NEJM Formatting Requirements: Complete Author Guide

NEJM's abstract requirements are precise and editors screen them carefully.

• 250 words maximum in four labeled paragraphs: Background (the problem), Methods (how the study was done), Results (the salient findings), Conclusions (what the authors conclude)

• Conclusions should be clinical interpretation, not a restatement of results. What should doctors do differently? If your Conclusions reads like compressed Results, it needs rewriting.

• Trial registration number closes the abstract for clinical trials, and the manuscript must also name the registry. NEJM accepts any ICMJE-approved registry or WHO ICTRP primary registry, not only ClinicalTrials.gov

The 5-item cap on tables and figures is where NEJM formatting becomes genuinely strategic. A randomized trial that shows its CONSORT flow diagram, Table 1 of baseline characteristics, and a Kaplan-Meier curve in the main text has three of five slots gone before the secondary outcomes get a look.

Source: NEJM new-manuscript instructions and Technical Guidelines for Figures and Multimedia (accessed October 8, 2026)

Table rules: Tables go at the end of the manuscript text file, double-spaced, footnotes included, each with a title. P values do not belong in the traditional Table 1 of a randomized trial; note chance imbalances in potential confounders instead. Conventional units go in figures and tables too, with conversion factors to SI in the legends or footnotes.

NEJM redraws all accepted figures. If a paper is accepted, NEJM's production team works with the authors to create every final figure, and an NEJM medical illustrator builds any illustration. Your submission figures must contain all necessary data and labeling, but they don't need to match NEJM's exact published look. Two image rules are strict: disclose every adjustment to brightness, contrast, color balance, cropping, or annotation, and never use generative AI to create or alter content inside a scientific or medical image.

NEJM uses numbered references, numbered consecutively in order of first citation and double-spaced in the list.

1. Author AB, Author CD, Author EF, et al. Title of article. Journal Abbrev Year;Volume:Pages.

Specifics that trip people up:

• Name every author when there are six or fewer; with seven or more, give the first three, then "et al." (a six-author paper gets all six names, a seven-author paper gets three)

• No periods after author initials

• Journal names abbreviated as in NEJM's own samples (for example, N Engl J Med)

• No issue numbers: NEJM's own sample runs "N Engl J Med 2000;343:230-8", with an abbreviated end page

• A reference first cited in a table or figure legend takes its number from the point where that table or figure is first mentioned in the text

• Personal communications, unpublished data, and manuscripts "in preparation" or "submitted" cannot be numbered references; if essential, work them into the text

Worked export check (hypothetical seven-author paper):

Generic manager export: Rivera, J., Okafor, N., Lindqvist, P., Chen, H., Duarte, M., Haddad, S., Moreau, L. Early anticoagulation after ischemic stroke. *N Engl J Med*. 2024;390(12):1101-1110.
NEJM target: Rivera J, Okafor N, Lindqvist P, et al. Early anticoagulation after ischemic stroke. N Engl J Med 2024;390:1101-10.

Five things changed: seven names became three plus et al., the comma-and-period initials collapsed, the journal lost its italics and trailing period, the issue number went, and the end page shortened. A generic Vancouver style in EndNote or Zotero can miss any one of them, so read the first few entries of the exported list by eye.

The 40-reference cap still requires careful curation. Every reference should directly support a specific claim.

NEJM requires a detailed title page with elements many authors forget:

• Title: Concise. NEJM strongly discourages abbreviations anywhere except units, so keep them out of the title.

• Author names with degrees: each author's name, highest degree (MD, PhD, MPH), and affiliation or institution.

• One corresponding author, with contact details. NEJM does not allow two corresponding authors; that person handles every email about forms, authorship, and files. Co-first or co-senior authors are fine, flagged with a footnote on the title page.

