Neonatal outcomes after prenatal exposure to selective serotonin reuptake inhibitors: extraction dataset and analysis code
Year: 2026
Type: dataset
Abstract: The complete data extraction and analysis code for the systematic review and meta-analysis "Neonatal outcomes after prenatal exposure to selective serotonin reuptake inhibitors" (PROSPERO CRD420251243990). The registered searches were run on 14 August 2026 from database inception, and all four registered sources were searched again on 7 September 2026 on a broadened exposure string adding antidepressant terms, also from inception. 127 observational studies drawing on 82 independent data sources were included. SSRI_extraction_dataset.xlsx holds the extraction in full. The Estimates sheet has one row per extracted estimate, 844 in all, of which 514 are the adjusted estimates that enter the analyses. Every one carries the verbatim sentence or table row it was taken from and the table or figure number in the source paper, so any extraction can be checked against its source without contacting the authors. Further sheets hold study characteristics, Newcastle-Ottawa domain scores with the adjustment-set assessment, and the reports excluded at full text with their reasons. SSRI_zenodo_deposit.zip holds the same dataset, the native JSON file the analysis reads, the three analysis output files (pooled estimates, subgroup and sensitivity results, GRADE judgements), and the Python scripts that produce every estimate, table and figure in the paper, with a README giving the order to run them in. SSRI_zenodo_README.md is that README, provided separately so it can be read without downloading the archive. Python 3.11 with numpy, scipy, pandas and matplotlib. No proprietary software and no random seeds: every result is deterministic and reproducible from the deposited files alone. Version 1.14, 11 September 2026: incorporates the broadened re-search. The registered exposure block named the individual SSRIs and the phrase itself but not the word antidepressant, while the review’s prespecified rules admit reports defining exposure more broadly and assign them to a mixed-exposure stratum, so the search was narrower than the eligibility criteria it fed. All four registered sources were searched again on 7 September 2026, identifying 18448 records of which 4456 were screened in duplicate, together with 491 records carrying no abstract whose title alone names both an antidepressant and pregnancy. Of the 898 unique reports sought, 630 were removed on record type before retrieval, 199 of the remaining 268 were retrieved and assessed at full text in duplicate, 69 could not be retrieved through any institutional or open-access route and are listed in supplementary table S7d rather than assumed ineligible, and 16 studies were added. The review moves from 111 studies on 74 data sources to 127 on 82. Every newly included study was assessed for participant overlap against all 111 studies already in the review and against the other fifteen. Seven cluster decisions follow, all recorded in supplementary table S8: Laine 2019 joins the Finnish national cluster; Rommel 2024 and Munk-Olsen 2021 join the Danish cluster, the latter having previously carried no cluster label at all; new clusters are created for Kaiser Permanente Northern California (Cornet 2024, Li 2025, Nance 2023), for the Taiwan National Health Insurance Research Database (Chen 2021, Su 2023) with the representation split by analysis, and for the Lombardy healthcare utilisation databases (both Cantarutti reports); Ankarfeldt 2023 joins the multi-country Nordic cluster; and Wartko 2021 and Simon 2002 are deliberately not clustered, sharing an institution but no calendar window. Two results change materially. Mean birth weight against an unexposed comparator moves from -56 g (-90 to -22) on nine data sources to -60 g (-85 to -34) on twelve, and no longer depends on any single study: every leave-one-out estimate lies between -78 g and -39 g, where previously omitting one source returned it to -34 g (-69 to +1). The prespecified SSRI-specific analysis gives -39 g (-67 to -10) on eleven sources with an I-squared of 0.0%. This body of evidence moves from very low to low certainty. Caesarean delivery against an unexposed comparator moves from 1.10 (0.97 to 1.24) on five sources to 1.11 (1.01 to 1.23) on six and now excludes no effect, but four of the six leave-one-out analyses return it to no effect, and it is reported with that fragility stated. Persistent pulmonary hypertension after gestational week 20 is 2.59 (1.25 to 5.37) on five sources and all windows 1.77 (1.18 to 2.66) on nine, following the admission of Munk-Olsen 2021 decided on 10 September 2026. Two further changes are reported as changes rather than presented as though the earlier result had not occurred. The exposure-timing stratification no longer resolves most of the heterogeneity in the pulmonary hypertension analysis: I-squared is 68.6% in the late stratum against 63.9% undifferentiated. And Egger’s test no longer crosses the prespecified threshold for any analysis; preterm