Evaluating the Association Between Fibrinogen and Rotational Thromboelastometry and the Progression to Severe Obstetric Hemorrhage.

AIM: This study aims to investigate the ability of fibrinogen and rotational thromboelastometry (ROTEM) parameters measured at obstetric hemorrhage protocol initiation to predict severe hemorrhage. METHODS: In this retrospective study we included patients whose hemorrhage was managed with an obstetric massive transfusion protocol. Fibrinogen and ROTEM parameters EXTEM clotting time (CT), clot formation time (CFT), alpha angle, A10, A20, lysis index 30 min after CT (LI30), FIBTEM A10, A20, were measured at initiation of the protocol with transfusion based on a predefined algorithm. Patients were grouped into either severe or nonsevere hemorrhage based on: peripartum fall in hemoglobin ≥4 g/dL, transfusion of ≥4 units of blood product, invasive procedures for hemorrhage control, intensive care unit admission, or death. RESULTS: Of the 155 patients included, 108 (70%) progressed to severe hemorrhage. Fibrinogen, EXTEM alpha angle, A10, A20, FIBTEM A10, A20 were significantly lower in the severe hemorrhage group while the CFT was significantly prolonged in the severe hemorrhage group. In univariate analysis, predicted progression to severe hemorrhage yielded areas under the receiver operating characteristic curve (95% confidence interval [CI]) of: fibrinogen: 0.683 (0.591-0.776), CFT: 0.671 (0.553, 0.789), EXTEM alpha angle: 0.690 (0.577-0.803), A10: 0.693 (0.570-0.815), A20: 0.678 (0.563-0.793), FIBTEM A10: 0.726 (0.605-0.847), and A20: 0.709 (0.594-0.824). In a multivariable model, fibrinogen was independently associated with severe hemorrhage (odds ratio [95% CI] = 1.037 [1.009-1.066]) for every 50 mg/dL decrease in fibrinogen drawn at obstetric hemorrhage massive transfusion protocol initiation. CONCLUSION: Both fibrinogen and ROTEM parameters measured at the initiation of an obstetric hemorrhage protocol are useful parameters for predicting severe hemorrhage.

Comparison of Creatinine, Urea, Ionized Calcium, and Bicarbonate Methods by 2 POC Systems and a Mainframe Chemistry Analyzer.

BACKGROUND: The i-STAT (iST) and GEM Premier ChemSTAT (ChmST) are point-of-care systems that measure electrolytes, metabolites, acid-base parameters, and hematocrit on blood. We compared results for creatinine, urea (blood urea nitrogen [BUN]), total carbon dioxide (tCo2), and ionized calcium (ion Ca) on blood by the ChmST and iST to Beckman DxC600 (DxC600) results for creatinine and urea on plasma and GEM Premier 5000 (GEM5K) results for ion Ca and tCo2 on blood. METHODS: 107 heparinized blood specimens were analyzed on the ChmST and iST systems, with plasma tested by DxC600 for creatinine and BUN and blood tested by GEM5K for ion Ca and tCo2. We evaluated the methods primarily by the mean and SD of the biases at clinically relevant concentrations. RESULTS: The correlations of ChmST and iST results on blood for creatinine, BUN, ion Ca, and tCo2 correlated to plasma by the DxC600 and to blood by the GEM5K were r ≥ 0.98 for all analytes except for ion Ca on the iST (r = 0.93). The mean and SD of biases were within clinically and analytically acceptable limits for all methods except for tCo2 on the ChmST, which measures bicarbonate with a bicarbonate-sensing electrode. Also, creatinine and BUN by the ChmST were less affected by icterus or hemolysis than were the DxC600 (icterus) and the iST (hemolysis). CONCLUSIONS: The ChmST and iST results on blood demonstrated strong correlations with each other and with the DxC600 results on plasma. We conclude the ChmST provides reliable results for whole blood creatinine, urea, ion Ca, and tCo2.

Retrospective Analysis of Quality Control Data Using Pooled Blood to Compare Agreement within Two Models of Blood Gas Analyzers.