• Placed elsewhere in the manuscript, but often forgotten: the trial registration number and registry name, the ethics-approval and written-consent statement where it applies, and who analyzed the data

Disclosure forms: NEJM follows ICMJE, lists study sponsorship, and posts the disclosure form filed by every author alongside the published research article. The form is part of the public record for every co-author, not just the corresponding author.

AI use: authors must say at submission whether AI-assisted tools (large language models, chatbots, image generators) were used to produce the work, and describe that use in both the cover letter and the manuscript.

Author contributions: NEJM follows ICMJE authorship criteria: (1) substantial contributions to conception/design or data acquisition/analysis, (2) drafting or critical revision, (3) final approval, and (4) accountability for the work.

This is where many technically sound manuscripts fail at NEJM.

Protocol and statistical analysis plan (SAP): NEJM's statistical guidelines ask for a protocol and an SAP with every manuscript, and both are required for all interventional trials. Send the original and final versions if they changed, dated, with each change and its rationale described. Observational studies are not exempt: an SAP must be submitted even if it was written after the fact, and it should say whether it came before or after preliminary analyses. If the paper is published, the protocol and SAP are posted with it.

Reporting rules:

• P values two-sided unless the design calls for one-sided tests (noninferiority trials, for example)

• P values above 0.01 to two decimal places, between 0.01 and 0.001 to three, and anything smaller as P<0.001

• Every significance test paired with an effect estimate and its standard error or confidence interval

• A prespecified plan to control type I error for confirmatory end points; without one, secondary end points are reported as point estimates with 95% confidence intervals and no P values

• Per-protocol analyses built by dropping patients after randomization are generally not allowed

• A CONSORT-format flow diagram for randomized trials, with the CONSORT checklist items supplied in the manuscript or appendix

Study registration check: NEJM screens every manuscript at submission to determine whether the study required prospective registration. No manuscript enters the editorial process until it passes this screen. If your trial should have been registered before the first patient enrolled and wasn't, NEJM turns the paper away before peer review.

The NEJM author center gives you the rules. Here's what it doesn't emphasize:

The desk screen is fast. NEJM says initial in-house review is typically complete in under a week. The screen weighs novelty and clinical impact, and a manuscript far over its word limit is returned unread. A perfectly formatted paper on a low-impact topic still gets desk-rejected. But a high-impact paper with sloppy formatting also gets desk-rejected, and that's the preventable failure.

Conventional units first. NEJM wants conventional units for lab values (mg/dL, cells/mm^3) with SI units in parentheses throughout the text. NEJM itself says a manuscript in SI units only will not be sent back for that reason, but converting before submission saves a round of copyediting later.

The limits tighten after the first round. NEJM states that its article-type guidelines are for initial submissions, and that a paper selected for revision will be asked to meet much stricter print and online standards. A first draft that squeaks in at 2,700 words with five dense display items leaves no room for the reviewers' requested analyses.

LaTeX is possible but awkward. NEJM wants one double-spaced file holding the text, references, figure legends, and tables, preferably Word, with plain text (.txt) and Rich Text Format (.rtf) also accepted. A compiled PDF is not on that list. If you use LaTeX, convert to Word and make sure your BibTeX style outputs NEJM's first-three-authors format, not a generic six-author one.

No graphical abstract is requested. Unlike Cell (where it's mandatory), NEJM's instructions list no graphical abstract among the materials to prepare, so don't add one unless the journal asks.

The Supplementary Appendix is not copyedited. NEJM presents it online as the authors supplied it, without editing for style, so every typo and mislabeled table in it goes out under the journal's name. Treat it as part of the paper when you proofread.

The Supplementary Appendix is where most of your detailed work lives at NEJM, and it's more structured than at other journals. The protocol and SAP travel as their own submission documents; the appendix carries everything else.

Order NEJM specifies: a table of contents, then the list of investigators (if there is one), then text such as extended methods, then figures, then tables, then references. The appendix references are self-contained: they are numbered and listed within the appendix, separate from the main paper's list.

