosstabulations and estimation of effect sizes using Cramer’s V (V: 0.10∼small, 0.30∼ medium, and 0.50∼large effect size). We estimated risk ratios (RR) and 95% confidence intervals (CI) of healthcare utilization between patients receiving RSM and controls using modified Poisson regression with robust standard errors. Penalized logistic regression was used in subgroup analysis by stage (early [I-III] vs. metastatic [de novo or recurrent]) to estimate odds ratios (OR) and CIs of healthcare utilization for RSM-enrolled patients compared to controls. Models controlled for age at RSM enrollment, race, sex, cancer stage (Poisson models only), insurance status, prior treatment, number of comorbidities, rurality of residence, area deprivation index, and an indicator for follow-up time during COVID-19 pandemic (April 2020-July 2021). Results: From May 2021 to February 2024, 328 patients with breast cancer were enrolled in RSM; 30% were Black, 13% resided in a rural area, and 24% lived in a highly disadvantaged neighborhood. Patients receiving RSM were similar to controls treated from January 2017 to January 2021 (n=494) in terms of race and residence (28% Black, 18% rural residence, 27% lived in a highly disadvantaged neighborhood). However, patients receiving RSM more often had metastatic disease at diagnosis compared to controls (25% vs. 11%). Compared to controls, patients receiving RSM were less often hospitalized at 3 months (10% vs. 13%; V=0.02) and 6 months (19% vs. 26%; V=0.07) post-enrollment. Patients receiving RSM visited the ER less often compared to control patients (3 months: 15% vs 16%; V=0.02; 6 months: 21% vs. 25%; V=0.04). In adjusted models, 6-month risk of hospitalizations and ER visits were 30% and 31% lower, respectively, for patients receiving RSM compared to controls (RR 0.70, CI 0.52-0.93; RR 0.69, CI 0.53-0.90). Though not statistically significant, the 3-month risk of hospitalizations and ER visits were also 21% and 24% lower, respectively, for patients receiving RSM compared to controls (RR 0.79, 95% CI 0.52-1.21; RR 0.76, 95% CI 0.55-1.06). In subgroup analyses, 6-month odds of hospitalizations and ER visits were 37% and 39% lower for early-stage patients receiving RSM compared to controls (OR 0.63, 95% CI 0.41-0.96; OR, 0.61, 95% CI 0.39-0.94, respectively). Though not statistically significant, 6-month odds of hospitalizations or ER visits also trended lower for metastatic breast cancer patients receiving RSM compared to controls (OR 0.49, 95% CI 0.18-1.33; OR 0.49, 95% CI 0.18-1.30, respectively). Conclusions: The use of ePRO-captured RSM reduced the risk of hospitalizations and ER visits for breast cancer patients in real-world settings, with the greatest impact seen for patients with longer engagement in RSM. Further work to expand this intervention nationally is needed. Citation Format: Gabrielle Rocque, Jeffrey A. Franks, Luqin Deng, Nicole E. Caston, Andres Azuero, Bradford E. Jackson, Nicole L. Henderson, D’Ambra Dent, Stacey Ingram, Noon Eltoum, Indya Starks, Joud El Dick, Nusrat Jahan, Erica Stringer-Reasor, Katia Khoury, Angela M. Stover, Ethan Basch, Jennifer Young Pierce, Courtney P. Williams. Real-world impact of remote symptom monitoring program on hospitalizations and emergency room visits in women with breast cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P3-06-05.

There is controversy regarding the most appropriate goals for blood pressure control. We assess the benefits and risks of chossing different therapeutic thresholds. We perform a systematic review and meta-analysis of large clinical trials in order to assess the impact of different therapeutic strategies on the reduction of cardiovascular events and the development of serious adverse effects. Four trials with 29,820 participants were included, with mean age of 65 ± 7.9 years; 42.2% were women and 22% were diabetic. Overall, intensive goals showed a non-significant trend towards reducing cardiovascular mortality (relative risk [RR]: 0.89; 95% confidence interval [CI]: 0.68-1.07; p=0.16) with non-impact in total mortality (p=0.45) and with moderate heterogeneity among the included trials (I index [I]2:44 and 59% respectively). Nevertheless, intensive goals reduce non-fatal cardiovascular events (RR: 0.82; 95% CI: 0.70-0.91; p=0.0003), being consistent in all the analyzed trials (I2: 0%). Regarding adverse effects, intensive goals caused more emergency consultations or hospitalization (RR: 1.98; 95% CI: 1.59-2.46; p<0.0001; I2: 14%), with no clear increase in renal failure (RR: 1.65; 95% CI: 0.94-2.89) but with increase in falls and syncope (RR: 2.39; 95% CI: 1.56-3.67; p<0.0001; I2: 28%). Intensive blood pressure goals reduce non-fatal cardiovascular events, without reduction in mortality, and with an increase of risk of adverse events. These results suggest that individual goals should be set according to the risk of each patient. ­ PROSPERO (CRD42020149134).

