{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T06:08:24Z","timestamp":1782713304476,"version":"3.54.5"},"reference-count":38,"publisher":"MDPI AG","issue":"3","license":[{"start":{"date-parts":[[2026,2,1]],"date-time":"2026-02-01T00:00:00Z","timestamp":1769904000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Diagnostics"],"abstract":"<jats:p>Background: Unmeasured ions (UIs) contribute significantly to acid\u2013base disturbances in critically ill patients, yet the optimal parameter for their estimation remains uncertain. The most widely used indicators are the albumin-corrected anion gap (AGc), the strong ion gap (SIG), and the base excess gap (BEGap). Methods: In this retrospective cohort study, a total of 2274 ICU patients (2018\u20132022) were included in the development cohort, and an independent external validation cohort of 1202 patients (2023\u20132025) was used to assess temporal generalizability. Three approaches to blood gas analysis\u2014traditional (PaCO2, HCO3\u2212, AGc), Stewart (PaCO2, SIDa, ATOT, SIG), and partitioned base excess (PaCO2, BECl, BEAlb, BELac, BEGap)\u2014were evaluated. Multivariable linear regression (MLR) and machine learning (ML, random forest [RF], extreme gradient boosting [XGBoost], and support vector regression [SVR]) were applied to evaluate the explanatory performance of analytical approaches with respect to arterial pH. Model performance was assessed using adjusted R2, RMSE, and MAE. Variable importance was quantified with tree-based methods, SHAP values, and permutation importance. All modeling and reporting steps followed the PROBAST-AI guideline. Results: In multiple linear regression (MLR), the partitioned base excess (BE) approach achieved the highest explanatory performance (adjusted R2 = 0.949), followed by the traditional (0.929) and Stewart approaches (0.926). In ML analyses, model fit was high across all approaches. For the traditional approach, R2 values were 0.979 with RF, 0.974 with XGBoost, and 0.934 with SVR. The Stewart\u2019s approach showed lower overall explanatory performance, with R2 values of 0.876 (RF), 0.967 (XGBoost), and 0.996 (SVR). The partitioned BE approach again demonstrated the strongest explanatory performance, achieving R2 values of 0.975 with XGBoost and 0.989 with SVR. Across all analytical models, BEGap consistently emerged as a strong and independent determinant of arterial pH, outperforming SIG and AGc. SIG showed an intermediate contribution, whereas AGc provided minimal independent explanatory value. Among ML models, XGBoost showed the most stable and accurate explanatory performance across approaches. Conclusions: This study demonstrates that BEGap is a practical, physiologically informative, and bedside-applicable parameter for assessing unmeasured ions, outperforming both AGc and SIG across linear and non-linear analytical models.<\/jats:p>","DOI":"10.3390\/diagnostics16030427","type":"journal-article","created":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T14:12:46Z","timestamp":1770041566000},"page":"427","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Unveiling the Gaps: Machine Learning Models for Unmeasured Ions"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0534-7973","authenticated-orcid":false,"given":"Furkan","family":"Tontu","sequence":"first","affiliation":[{"name":"Department of Anesthesiology and Reanimation, Ba\u015fak\u015fehir \u00c7am and Sakura City Hospital, Istanbul 34480, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8893-3977","authenticated-orcid":false,"given":"Zafer","family":"\u00c7ukurova","sequence":"additional","affiliation":[{"name":"Department of Reanimation, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul 34147, Turkey"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,2,1]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"102","DOI":"10.2215\/CJN.04500422","article-title":"Acid-Base Disorders in the Critically Ill Patient","volume":"18","author":"Achanti","year":"2023","journal-title":"Clin. J. Am. Soc. Nephrol."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"113","DOI":"10.1186\/cc6768","article-title":"Unmeasured Anions: The Unknown Unknowns","volume":"12","author":"Venkatesh","year":"2008","journal-title":"Crit. Care"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"220","DOI":"10.1186\/cc4954","article-title":"Unmeasured Anions in Metabolic Acidosis: Unravelling the Mystery","volume":"10","author":"Forni","year":"2006","journal-title":"Crit. Care"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"1434","DOI":"10.1056\/NEJMra1003327","article-title":"Physiological Approach to Assessment of Acid\u2013Base Disturbances","volume":"371","author":"Berend","year":"2014","journal-title":"N. Engl. J. Med."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"1109","DOI":"10.1046\/j.1365-2044.2002.02782_2.x","article-title":"Estimating Unmeasured Anions in Critically Ill Patients: Anion-Gap, Base-Deficit, and Strong-Ion-Gap","volume":"57","author":"Story","year":"2002","journal-title":"Anaesthesia"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"247","DOI":"10.1007\/s12630-008-9037-y","article-title":"Diagnosing Metabolic Acidosis in the Critically Ill: Bridging the Anion Gap, Stewart, and Base Excess Methods","volume":"56","author":"Fidkowski","year":"2009","journal-title":"Can. J. Anesth.