birth falls from z = 2.05 on 22 data sources to z = 1.55 on 24, so the publication-bias downgrade is removed. The archive gains supplementary tables S7c (the 183 wave 4 full-text exclusions with reasons), S7d (the 69 reports that could not be retrieved, with the status of the last attempt) and S10 (comparator C, the three adjusted estimates that exist and why none is pooled), together with merge_wave4.py, which applies the extraction rulings, and reconstruct_w4_prisma.py, which recomputes the identification cascade of the broadened search from the raw harvests. Version 1.3, 3 September 2026: adds Bellissima 2015 (ILLiad-retrieved after version 1.2), included with six unadjusted outcome rows, none pool-eligible; Bellissima 2020 recorded as a duplicate report of that cohort. No pooled estimate changes. Version 1.4, 3 September 2026: corrects the Apgar below 7 at five minutes threshold subgroup, which now rests on nine data sources at 2.06 (1.69 to 2.51) with an I-squared of 87%, after Goaz Melet 2026 was correctly recorded as using a below-7 threshold; removes a GRADE large-effect upgrade that had been applied to a body of evidence already carrying downgrades, which GRADE guidance does not permit, with no resulting change to any certainty rating; fills four study-characteristic fields left blank in version 1.3 (Manakova 2011, Salisbury 2011, Itkonen 2025 and the Ericson 1999 Newcastle-Ottawa note); and normalises country naming and tidies the analysis scripts. No pooled estimate changes. Version 1.5, 3 September 2026: adds the verbatim source quotations for the three comparator B estimates from Goaz Melet 2026 (preterm birth, Apgar below 7 at five minutes and neonatal care admission), which had been recorded with their arm counts and adjusted values but without the quotation, and corrects their effect measure from OR to aOR; all three were verified against table 3 of the source paper and no value changed. Adds make_table_s9.py, which generates the new supplementary table S9 printing every estimate that enters a pooled analysis with its source. Adds build checks for the country count, the number of bodies of evidence downgraded for risk of bias, the presence of a source quotation on every pooled input, and agreement between the scripts in this deposit and those printed in the paper. No pooled estimate changes. Version 1.7, 4 September 2026: makes the archive byte-reproducible, so that the copy submitted with the manuscript and the copy deposited here can be compared rather than merely assumed to match. Two things had made every rebuild produce a different file from identical inputs: openpyxl stamps the workbook with the moment it was written, and the archive was assembled in filesystem order with live modification times. The workbook properties are now pinned and the saved file normalised, and archive entries are written in sorted order with a fixed timestamp. Three consecutive rebuilds now produce the same file. No data, no analysis output, no script and no result changes; the extraction workbook differs from version 1.6 only in that internal timestamp. This version of the archive is the one supplied with the manuscript as supplementary file 3. Version 1.13, 7 September 2026: the retrieval arm of the flow diagram is corrected and the excluded-report table completed. Fifteen priority-set reports had been recorded as not retrieved. Every one has now been resolved: four were already held, two more were obtained (a Pneumologie news summary through the publisher and an American Journal of Nursing column located at article level by the library link resolver), eight are congress abstracts whose supplement registers a single identifier for the whole volume so that the abstract is the complete published report, and one is a German-language item excluded under the registered language criterion, which is determinable from the record. None was unobtainable. The priority arm therefore reads 119 reports sought, 0 not retrieved, 3 duplicate reports, 116 assessed and 27 excluded, and table S7 now cites all 27 with reasons rather than 12. Congress abstracts published without an author list or a per-abstract identifier are cited by Web of Science accession and are marked as unattributed, and the citation gate accepts that form. Bellissima 2020 is recorded as a duplicate report of Bellissima 2015: the full text, obtained by interlibrary loan, confirms a paroxetine-specific re-analysis of the same February 2010 to March 2018 cohort, a cross-sectional design, and a control group defined as term deliveries and so selected on the absence of preterm birth. No study count, pooled estimate or result changes. Version 1.12, 6 September 2026: one extraction correction, found by the duplicate extraction of every pooled input. Malm 2026 is removed from all preterm birth analyses. The study publishes no adjusted estimate for preterm birth before 37 weeks; every preterm estimate it reports belongs to one of two disjoint bands, late preterm 32 to 36+6 weeks or very pre
Language: English
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7204649304