Several studies have investigated the effect of Urtica dioica (UD) consumption on metabolic profiles in patients with type 2 diabetes mellitus (T2DM); however, the findings are inconsistent. This systematic review and meta-analysis of clinical trials were performed to summarize the evidence of the effects of UD consumption on metabolic profiles in patients with T2DM. Eligible studies were retrieved from searches of PubMed, Embase, Scopus, Web of Science, Cochrane Library, and Google Scholar databases until December 2019. Cochran (Q) and I-square statistics were used to examine heterogeneity across included clinical trials. Data were pooled using a fixed-effect or random-effects model and expressed as weighted mean difference (WMD) and 95% confidence interval (CI). Among 1485 citations, thirteen clinical trials were found to be eligible for the current metaanalysis. UD consumption significantly decreased levels of fasting blood glucose (FBG) (WMD = - 17.17 mg/dl, 95% CI: -26.60, -7.73, I2 = 93.2%), hemoglobin A1c (HbA1c) (WMD = -0.93, 95% CI: - 1.66, -0.17, I2 = 75.0%), C-reactive protein (CRP) (WMD = -1.09 mg/dl, 95% CI: -1.64, -0.53, I2 = 0.0%), triglycerides (WMD = -26.94 mg/dl, 95 % CI = [-52.07, -1.82], P = 0.03, I2 = 90.0%), systolic blood pressure (SBP) (WMD = -5.03 mmHg, 95% CI = -8.15, -1.91, I2 = 0.0%) in comparison to the control groups. UD consumption did not significantly change serum levels of insulin (WMD = 1.07 μU/ml, 95% CI: -1.59, 3.73, I2 = 63.5%), total-cholesterol (WMD = -6.39 mg/dl, 95% CI: -13.84, 1.05, I2 = 0.0%), LDL-cholesterol (LDL-C) (WMD = -1.30 mg/dl, 95% CI: -9.95, 7.35, I2 = 66.1%), HDL-cholesterol (HDL-C) (WMD = 6.95 mg/dl, 95% CI: -0.14, 14.03, I2 = 95.4%), body max index (BMI) (WMD = -0.16 kg/m2, 95% CI: -1.77, 1.44, I2 = 0.0%), and diastolic blood pressure (DBP) (WMD = -1.35 mmHg, 95% CI: -2.86, 0.17, I2= 0.0%) among patients with T2DM. UD consumption may result in an improvement in levels of FBS, HbA1c, CRP, triglycerides, and SBP, but did not affect levels of insulin, total-, LDL-, and HDL-cholesterol, BMI, and DBP in patients with T2DM.

Real-world data on cardiovascular immune-related adverse events (CV-irAEs) in cancer patients treated with immune checkpoint inhibitors (ICIs) present findings that differ from those reported in meta-analyses of clinical trials. This study aims to estimate the incidence of CV-irAEs from observational studies among patients undergoing ICI therapy for various malignancies and investigate the discrepancy between the results of meta-analyses and observational studies. A systematic literature review and meta-analysis were conducted according to the Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) guidelines. The PubMed database was searched for observational studies that included cancer patients treated with ICIs. Α single-arm meta-analysis using the metaprop command in Stata Statistical Software: Release 16 (StataCorp LLC, College Station, Texas, United States) was performed for the following outcomes: myocarditis, pericardial disease, arrhythmias, cardiac failure, Takotsubo cardiomyopathy, ischemic heart disease, heart valve disease, venous thromboembolism and artery disease. ICI treatment agents were classified into three major classes: PD-1 inhibitors, PD-L1 inhibitors, CTLA4 inhibitors, or their combinations. Heterogeneity was quantified using the I2 statistic and small study effect, and potential publication bias was assessed by inspecting funnel plots, as well as by Egger’s test. A total of 42 studies were included. The incidence of CV-irAEs within the entire population undergoing treatment with ICIs was assessed as follows: total CV-irAEs: 8% (95% confidence interval (CI): 6%, 10%), arrhythmias: 18% (95%CI: 10%, 27%), myocarditis: 11% (95%CI: 5%, 18%), cardiac failure: 8% (95%CI: 2%, 15%), ischemic heart disease: 6% (95%CI: 3%, 11%), pericardial disease: 5% (95%CI: 1%, 10%), artery disease 5% (95%CI: 1%, 12%), and venous thromboembolism: 3% (95%CI: 0%, 8%); cardiomyopathy and heart valve disease had minimal number of observed episodes, thus the pooled incidence results are referring as zero, 0% (95%CI: 0%, 0%) and total CV deaths: 1% (95%CI: 0%, 3%). Median time to CV-irAEs was estimated at 119 days (interquartile range (IQR) 53-180). The most common CV-irAEs were arrhythmias, myocarditis, and cardiac failure with life-threatening complications. Data derived from meta-analyses of clinical trials in most cases indicated that the total incidence of CV-irAEs varied between 0.05% and 1.30%, while in large pharmacovigilance databases, it ranged from 0.125% to 6.7%. In our meta-analysis of post-market surveillance studies, higher estimates were obtained, which offer an insight into the long-term prevalence and outcomes for patients experiencing CV complications associated with ICIs. Longer follow-up period and different definitions of cardiotoxicity may account for the higher cardiotoxicity rates that seem to reflect an emerging threat.