\/J. Can. D\u2019anesth\u00e9sie"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"162","DOI":"10.2215\/CJN.03020906","article-title":"Serum Anion Gap: Its Uses and Limitations in Clinical Medicine","volume":"2","author":"Kraut","year":"2007","journal-title":"Clin. J. Am. Soc. Nephrol."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"526","DOI":"10.1136\/adc.87.6.526","article-title":"Correction of the Anion Gap for Albumin in Order to Detect Occult Tissue Anions in Shock","volume":"87","author":"Hatherill","year":"2002","journal-title":"Arch. Dis. Child."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Chawla, L.S., Shih, S., Davison, D., Junker, C., and Seneff, M.G. (2008). Anion Gap, Anion Gap Corrected for Albumin, Base Deficit and Unmeasured Anions in Critically Ill Patients: Implications on the Assessment of Metabolic Acidosis and the Diagnosis of Hyperlactatemia. BMC Emerg. Med., 8.","DOI":"10.1186\/1471-227X-8-18"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"205","DOI":"10.5935\/0103-507X.20130036","article-title":"Anion Gap Corrected for Albumin, Phosphate and Lactate Is a Good Predictor of Strong Ion Gap in Critically Ill Patients: A Nested Cohort Study","volume":"25","author":"Zampieri","year":"2013","journal-title":"Rev. Bras. Ter. Intensiv."},{"key":"ref_11","unstructured":"Kellum, J.A., and Elbers, P.W. (2009). Stewart\u2019s Textbook of Acid-Base, Lulu Press, Inc."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"43","DOI":"10.1186\/s40560-016-0166-z","article-title":"A Comparison of Prognostic Significance of Strong Ion Gap (Sig) with Other Acid-Base Markers in the Critically Ill: A Cohort Study","volume":"4","author":"Ho","year":"2016","journal-title":"J. Intensive Care"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"51","DOI":"10.1016\/0883-9441(95)90016-0","article-title":"Strong Ion Gap: A Methodology for Exploring Unexplained Anions","volume":"10","author":"Kellum","year":"1995","journal-title":"J. Crit. Care"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"219","DOI":"10.1186\/cc2189","article-title":"Closing the Gap on Unmeasured Anions","volume":"7","author":"Kellum","year":"2003","journal-title":"Crit. Care"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"860","DOI":"10.1097\/ALN.0000000000004712","article-title":"Acid\u2013Base Analysis in the Operating Room: A Bedside Stewart Approach","volume":"139","author":"Story","year":"2023","journal-title":"Anesthesiology"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"511","DOI":"10.1213\/ANE.0000000000001261","article-title":"Stewart Acid-Base: A Simplified Bedside Approach","volume":"123","author":"Story","year":"2016","journal-title":"Anesth. Analg."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"349","DOI":"10.1007\/s10877-011-9324-y","article-title":"Partitioning Standard Base Excess: A New Approach","volume":"25","author":"Morgan","year":"2011","journal-title":"J. Clin. Monit. Comput."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"R464","DOI":"10.1186\/cc3475","article-title":"Validation of a Method to Partition the Base Deficit in Meningococcal Sepsis: A Retrospective Study","volume":"9","author":"Tibby","year":"2005","journal-title":"Crit. Care"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"2282708","DOI":"10.1080\/0886022X.2023.2282708","article-title":"Association between Albumin Corrected Anion Gap and All-Cause Mortality in Critically Ill Patients with Acute Kidney Injury: A Retrospective Study Based on Mimic-Iv Database","volume":"45","author":"Gao","year":"2023","journal-title":"Ren. Fail."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"864","DOI":"10.1007\/s00134-002-1318-2","article-title":"The Strong Ion Gap Does Not Have Prognostic Value in Critically Ill Patients in a Mixed Medical\/Surgical Adult Icu","volume":"28","author":"Cusack","year":"2002","journal-title":"Intensive Care Med."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"2131","DOI":"10.1097\/01.CCM.0000079819.27515.8E","article-title":"Unmeasured Anions in Critically Ill Patients: Can They Predict Mortality?","volume":"31","author":"Rocktaeschel","year":"2003","journal-title":"Crit. Care Med."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"678","DOI":"10.1097\/EJA.0b013e32836412db","article-title":"The Base Excess Gap Is Not a Valid Tool for the Quantification of Unmeasured Ions in Cardiac Surgical Patients: A Retrospective Observational Study","volume":"30","author":"Agrafiotis","year":"2013","journal-title":"Eur. J. Anaesthesiol."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Gucyetmez, B., Gucyetmez, K., Sarikaya, Z.T., and Telci, L. (2024). Independent Variables of Ph: Ten Knights of the Hydrogen Ion Kingdom-Part I. A Prospective Observational Study. PLoS ONE, 19.","DOI":"10.1371\/journal.pone.0306756"},{"key":"ref_24","first-page":"7","article-title":"Blood Gas Analysis: Clinical Applications, Interpretation and Future Directions","volume":"6","author":"Osredkar","year":"2025","journal-title":"Med. Int."