Negative Anion Gaps
Supplementary file 1: Breakdown of the data sources in the study which are included in Supplemental files 2-9. Supplementary file 2: Data from 133 serum/plasma specimens from 116 unique patients (52 female, 64 male) that had a negative anion gap (-1 or lower). The retrospective timeframe is May 1, 2009 through August 18, 2021. Specific data fields include: unique patient identification number (deidentified), order description which generated the anion gap calculation, patient age in years at time of specimen collection, sex of patient as in the electronic medical record, [Na+] in mEq/L, [K+] in mEq/L, [Cl-] in mEq/L, [HCO3-] in mEq/L, calculated anion gap, identification of a recognized cause of low anion gap in the patient, comments on processes likely underlying the negative anion gap, classification of the electrolyte abnormalities that would result in a negative anion gap (low [Na+], high [HCO3-], and/or high [Cl-]), presence or absence of hypoalbuminemia (based on age-related reference range), and corrected anion gap (if hypoalbuminemia was present). Supplementary file 3: Data for 886 whole blood specimens from 387 unique patients (166 female, 221 male) that had a negative anion gap (-1 or lower). The retrospective timeframe is May 1, 2009 through August 2, 2014. Calculations are as described in Supplementary file 2. Supplementary file 4: Data for 776,504 specimens from 167,763 unique patients who had the Basic Metabolic Panel (sodium, potassium, chloride, bicarbonate, creatinine, blood urea nitrogen, glucose, total calcium) ordered which yielded an anion gap or 11 or lower. The retrospective timeframe is January 11, 2010 through August 18, 2021. Supplementary file 5: Data for 804,639 specimens from 184,466 unique patients who had the Basic Metabolic Panel ordered which yielded an anion gap or 12 or higher. The retrospective timeframe is January 11, 2010 through August 18, 2021. Supplementary file 6: Data for 809,704 specimens from 183,976 unique patients who had discrete (individual orders) for sodium, chloride, and bicarbonate on the same specimen. The retrospective timeframe is May 1, 2009 through August 18, 2021. Supplementary file 7: Data for 451,016 specimens from 121,862 unique patients who had orders for chemistry panel. The retrospective timeframe is January 12, 2010 through August 18, 2021. Supplementary file 8: Data for 93,987 specimens from 15,715 unique patients who had blood gas analysis performed on whole blood specimens where concentrations for sodium, chloride, and bicarbonate were all determined. The retrospective timeframe is May 1, 2009 through August 2, 2014. Supplementary file 9: Data for 361,199 specimens from 127,230 unique patients who had basic metabolic panels performed on either Piccolo Express (1,575 specimens) or in the central laboratory (185,277 specimens) or comprehensive metabolic panels performed on either Piccolo Express (11,139 specimens) or in the central laboratory (163,198 specimens).
Steps to reproduce
The retrospective analysis period was May 1, 2009 through August 18 , 2021. Supplemental file 1 lists the start date for when the other panels (Electrolyte Panel, Chemistry 6 Panel, Chemistry 7 Panel, and Comprehensive Metabolic Panel) were introduced in the electronic order entry system. Prior to January 11, 2010, all electrolyte orders on serum/plasma were only available as discrete, individually ordered tests. A reporting tool within the electronic medical record, known as Epic Reporting Workbench, was used to identify and then download all of the clinical chemistry tests in our retrospective study as we have described in previous report. The blood gas analyzer data was retrieved from the laboratory information system used in our clinical laboratories until October 2014 (Cerner Classic, Kansas City, MO, USA). Unlike with Epic Reporting Workbench reports for the serum/plasma data described above, we were unable to obtain patient location at the time of testing for the blood gas analyzer data. All data in the present study was obtained from patient data in the electronic medical record or laboratory information system from the University of Iowa Hospitals and Clinics (Iowa City, Iowa, United States); no data was obtained from diagnostic vendors or electronic medical records of other hospital systems. Calculated anion gap used the equation Anion Gap = [Na+] – [HCO3-] – [Cl-], with the electrolyte concentrations in mEq/L. Correction anion gap used the following equation: Corrected Anion Gap = Anion gap + (2.5 x ([Normal Albumin] – [Measured Albumin])), with [Albumin] measured in concentrations of g/dL. Detailed chart review was performed for all patients who had a negative anion gap on a serum/plasma specimen. For assessment whether a spurious event such as laboratory error or specimen contamination was likely to have caused the negative anion gap, the specimen must have been associated with at least 2 of the following: (a) no other explanation likely given clinical history, (b) result inconsistent with baseline and/or follow-up laboratory studies, (c) clinical documentation attributing the result to “laboratory error” or similar language, or (d) cluster of similar suspicious results on same day. At the start of the retrospective analysis timeframe, electrolyte analysis (including sodium, chloride, and bicarbonate) on serum/plasma specimens was performed on Roche Diagnostics (Indianapolis, IN, USA) Modular P analyzers, with front-end automation using a Modular Pre-Analytic (MPA)-7 unit. The modular P analyzers running electrolytes were replaced in early 2013 with cobas 8000 system c702 analyzers, which have been used up until the end of the retrospective analysis timeframe. Blood gas analysis was performed on Radiometer (Copenhagen, Denmark) ABL835 analyzers in the retrospective timeframe of this report. The point-of-care chemistry analysis was performed on the Abaxis Piccolo Xpress.