BACKGROUND: Two previous reports from the same group concluded that the analytical reliability of Instrumentation Laboratory GEM4000 analyzers (GEM4K) deteriorated during a 24-hour period, based on results between samples from the same patient but collected at different times. Our routine blood gas Between-Laboratory Survey is done every 2 weeks using a freshly pooled heparinized blood sample taken to each analyzer location to verify comparability among our GEM4K and Radiometer ABL90 (Rad90) blood gas analyzers. Because another report found a few very large differences in glucose results between the GEM4K and central laboratory analyzers, we reviewed the glucose results on plasma from our Chemistry Between-Laboratory Surveys that includes comparisons between our central laboratory analyzers (Beckman DxC800; DxC800) and our GEM4K and Rad90 blood gas analyzers. METHOD: Using data from our Blood Gas and Chemistry Surveys over a 55-week period, we calculated the mean, standard deviation (SD), and concentration intervals of the 27 sets of results by the GEM4K or Rad90 analyzers. Agreement in plasma glucose between DxC800 and blood gas analyzers was evaluated by the limits of agreement and intraclass correlation coefficient analysis. RESULTS: For each analyte, the Rad90 had lower average SD than the GEM4K for the 55-week period, although both brands of analyzers showed acceptable performance. For plasma glucose results on our Chemistry Survey, the GEM4K results agreed more closely with the DxC800 results than the Rad90 results. CONCLUSIONS: Based on both our Blood Gas and Chemistry Surveys, we conclude that both brands of analyzers performed within analytically and clinically acceptable limits throughout the year, with no evidence for the type of errors reported previously.

Use of Earlier-Reported Rotational Thromboelastometry Parameters to Evaluate Clotting Status, Fibrinogen, and Platelet Activities in Postpartum Hemorrhage Compared to Surgery and Intensive Care Patients.

BACKGROUND: Rotational thromboelastometry (ROTEM) can provide clinical information in 10-20 minutes for guiding administration of fibrinogen, platelets, and fresh frozen plasma products. While ROTEM testing is well established for cardiac and other surgeries, it is less characterized for use in postpartum hemorrhage (PPH) patients. We wanted to determine if the earlier-measured ROTEM parameters (α-angle and amplitude at 10 minutes [A10]) could replace the later parameters (amplitude at 20 minutes and maximum amplitude [maximum clot firmness {MCF}]) in all patient groups studied. We also correlated the A10 and α-angle of the EXTEM and FIBTEM tests to the fibrinogen levels and platelet counts in these patients. METHODS: We retrospectively analyzed 100 sets of EXTEM and FIBTEM results ordered on patients undergoing operations for PPH, patients in intensive care units (ICU), and those undergoing cardiothoracic surgery (cardiothoracic operating room [C/T OR]). We determined if the correlations among the various parameters were similar among the PPH, ICU, and C/T OR patients. RESULTS: As expected, the EXTEM A10 (A10EX) and FIBTEM A10 (A10FIB) correlated highly to the EXTEM MCF and FIBTEM MCF in all patient groups. The A10EX parameter correlated significantly to both fibrinogen and platelet levels, and the A10FIB correlated to the fibrinogen levels. The difference between the A10EX and the A10FIB (PLTEM) is related to platelet activity, and we found that the PLTEM and platelet count correlated highly for all 100 PPH patients (r = 0.80), C/T OR patients (r = 0.70), and ICU patients (r = 0.66), despite 4 high platelet counts with relatively low PLTEM values in the ICU group. The earlier-reported parameter EXTEM α angle (α-EX) is an excellent indicator of the A10EX, with an α-EX ≥65° (ie, normal) giving a >96% probability that the A10EX was ≥44 mm, and an α-EX value below 65 mm giving an 86% probability that the A10EX was <44 mm. CONCLUSIONS: The correlations among the ROTEM parameters for the PPH comparisons were equivalent to the C/T OR patients studied, and the A10EX and A10FIB could replace the MCF results in all patient groups. Also, the α-EX was an early indicator of the A10EX and had good correlations to the A10FIB and the fibrinogen in all patient groups. Finally, in a separate group of 62 comparisons, the FIBTEM α angle showed promise as an early indicator of the A10FIB and the fibrinogen levels.

Quality Control Practices for Chemistry and Immunochemistry in a Cohort of 21 Large Academic Medical Centers.

