Changelog

Source: NEWS.md

egfr 2.0.0

This release corrects two equations that returned clinically wrong values for paediatric patients, and removes one function that implemented an experimental equation.

Breaking changes

• egfr_ekfc_cr() now uses the correct reference Q specification published in Pottel et al. (2021): the age polynomial applies to ages 2-25 inclusive, with the adult Q only above age 25. Verified against the source paper and cross-validated against the eGFR.app implementation.
• egfr_fas_cr() now applies the published age- and sex-specific Q values from Table 1 of Pottel et al. (2016). Previous versions used the adult Q (0.90 / 0.70 mg/dL) at every age, which overestimated eGFR by up to ~246% in infants and ~76% in a 10-year-old. Q is shared between boys and girls for ages 1-14, becomes sex-specific for ages 15-19, and reaches the adult values from age 20. Adult results (age >= 20) are unchanged.
• egfr_ckid_u25_cr_extended() has been removed. Its kappa extension to age 30 (0.995^(age - 25)) was an experimental feature. Use egfr_ckid_u25_cr() within its validated range.

Bug fixes

• ckd_stage() no longer errors on a logical NA (e.g. ckd_stage(NA)) and returns NA as expected.

Documentation

• Corrected the citation for egfr_neonatal().
• egfr_ckdepi_cys_2021() now documents its use of the four-decimal age factor 0.9962 from Table 2 of Inker et al. (2021), and the resulting ~1.2-1.6% difference at ages 60-80 relative to calculators that use the rounded 0.996 printed in the 2012 paper and by the National Kidney Foundation.
• Added published validity ranges to the documentation of egfr_ekfc_cr() (ages 2-90; the previous claim of 2-120 was incorrect), egfr_ekfc_cys(), egfr_fas_cr(), egfr_bis_cr() (age >= 70), egfr_lund_malmo() (adults), egfr_schwartz() (ages 1-16), egfr_ckid_u25_cr() (ages 1-25), egfr_capa(), and egfr_neonatal() (term-born, postnatal days 0-28).
• egfr_cockcroft_gault() now documents that it returns creatinine clearance (which overestimates GFR through tubular secretion) and was derived using non-standardised Jaffe creatinine.

Testing

• Added test-reference-values.R, which validates every equation against coefficients written out directly from the original publications, including the EKFC and FAS reference Q tables and continuity checks at age knots. The previous suite largely re-derived values using the package’s own internals and could not detect a mis-transcribed equation.

egfr 1.1.2

• The CITATION file now reports the installed package version dynamically instead of a hardcoded value.
• Added the maintainer’s ORCID and Zenodo archive metadata (.zenodo.json) so the package can be cited via a DOI.

egfr 1.1.0

• Fixed the Full Age Spectrum (FAS) creatinine equation (egfr_fas_cr()) for ages over 40: the creatinine-ratio exponent is now correctly -1 (a plain reciprocal) as published in Pottel et al. (2016), rather than -0.88. This changes returned eGFR values for patients older than 40.
• egfr_ekfc_cr(), egfr_ekfc_cys(), and egfr_ekfc_cr_cys() gain an optional q argument (q_cr / q_cys for the combined equation) to supply a population-, assay-, or individual-specific reference Q value. The default (NULL) continues to use the built-in EKFC reference Q, so existing code is unaffected.

egfr 1.0.0

egfr 0.1.0

• Initial release.
• 20 eGFR / creatinine-clearance equations: CKD-EPI 2021 (creatinine, cystatin C, combined), CKD-EPI 2009, CKD-EPI 2012 cystatin C, MDRD, Cockcroft-Gault, EKFC (creatinine, cystatin C, combined), FAS, Lund-Malmoe, Berlin Initiative Study, Schwartz bedside, CKiD U25 (creatinine, cystatin C, combined, extended), CAPA, and neonatal eGFR.
• Helpers: bsa(), ckd_stage(), and creatinine unit conversion.