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"278","DOI":"10.1111\/joim.13483","article-title":"Predictive Models for Clinical Decision Making: Deep Dives in Practical Machine Learning","volume":"292","author":"Eloranta","year":"2022","journal-title":"J. Intern. Med."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"60","DOI":"10.4103\/2045-9912.326002","article-title":"Prediction of Diagnosis and Prognosis of Covid-19 Disease by Blood Gas Parameters Using Decision Trees Machine Learning Model: A Retrospective Observational Study","volume":"12","author":"Huyut","year":"2022","journal-title":"Med. Gas Res."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"277","DOI":"10.1007\/s11845-024-03767-6","article-title":"Interpretation of Acid\u2013Base Metabolism on Arterial Blood Gas Samples via Machine Learning Algorithms","volume":"194","author":"Ozdemir","year":"2025","journal-title":"Ir. J. Med. Sci. (1971-)"},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Phetrittikun, R., Suvirat, K., Horsiritham, K., Ingviya, T., and Chaichulee, S. (2023). Prediction of Acid-Base and Potassium Imbalances in Intensive Care Patients Using Machine Learning Techniques. Diagnostics, 13.","DOI":"10.3390\/diagnostics13061171"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Helleberg, J., Sundelin, A., M\u00e5rtensson, J., Rooyackers, O., and Thobaben, R. (2025). Beyond Labels: Determining the True Type of Blood Gas Samples in Icu Patients through Supervised Machine Learning. BMC Med. Inform. Decis. Mak., 25.","DOI":"10.1186\/s12911-025-03115-3"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"120691","DOI":"10.1016\/j.cca.2025.120691","article-title":"Artificial Intelligence for Arterial Blood Gas Interpretation","volume":"579","author":"Mousavinejad","year":"2025","journal-title":"Clin. Chim. Acta"},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"e082505","DOI":"10.1136\/bmj-2024-082505","article-title":"Probast+Ai: An Updated Quality, Risk of Bias, and Applicability Assessment Tool for Prediction Models Using Regression or Artificial Intelligence Methods","volume":"388","author":"Moons","year":"2025","journal-title":"BMJ"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"25","DOI":"10.4274\/tybd.galenos.2022.83007","article-title":"Stewart\u2019s Approach for Acid-Base Disorders: Does the Strong Ion Difference and Effects Have an Impact on Intensive Care Unit Mortality?","volume":"21","author":"Tontu","year":"2023","journal-title":"Turk. J. Intensive Care"},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Gucyetmez, B., Tuzuner, F., Atalan, H.K., Sezerman, U., Gucyetmez, K., and Telci, L. (2021). Base-Excess Chloride; the Best Approach to Evaluate the Effect of Chloride on the Acid-Base Status: A Retrospective Study. PLoS ONE, 16.","DOI":"10.1371\/journal.pone.0250274"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"1120","DOI":"10.1097\/01.CCM.0000125517.28517.74","article-title":"Initial Ph, Base Deficit, Lactate, Anion Gap, Strong Ion Difference, and Strong Ion Gap Predict Outcome from Major Vascular Injury*","volume":"32","author":"Kaplan","year":"2004","journal-title":"Crit. Care Med."},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Dyba\u0142a, M., Bartkowska-\u015aniatkowska, A., Pietrzkiewicz, K., Wiernik, A., Rosada-Kurasi\u0144ska, J., Piontek, T., Oleksiak, A., and Czyrski, A. (2025). Machine Learning Techniques for the Analysis of the Influence of Blood Gasometry Parameters on Acid\u2013Base Homeostasis in Pediatric Patients. Diagnostics, 15.","DOI":"10.3390\/diagnostics15243166"},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"371","DOI":"10.3892\/etm.2023.12070","article-title":"Serum Lactate and the Mortality of Critically Ill Patients in the Emergency Department: A Retrospective Study","volume":"26","author":"Ma","year":"2023","journal-title":"Exp. Ther. Med."},{"key":"ref_37","unstructured":"(2016, January 13\u201317). Xgboost: A Scalable Tree Boosting System. Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"817","DOI":"10.1007\/s10877-021-00713-0","article-title":"Comparison of a Modified Story Approach to Traditional Evaluation of Acid\u2013Base Disturbances in Patients with Shock: A Cohort Study","volume":"36","author":"Blaya","year":"2022","journal-title":"J. Clin. Monit. Comput."}],"container-title":["Diagnostics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/3\/427\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,2]],"date-time":"2026-02-02T14:27:58Z","timestamp":1770042478000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/3\/427"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,1]]},"references-count":38,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2026,2]]}},"alternative-id":["diagnostics16030427"],"URL":"https:\/\/doi.org\/10.3390\/diagnostics16030427","relation":{},"ISSN":["2075-4418"],"issn-type":[{"value":"2075-4418","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,1]]}}}