OBJECTIVES: In the United States, minimum standards for quality control (QC) are specified in federal law under the Clinical Laboratory Improvement Amendment and its revisions. Beyond meeting this required standard, laboratories have flexibility to determine their overall QC program. METHODS: We surveyed chemistry and immunochemistry QC procedures at 21 clinical laboratories within leading academic medical centers to assess if standardized QC practices exist for chemistry and immunochemistry testing. RESULTS: We observed significant variation and unexpected similarities in practice across laboratories, including QC frequency, cutoffs, number of levels analyzed, and other features. CONCLUSIONS: This variation in practice indicates an opportunity exists to establish an evidence-based approach to QC that can be generalized across institutions.

Effect of beef ingestion by humans on plasma concentrations of creatinine, urea, and cystatin C.

PURPOSE: The effect of eating meat on serum concentrations of creatinine has varied among previous reports, with some finding no effect and others finding 50-100% increases, which appears related to how the beef is cooked. For other analytes related to kidney function, urea is well known to increase following a protein meal, and the effect of eating meat on cystatin C concentrations has been studied once. METHODS: We had 32 participants eat a measured amount of cooked beef (5-6 or 10-12 oz; 142-170 or 284-340 g) and collected blood for measurements at 1 h before and immediately before eating beef, then at 1, 2, and 4 h after eating the beef. We measured creatinine using both alkaline picrate and enzymatic methods, cystatin C using a nephelometric immunoassay, and urea using an enzymatic method. RESULTS: For creatinine, both the picrate and enzymatic methods showed similar responses, with a peak average increases of 5.9 µmol/L (0.07 mg/dL) and 4.6 µmol/L (0.05 mg/dL), respectively, at 2 h. Cystatin C had a very slightly maximal decrease of -0.037 mg/L at 2 h. Urea had the largest change, increasing by 0.30 and 0.77 mmol/L at 2 and 4 h respectively. CONCLUSIONS: Healthy individuals were found to have minor increases in serum creatinine (~5 µmol/L) following the ingestion of 5/6 or 10/12 oz of fried beef. Cystatin C appears to decrease very slightly in some people after beef ingestion, possibly due either to circadian variation or to a hormonal effect of eating. We conclude that ingesting these amounts of fried beef has a small effect on plasma creatinine concentrations. Although these increases would likely not affect the diagnosis of a kidney impairment in this population or in those with kidney disease, eating meat before collecting blood for creatinine measurement should be avoided.

Relationships and Clinical Utility of Creatinine, Cystatin C, eGFRs, GFRs, and Clearances.

BACKGROUND: This review addresses techniques for glomerular filtration rate (GFR), either measured by clearance tests such as with creatinine, iothalamate, inulin, or iohexol [measured GFR (mGFR)] or calculated by equations that determine the estimated GFR (eGFR) from serum measurements of creatinine and/or cystatin C. However, mGFR tests are slow and impractical for routine use. Therefore, calculations of eGFRs have been developed that have advantages over the mGFRs. CONTENT: The eGFR is a serum creatinine and/or cystatin C adjusted for age, sex, and race, with mathematical manipulations to produce an average numerical agreement with the mGFR. However, all comparisons between eGFR and mGFR show wide scatter that appears to be related to the large variability of the mGFR. Procedures for mGFR often do not agree with each other and have both wide population variation (similar to plasma creatinine and cystatin C) and within-individual variation that is much larger than creatinine or cystatin C. Whether the measured GFR is even equivalent to serum creatinine and/or cystatin C for detecting early clinical changes in chronic kidney disease will be addressed. SUMMARY: Procedures for measuring GFR are tedious and expensive, and have both wide population variation (similar to plasma creatinine and cystatin C), and within-individual variation that is much larger than either creatinine or cystatin C. Because the normal range for mGFR overlaps considerably with the stages 1 and 2 of chronic kidney disease, mGFR has significant clinical limitations. Instead of trying to mimic mGFRs, the focus should be on using eGFRs on their own clinical merits to detect impaired kidney function.

The analytical change in plasma creatinine that constitutes a biologic/physiologic change.

