If your systematic review or meta-analysis was rejected, the cause is almost always methodological, predictable, and fixable. Journals turn these manuscripts away for a short list of recurring reasons: a question that is not new, an incomplete or unreproducible search, missing independent screening, the wrong risk of bias tool, mishandled heterogeneity, and reporting that falls short of the PRISMA 2020 standard. The rejection feels personal. The reasons behind it are not. They are technical, documented, and correctable.
This guide walks through the reasons evidence syntheses get rejected, each backed by published evidence, and gives you a concrete fix for every one. It also explains how to read your rejection letter, how to respond to a revise and resubmit decision, and how to audit a manuscript before you send it anywhere else.
Quick answer
Most systematic reviews and meta-analyses are rejected for methodological and reporting failures, not for the topic itself. The most common reasons are redundancy with an existing review, an incomplete literature search, no dual independent screening, an inappropriate or missing risk of bias assessment, inappropriate statistical pooling of heterogeneous studies, missing certainty of evidence (GRADE), and incomplete PRISMA 2020 reporting. Each one has a specific, well-documented fix.
First, read the letter: desk rejection or peer-review rejection?
Before you change a single line, work out which kind of rejection you received, because the two point to very different problems.
A desk rejection comes from the editor, usually within days, and without external peer review. It signals a problem the editor could see on a first read: the topic is outside the journal's scope, the question is not new enough to justify another review, the manuscript ignores the journal's formatting or registration requirements, or there is an ethics or reporting issue. Desk rejection is fast and impersonal, and it is rarely about the quality of your analysis because no one has assessed your analysis yet.
A peer-review rejection arrives later, often after weeks or months, and comes with reviewer comments. This is the rejection that tells you something about your methods. Reviewers and methodological editors assess your search, your screening, your risk of bias judgments, your statistics, and your reporting. The fixes here are deeper, but the comments are also a roadmap.
Sort your rejection into one of these two buckets first. A desk rejection often means you targeted the wrong journal or under-justified the need for the review, both of which you can fix quickly. A peer-review rejection means you have methodological work to do before resubmitting anywhere.
The context every author should understand
It helps to know what you are up against. The volume of evidence syntheses has grown faster than their quality, and editors know it.
John Ioannidis of Stanford, in a widely cited 2016 analysis in The Milbank Quarterly, argued that the production of systematic reviews and meta-analyses had reached epidemic proportions and that many function mostly as publishable units or marketing tools rather than as useful evidence. His estimate that only about 3 percent are both non-redundant and clinically useful is a deliberately provocative figure, and it has been debated since, so treat it as an expert's estimate rather than a measured rate. The direction of the argument, though, is not seriously disputed.
The more recent data is harder to wave away. A 2024 meta-research study by Bojcic, Todoric, and Puljak in the Journal of Clinical Epidemiology examined systematic reviews that explicitly claimed to follow AMSTAR 2, the standard appraisal tool. Of the 43 reviews assessed, 35 were rated critically low confidence, 7 were low, and only one was high. None reached moderate. In other words, even reviews that advertised their own rigor mostly failed to deliver it.
For you, the author, this is the takeaway: editors and reviewers approach evidence syntheses expecting weaknesses. A manuscript that anticipates and closes those weaknesses stands out precisely because so many do not.
The reasons systematic reviews and meta-analyses get rejected, and how to fix each one
1. The question is not new, or the review is not needed
The fastest route to a desk rejection is proposing a review that already exists. Editor's check. If a recent systematic review answers your question, or if your question is so broad that it answers nothing in particular, the manuscript stops at the editor's desk.
The fix. Before you commit, run a scoping search of PROSPERO and the major databases to confirm no current review covers your exact question. If one exists, sharpen your question so it addresses a real gap, a different population, a newer body of evidence, or an unresolved inconsistency. State that gap explicitly in your introduction. Register your protocol on PROSPERO before you begin searching, which both protects against duplication and signals to editors that the review is planned rather than retrofitted.
2. The literature search is incomplete or cannot be reproduced
This is the single most common methodological reason reviews are sent back. A 2019 survey of librarian methodological peer reviewers, published in Research Integrity and Peer Review, found that the most frequently cited basis for recommending rejection or revision was the search methodology itself. The search is the foundation, and reviewers look there first.