Labeling: supplementary tables run Table S1, Table S2, and so on, each with a title and footnotes as needed. Each supplementary figure carries its title and legend on the same page. For any outcome scale, give its range, its direction, and the minimally important difference if known, with a citation for the scale.

Formatting: paginated, and submitted twice, once as a PDF and once in Word or another editable format.

Data sharing statement: NEJM follows the ICMJE requirement for a data sharing statement with clinical trial reports, and trials registered from 2019 on also need a data sharing plan in the registry record. Say what will be shared, with whom, and how.

The Supplementary Appendix is also where you recover from the 5-item limit. NEJM publishes extensive tables and extra material as supplementary content with the online article, so plan your display strategy across both the main manuscript and the appendix.

Three risks sit where NEJM formatting meets editorial judgment. None of them is fixed by a template.

Scope-fit ambiguity in the abstract. NEJM editors move fastest on manuscripts whose contribution is obviously aligned with clinical evidence that could change what a practicing clinician does tomorrow, not eventually. The named failure pattern: a mechanistic paper whose abstract never says what a clinician would do differently, which then has to make that case during an in-house screen NEJM typically finishes in under a week. Check whether your abstract reads to NEJM's scope

An SAP that leaves the primary analysis vague. NEJM's statistical editors review the manuscript against the methods in the protocol and SAP, and they can ask for changes to the analysis as submitted. A trial whose SAP never states the primary end point analysis, or whose Methods skip the sample-size reasoning, invites exactly that request. Check if your methods package is reviewer-complete

Reference-list and clean-citation failure mode. NEJM's instructions do not describe a retraction screen, but a retracted or superseded trial sitting in a 40-reference list is hard to defend once reviewers find it. Check the list against Crossref and Retraction Watch before upload. Check whether your reference list is clean against Crossref + Retraction Watch

Across our NEJM pre-submission reviews, the formatting problems that cost authors a submission cycle are the ones they could have checked in ten minutes. Five come up again and again.

Abstract over 250 words, or built on another journal's headings. NEJM Original Articles need a structured abstract of no more than 250 words, in four labeled paragraphs: Background, Methods, Results, and Conclusions. A draft recycled from a Lancet submission still says "Findings" and "Interpretation," and one recycled from a BMJ journal still has Design, Setting, and Participants as separate headings. Both tell an NEJM editor the paper was written for somewhere else first.

A sixth table or figure in the main text. NEJM's normal limit for an Original Article is five tables and figures in total. The usual way a trial reaches six: CONSORT diagram, Table 1, the primary-outcome figure, a subgroup forest plot, an adverse-events table, and a secondary-outcomes table. Decide which two carry the least weight for a clinician and move them into the Supplementary Appendix before submission, not after the editor asks.

Primary analysis that does not match the prespecified plan. NEJM asks that the analysis of the primary outcome match the final protocol and SAP except in unusual circumstances, and that any departure be justified in the Methods. NEJM's statistical review takes the protocol and SAP into account, so an unexplained change in the primary end point, or a secondary end point promoted to equal billing, is visible the moment the documents are read side by side.

P values where NEJM does not want them. NEJM's statistical guidelines keep P values out of Table 1 of a randomized trial, out of subgroup-interaction tests in forest plots that the multiplicity plan does not cover, and out of hierarchical testing after the first nonsignificant comparison. These are easy to miss because many journals print all three. Strip them before submission and add the absolute event counts the editors prefer to see ahead of relative risks.

Clinical significance framing applies to a narrow specialty population. NEJM's stated interest is original research that changes clinical practice. Submissions whose clinical implications are meaningful primarily to subspecialists in one narrow field, rather than to the broad readership of practicing clinicians across disciplines, are the ones most exposed at the desk. A useful test before you upload: would a hospitalist or primary care physician do anything differently next week because of this paper?

A NEJM formatting and readiness check evaluates manuscript structure, abstract compliance, trial registration consistency, and clinical significance framing against these desk-rejection patterns before you submit.

Or see example reports before you finalize.