PURPOSE: Accurate and precise measurements of creatinine are necessary to evaluate changes in kidney function related to a decreased glomerular filtration rate (GFR). When serial measurements of creatinine are monitored in an individual, it is useful to know what magnitude of an analytical change in creatinine indicates a true physiologic/biologic change in plasma creatinine that might warrant clinical intervention. METHODS: We compared results between three different methods for creatinine using large chemistry analyzers, two based on alkaline picrate (AP1 and AP2), and one based on dry-slide enzymatic conversion (ENZ). On each of three different segments or days of the study spaced 1-2months apart, we selected 10 different plasma samples having creatinine concentrations ranging from about 0.5mg/dL to 4.5mg/dL (44 to 400µmol/L). Each sample was analyzed in triplicate on each of two same-model analyzers at each institution, then from this data we determined the precision of each model of analyzer. The within-instrument precision of each analyzer was evaluated from the differences between the triplicate results on each sample by each analyzer (mean and SD of the differences). The between-instrument precision was evaluated as the differences between results on the same sample (1, 2, 3, etc.) analyzed on different analyzers of the same model (A and B). This between-analyzer precision data was used to determine both the range and mean±2SD of the differences that could be used to indicate that greater changes in creatinine concentrations would represent a biologic change. RESULTS: The within-instrument precision was best for the ENZ method in comparison to the two alkaline picrate rate methods. The between-instrument precision of the 90 consecutive measurements (30 samples×triplicate analyses) between the same-model analyzers were (mean and SD of differences in mg/dL): -0.018 and 0.029 (ENZ); 0.016 and 0.11 (AP1), and -0.058 and 0.071 (AP2). CONCLUSIONS: While all three of the creatinine methods studied had good precision, the ENZ method had the best precision, such that a change of 0.07mg/dL (6µmol/L) in serial creatinine concentrations up to 1.5mg/dL on a patient could indicate a biologic change had occurred. For the alkaline picrate methods, a measured change of creatinine of 0.23mg/dL for AP1 or 0.11mg/dL for AP2 would indicate that a physiologic change in serum/plasma creatinine has occurred. While a definite biologic change may simply represent daily variations, detecting a biologic change in creatinine more rapidly could impact the ability of creatinine to detect early and clinically significant changes in renal function.

Clinical evaluation of the FreeStyle Precision Pro system.

BACKGROUND: A new version of international standard (ISO 15197) and CLSI Guideline (POCT12) with more stringent accuracy criteria are near publication. We evaluated the glucose test performance of the FreeStyle Precision Pro system, a new blood glucose monitoring system (BGMS) designed to enhance accuracy for point-of-care testing (POCT). METHODS: Precision, interference and system accuracy with 503 blood samples from capillary, venous and arterial sources were evaluated in a multicenter study. Study results were analyzed and presented in accordance with the specifications and recommendations of the final draft ISO 15197 and the new POCT12. RESULTS: The FreeStyle Precision Pro system demonstrated acceptable precision (CV <5%), no interference across a hematocrit range of 15-65%, and, except for xylose, no interference from 24 of 25 potentially interfering substances. It also met all accuracy criteria specified in the final draft ISO 15197 and POCT12, with 97.3-98.9% of the individual results of various blood sample types agreeing within ±12 mg/dl of the laboratory analyzer values at glucose concentrations <100mg/dl and within ±12.5% of the laboratory analyzer values at glucose concentrations ≥100 mg/dl. CONCLUSIONS: The FreeStyle Precision Pro system met the tighter accuracy requirements, providing a means for enhancing accuracy for point-of-care blood glucose monitoring.

Improving the clinical value of estimating glomerular filtration rate by reporting all values: a contrarian viewpoint.

The serious limitations of the estimating glomerular filtration rate (eGFR) appear related not to a shortcoming of the equation, but to the futility of trying to force agreement between two inherently different parameters: a blood marker of kidney function with a very stable concentration (creatinine) and a renal filtration parameter that fluctuates continually (glomerular filtration rate, GFR). Although GFR is regarded as the ultimate determinant of kidney function, it may be less ideal as an early clinical marker to detect declining kidney function. Another shortcoming of GFR is that it has significant overlap between health and kidney disease states categorized according to stage I, II, etc. Serum creatinine has a real and measurable increase as kidney function declines, but this is often masked when creatinine is plotted on a scale of 1.0 mg/dl (88 micromol/l), which is well above the detection limit of modern creatinine methods of about 0.05 mg/dl. A new equation to estimate GFR, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine equation, modestly improves accuracy from 80.6% of the Modification of Diet in Renal Disease eGFRs being within 30% of the measured GFR, to 84.1% of the CKD-EPI eGFRs being within 30% of the measured GFR. Creatinine methods have recently been standardized to an isotope dilution mass stectrometry reference method. While this will lessen the systematic bias between methods, it will have no effect on either the imprecision of a particular creatinine method or on the inherent random differences between serum creatinine (or eGFR) and actual GFR. Finally, the eGFR is not recommended for reporting until it is well below a reference range for those with no kidney disease. However, if the eGFR were properly regarded as an age-, gender-, and race-adjusted serum creatinine, it could be reported at all values and become a more clinically useful parameter.