The scale of the problem is well documented. Bramer and colleagues, writing in Systematic Reviews in 2017, estimated that around 60 percent of published systematic reviews fail to retrieve 95 percent of the relevant references, largely because authors searched too few databases. Their analysis found that the combination of Embase, MEDLINE, Web of Science Core Collection, and Google Scholar achieved 98.3 percent recall, while thinner searches missed substantial portions of the evidence.
The fix. Search at least Embase, MEDLINE, Web of Science, and Google Scholar as a baseline, and add a topic-matched specialist database such as CINAHL for nursing and allied health or PsycINFO for behavioral and mental health work. Bring in an information specialist or librarian to build and peer-review the search string. Report the full strategy for every database, including dates and result counts, so a reader could reproduce it exactly. Reporting the search against the PRISMA-S extension closes most of the gaps that reviewers flag.
3. There was no independent dual screening or extraction
A systematic review is not a solo exercise, and reviewers know that a single person screening and extracting introduces error and bias that no amount of care can fully remove. Major research libraries are explicit that screening and data extraction must be done independently by at least two people. A manuscript that describes one person doing this work invites rejection on rigor alone.
The fix. Use at least two reviewers who screen titles, abstracts, and full texts independently, with a third person or a consensus process to resolve disagreements. Report your inter-reviewer agreement. Manage the workflow in software built for it, such as Covidence or the free tool Rayyan, both of which support blinded dual screening and document the process for your methods section.
4. The risk of bias assessment is missing or uses the wrong tool
Reviewers reject manuscripts that skip quality appraisal and, almost as often, manuscripts that appraise studies with the wrong instrument. Applying a tool built for observational studies to randomized trials, or vice versa, tells a methodological reviewer that the authors did not understand the assessment.
The fix. Match the tool to the design. Use the Cochrane risk of bias tool, RoB 2 (Sterne and colleagues, BMJ, 2019), for randomized trials, and the Newcastle-Ottawa Scale for cohort and case-control studies. Assess risk of bias at the level of the individual result, not just the whole study, and carry those judgments through into your synthesis and conclusions rather than reporting them and then ignoring them. AMSTAR 2 (Shea and colleagues, BMJ, 2017) and ROBIS (Whiting and colleagues, Journal of Clinical Epidemiology, 2016) appraise the review itself, and reading your own manuscript against them before submission is one of the highest-value checks you can run.
5. The meta-analysis pools studies that should not be pooled
Statistical problems sink meta-analyses. The most damaging is combining studies that are too clinically or methodologically different into a single pooled estimate, then reporting that estimate as if it meant something. Closely related failures include ignoring high heterogeneity, choosing a fixed-effect model when a random-effects model is warranted, and running a meta-analysis at all when a narrative synthesis would have been more honest.
The fix. Justify pooling before you do it. If the populations, interventions, or outcomes are not reasonably comparable, do not force a combined estimate. Report the I-squared statistic with its confidence interval and interpret it rather than burying it, and bear in mind that I-squared is unstable when a meta-analysis contains only a handful of studies. Prespecify your choice between fixed-effect and random-effects models in the protocol, and follow the Cochrane Handbook's guidance on when each is appropriate.
6. There is no certainty of evidence rating or publication bias assessment
Modern reviewers expect a certainty of evidence assessment, and its absence reads as incomplete. They also scrutinize publication bias analyses, and one specific error recurs constantly: drawing a funnel plot or running a test for funnel-plot asymmetry when there are too few studies to support it.
The fix. Rate the certainty of evidence for each major outcome using GRADE, which classifies confidence as high, moderate, low, or very low based on risk of bias, imprecision, inconsistency, indirectness, and publication bias. For publication bias, follow the Cochrane Handbook and avoid funnel-plot asymmetry tests when fewer than ten studies are included, since with small numbers, the tests cannot distinguish real bias from chance or from genuine differences between large and small studies.
7. The conclusions overreach the evidence
Reviewers call this spin, and they are quick to spot it. A review that finds low-certainty or mixed evidence but concludes with a confident clinical recommendation undermines its own credibility. Overstated abstracts are a frequent trigger for rejection because the abstract is the first and sometimes only thing an editor reads closely.
The fix. Calibrate every claim to your GRADE certainty rating. If the evidence is of low certainty, say so plainly in the abstract and the conclusion. Report limitations in proportion to their seriousness, and resist the urge to recommend practice changes that your evidence base cannot support.
8. The reporting does not meet PRISMA 2020
Even a methodologically sound review gets rejected when it is reported poorly. Missing checklist items, an inaccurate flow diagram, or numbers that do not reconcile across the text, tables, and diagram all signal carelessness to a reviewer.
