PubMed 2025 Baseline Released

The National Library of Medicine (NLM) has released the 2025 production year PubMed baseline files. The complete 2025 baseline consists of files pubmed25n0001 through pubmed25n1274.xml. In addition to the XML files there are corresponding MD5 checksum files for each XML export file.

The first set of 2025 PubMed Update files have been posted. The first 2025 update file is pubmed25n1275.xml. Regular daily updates have resumed.

NOTE: All baseline files must be downloaded and processed PRIOR to loading the first and subsequent update files.

DTD Changes: On September 28, 2024, NLM announced the release of the 2025 PubMed DTD.

Additions/Changes since pubmed_240101.dtd:

1. Updated DescriptorName model to allow for optional AutoHM="Y" attribute.


2. Updated QualifierName model to allow for optional AutoHM="Y" attribute.

The AutoHM attribute will be used to distinguish between Descriptors/Qualifiers that have been added by curators or MTIX based on information in the article text (attribute not present), and Descriptors/Qualifiers that have been added automatically after a supplementary concept record (SCR) was indexed (AutoHM="Y"). Learn more about MeSH record types.


3. Updated IndexingMethod attribute to include allowed value "Manual" for fully human-indexed citations. This attribute has been retroactively applied to manually indexed citations in PubMed.

Please see pubmed_250101.dtd Documentation for annotations and examples for elements and attributes defined in the PubMed DTD. Learn more about downloading PubMed data.

If you have questions regarding this process, please write to the NLM Help Desk.