Retracted randomized trials from super-retractors and top-cited scientists with multiple retractions

This study reveals that a small, concentrated group of highly influential authors, particularly "super-retractors" and top-cited scientists with multiple retractions, are disproportionately responsible for a significant portion of retracted randomized clinical trials, which tend to be older, take longer to retract, and accumulate more citations than other retracted studies.

The Gist

Imagine the world of medical research as a massive, bustling library. Inside this library are millions of books (scientific studies) that doctors use to decide how to treat patients. Most of these books are trustworthy, but some are "fake books" filled with made-up stories or stolen ideas. When a fake book is discovered, the library staff (journals) pulls it off the shelf and puts a big red "RETRACTED" stamp on it. This is called a retraction.

This paper is like a detective report investigating who is writing the most of these fake books and how they are getting away with it for so long.

Here is the breakdown of the findings using simple analogies:

1. The "Super-Authors" (The Super-Retractors)

The researchers found that a tiny group of people are responsible for a huge chunk of the fake books.

• The Analogy: Imagine if 6 out of 100 authors in the entire library were responsible for writing 20% of all the fake books. That's like finding out that a few specific "ghostwriters" are churning out bad stories at a factory rate.
• The Finding: They identified 6 "Super-Retractors" (mostly from Japan, working in anesthesia and diabetes research) who co-wrote one-fifth of all the retracted medical trials. Another group of 18 famous scientists (who usually get a lot of praise and citations) also co-wrote one-quarter of these bad trials.

2. The "Zombie" Books

The paper talks about "Zombie RCTs."

• The Analogy: These are books that should have been burned immediately because the story makes no sense, but they stayed on the shelves for years. They are "undead" because they keep influencing doctors even though they are known to be false.
• The Finding: The books written by these Super-Authors were published a long time ago (around the year 2000) and didn't get pulled off the shelf until much later (around 2013). They were "zombies" for a long time, tricking people into reading them.

3. The "Fame Trap" (Citations)

You might think fake books get ignored, but the opposite happened.

• The Analogy: Because these authors were famous "celebrities" in the science world, their books were borrowed and read more than the average book, even before anyone knew they were fake. It's like a famous chef serving a poisoned meal; everyone eats it because they trust the chef, not realizing the food is bad until it's too late.
• The Finding: The retracted papers by these famous authors had more citations (people reading and referencing them) than the retracted papers by unknown authors. They were popular, but in a dangerous way.

4. The "Slow Motion" Problem

• The Analogy: Imagine a smoke alarm that takes 10 years to go off.
• The Finding: The papers by these Super-Authors took a very long time to be retracted (about 14 years from publication to retraction). The papers by regular authors were caught much faster (about 1.5 years). This suggests that because these authors were so famous and powerful, the library staff was slower to investigate them, or it took longer to gather the evidence to prove they were faking it.

5. The "Clan" Effect

• The Analogy: These bad actors didn't work alone; they worked in tight-knit families or clans.
• The Finding: Many of these Super-Authors were frequent co-authors with each other. If you pulled out one bad book, you often found the same names on the cover of other bad books. They were concentrated in specific fields (like Anesthesia) and specific countries (mostly Japan and the US).

The Big Takeaway

The main message is that bad science isn't spread evenly. It is concentrated in the hands of a few "super-famous" authors who often work together.

Because these authors were so influential, their fake studies got more attention, stayed on the shelves longer, and did more damage to medical guidelines than the fake studies written by unknown researchers. The paper suggests that to clean up the library, we need to pay extra attention to these specific "celebrity" authors and their close circles, rather than just looking at the library as a whole.

In short: A small group of famous scientists wrote a massive amount of fake medical research. Because they were famous, their fake work was read more, believed longer, and took much longer to be exposed than the fake work of lesser-known scientists.

Technical Summary

1. Problem Statement

Retracted Randomized Controlled Trials (RCTs) pose a severe threat to the evidence ecosystem. They distort systematic reviews and meta-analyses, which subsequently influence clinical practice guidelines. While guidelines (e.g., COPE) suggest that "zombie" RCTs (those with fabricated data or lack of credibility) should be retracted, many remain in the literature.

Previous research has identified "super-retractors" (authors with the highest number of retractions) and "top-cited scientists" who have accumulated multiple retractions not due to publisher error. However, there was no quantitative estimation of the specific contribution of these influential authors to the corpus of retracted RCTs. The study aims to determine if a small subset of highly influential authors disproportionately drives the volume of retracted clinical trial literature.

2. Methodology

Study Design: Retrospective cohort study following STROBE guidelines.