The fix. Report against the full PRISMA 2020 checklist (Page and colleagues, BMJ, 2021), all 27 items, and include a flow diagram whose numbers reconcile at every stage from identification through screening to inclusion. Use PRISMA-P for the protocol. Treating PRISMA 2020 as a writing checklist rather than a box to tick at the end removes a whole category of reviewer objections.
9. There is no protocol or prospective registration
Many journals now expect prospective registration, and its absence raises questions about whether the methods were decided before or after the authors saw the data. Registration is not always mandatory, but skipping it weakens the manuscript and, at some journals, ends it.
The fix. Register the protocol on PROSPERO before you begin your searches. Note that PROSPERO will not accept a registration once searching has started, and that scoping reviews are not eligible and should follow the JBI approach instead. Disclose any deviation from the registered protocol transparently in the manuscript.
Reasons for rejection at a glance
Table 1: Common reasons why Systematic Reviews get rejected
How to respond to a revise and resubmit
A revise and resubmit is not a rejection, and a rejection from one journal is not the end of the manuscript. Most evidence syntheses that are eventually published were turned away at least once.
Respond to the reviewers ' points point by point. Reproduce each comment, then answer it directly beneath, and state exactly what you changed and where, with page and line numbers. Where you disagree, say so respectfully and back your position with evidence and citation rather than assertion. Make the changes visible with tracked changes or a clearly marked revised file. Reviewers are far more forgiving of authors who engage seriously with every comment than of authors who address the easy points and quietly skip the hard ones.
If the rejection was a desk rejection on scope or novelty, the right move is often a different journal rather than a rebuttal. Match the manuscript to a journal whose readership and aims fit your question, and rework the framing so the contribution is unmistakable in the first paragraph.
A ten-minute pre-submission self-audit
Before you submit anywhere, run the manuscript against this list:
The question addresses a documented gap, and no current review already answers it.
The protocol was registered on PROSPERO before searching began.
The search covered at least four major databases plus any topic-specific database, and the full strategy is reported for each.
Screening and extraction were done independently by at least two people, and agreement is reported.
Risk of bias was assessed with the correct tool for the study designs included.
Any meta-analysis justifies its pooling, reports I-squared with interpretation, and prespecifies its model.
Publication bias was assessed appropriately, with no funnel-plot test under ten studies.
Certainty of evidence was rated with GRADE for each major outcome.
Conclusions match the certainty of the evidence, with no spin.
The manuscript meets all 27 PRISMA 2020 items, and the flow diagram numbers reconcile.
If you cannot check every box, fix the gap before a reviewer finds it.
When professional support is worth it
The stages that most often trigger rejection, the search strategy, the statistical synthesis, and PRISMA-compliant reporting, are also the stages that reward specialist expertise the most. A second methodological read before submission catches the issues reviewers would otherwise catch for you, and it does so while you can still fix them quietly. If you want an experienced evidence-synthesis specialist to review your search, your meta-analysis, or your reporting before you resubmit, ScribeLabWriter works with researchers preparing systematic reviews and meta-analyses for Tier 1 and Tier 2 journals. The goal is simple: a manuscript that gives reviewers nothing easy to reject.
Frequently asked questions
Can I resubmit a rejected systematic review to a different journal?
Yes. A rejection from one journal does not prevent submission to another, and most published reviews were rejected at least once. Use the reviewer comments to strengthen the methods first, then match the manuscript to a journal whose scope fits your question.
Is a desk rejection final?
For that journal, usually yes, unless the editor invites a reframed resubmission. A desk rejection most often reflects scope, novelty, or a formatting and registration issue rather than the quality of your analysis, so the practical response is to fix the framing and choose a better-matched journal.
Does PROSPERO registration stop my review from being rejected?
No single step prevents rejection, but prospective registration removes one common objection and signals that your methods were planned in advance. Registration is increasingly expected, and its absence weakens the manuscript at many journals.
How many databases do I need to search to avoid rejection on the search?
Evidence points to at least four, namely Embase, MEDLINE, Web of Science, and Google Scholar, plus a topic-specific database where relevant. Searching fewer is the most common reason reviews fail to retrieve the full evidence base.
Can one person complete a systematic review?
A defensible systematic review requires at least two people for independent screening and data extraction, because dual review reduces error and bias. A manuscript describing single-reviewer screening is likely to be rejected on methodological grounds.