Data Sources:

• VITALITY Cohort: A complete, openly available dataset of 1,330 retracted RCTs identified from Retraction Watch (updated November 2024).
• Author Lists: Three distinct groups of scientists were linked to the VITALITY cohort (updated August 2025):
  1. Super-retractors (Group 1): The top 30 researchers from the Retraction Watch Leaderboard (33–221 retractions each).
  2. Career-long Top-cited (Group 2): 163 scientists in the top 2% of citation impact over their entire career who had $\ge$10 retractions (excluding publisher errors).
  3. Single-Year Top-cited (Group 3): 174 scientists in the top 2% of citation impact in 2024 who had $\ge$10 retractions.

Data Extraction & Analysis:

• Linking: Researchers manually and algorithmically matched full author lists of the 1,330 retracted RCTs against the three author groups.
• Variables: Publication year, retraction year, time lag (days) between publication and retraction, and total citation counts (Scopus, Nov 2025).
• Statistical Methods:
  • Descriptive statistics (counts, proportions, medians, IQR).
  • Wilcoxon rank-sum tests for comparing continuous variables between groups.
  • Pearson correlation with Sidak correction for multiple comparisons.
  • Multivariable Linear Regression: To disentangle the effect of author status from "survival bias" (i.e., older papers naturally accumulate more citations). The model included author status, time to retraction, and their interaction term.

3. Key Contributions

• Quantification of Influence: The study provides the first comprehensive quantification of how much of the retracted RCT literature is attributable to a tiny fraction of "super-retractors" and highly cited scientists with multiple retractions.
• Overlap Analysis: It reveals a massive overlap between super-retractors and top-cited scientists, suggesting that the same individuals dominate both high-impact citation metrics and retraction counts.
• Temporal Dynamics: It characterizes the lifecycle of these retracted papers, showing they are published earlier, take significantly longer to retract, and accumulate more citations than other retracted trials.
• Field Concentration: It identifies specific medical subfields (Anesthesiology, Endocrinology & Metabolism) and countries (Japan, USA) where these retractions are disproportionately concentrated.

4. Key Results

Author Contribution:

• Super-retractors: 6 out of 30 super-retractors co-authored 290 (22%) of the 1,330 retracted RCTs. These 6 authors were from Anesthesiology and Endocrinology & Metabolism; all but one were affiliated with Japanese institutions.
• Career-long Top-cited: 18 out of 163 top-cited scientists (with $\ge$10 retractions) co-authored 327 (25%) of the retracted RCTs.
• Overlap: There was a significant intersection: 275 (84%) of the trials co-authored by career-long top-cited scientists were also co-authored by a super-retractor.
• Single-year Top-cited: 7 scientists co-authored 50 trials (4%); all were also in the career-long group.

Characteristics of Retracted Articles:
Articles involving super-retractors or top-cited scientists differed significantly from those without such authors:

• Publication Year: Published much earlier (Median: 2000 vs. 2020).
• Retraction Year: Retracted earlier (Median: 2013 vs. 2023).
• Time Lag: Took significantly longer to retract (Median: 5,111 days vs. 482 days).
• Citations: Accumulated more citations (Median: 21 vs. 5).

Regression Analysis:

• In multivariable models, time to retraction was the only factor significantly and positively associated with total citations ($\beta = 0.02, P < 0.001$).
• The presence of a super-retractor was not independently significant in the regression ($P = 0.098$) once time to retraction was controlled for.
• Interaction Effect: There was a significant negative interaction between time to retraction and author status. For non-super-retractors, citations increased over time; for super-retractors, this trend was absent, suggesting their papers were cited heavily early on but did not continue to accumulate citations at the same rate relative to their age.

5. Significance and Implications

• Systemic Concentration: A small number of influential authors, often co-authors of each other, account for a disproportionate share (approx. 25%) of retracted RCTs. This suggests that misconduct or systemic issues are not random but concentrated in specific networks, subfields, and geographic regions.
• The "Survival" Bias: The high citation counts of retracted papers by super-retractors are largely driven by the fact that these papers were published decades ago and remained in the literature for a long time before retraction. This highlights a critical window where flawed evidence influences clinical practice and guidelines.
• Scrutiny and Detection: The findings suggest that the work of super-retractors has been thoroughly scrutinized (leading to many retractions), whereas top-cited scientists with fewer retractions may still harbor undetected issues.
• Policy Recommendations: The study implies that monitoring author networks and specific subfields (like Anesthesiology) is more effective than random screening. It also underscores the need for faster retraction processes to prevent the accumulation of citations for fraudulent or flawed data.
• Limitations: The study relies on retracted papers as a proxy for misconduct, which may miss "zombie" trials that haven't been retracted yet. Additionally, citation data does not distinguish between citations before and after retraction.

Conclusion: The retracted RCT literature is heavily skewed toward a small, interconnected group of highly cited scientists, primarily from Japan and the US, concentrated in Anesthesiology and Endocrinology. These papers are older, take longer to retract, and have historically garnered more citations, posing a persistent risk to medical evidence synthesis.