Variation of serum creatinine, cystatin C, and creatinine clearance tests in persons with normal renal function.

BACKGROUND: To determine the potential sensitivity of several renal function tests for detecting early changes in renal function, we compared the within-individual (W-I) variation over 5 months of serum creatinine, serum cystatin C, and creatinine clearance. METHODS: On 31 healthy subjects, blood and timed urine specimens were collected once each month to get 6 collections. Creatinine (enzymatic) in serum and urine and cystatin C (immunonephelometric) in serum were measured and glomerular filtration rate (GFR) by creatinine clearance and the Modification of Diet in Renal Disease (MDRD) equation were calculated. To compare W-I variations between different creatinine methods, we also measured creatinine by both enzymatic and kinetic alkaline picrate methods on 15 sets of frozen samples. RESULTS: For the 31 volunteers, the mean W-I variations for serum creatinine (5.8%) and cystatin C (5.4%) were both much lower than the W-I variation of creatinine clearance (18.7%). As expected, the MDRD GFR had a similar W-I variation (6.7%) to that of serum creatinine and its values were markedly different than GFR by creatinine clearance. On the 15 sets of frozen samples, the W-I variation of creatinine measured by the enzymatic method (CV 5.2%) was slightly less than by the picrate method (CV 6.2%). CONCLUSIONS: The low W-I variation of both serum cystatin C and serum creatinine suggests that serial measurements of either would detect a changes in renal function earlier than would GFR by creatinine clearance or MDRD equation, which allows reporting only for GFRs<60 ml/min/1.7 m(2). While we measured only creatinine clearance, the large variability, difficulty, and cost of all clearance measurements make them impractical for routine monitoring of patients.

Misconceptions in reporting oxygen saturation.

BACKGROUND: We describe some misconceptions that have become common practice in reporting blood gas and cooximetry results. In 1980, oxygen saturation was incorrectly redefined in a report of a new instrument for analysis of hemoglobin (Hb) derivatives. Oxygen saturation (sO2) was redefined as the ratio of oxyhemoglobin (O2Hb) to total Hb instead of the ratio of O2Hb to active Hb (O2Hb + desoxyhemoglobin). In addition, the new terms "functional saturation" and "fractional saturation" were introduced. Since the new parameter was implemented in a widely used cooximeter, its use is now widespread and has caused misunderstandings. METHODS: In this report, we review the development of the definitions and measurements of sO2 and related quantities and contend that the misconceptions should be resolved by standardizing instrument read-outs and clinical reports, so that sO2, defined as the ratio of O2Hb to active Hb, should replace FO2Hb and be reported along with the total Hb concentration and the common dyshemoglobin fractions (%CO-Hb and % methemoglobin [metHb]). RESULTS: The redefinition of sO2 as the %O2Hb or FO2Hb did not address the confusion that might result from interchanging these two often-similar but different terms. The term fractional saturation is an inappropriate terminology and lacks clear physiological meaning. We see frequent cases of confusion: (a) the difference between the sO2 in pulse oximetry and the FO2Hb in cooximetry is called the "pulse oximeter gap;" (b) sO2 results are described as "method dependent;" and (c) reference ranges for these terms are substituted. CONCLUSIONS: Although either parameter could be used by clinicians who fully understand the relatively simple difference between these parameters, we find clear evidence that there is widespread confusion of these terms, even among experts in the field. Standardization of the reporting format would help, and instrument manufacturers could contribute by standardizing the reporting format for cooximetry results.

Validation of a quality assessment system for blood gas and electrolyte testing.

BACKGROUND: A recently-introduced quality assessment system (Intelligent Quality Management: iQM), was evaluated in routine clinical use at four different hospitals. The iQM technology is designed to replace conventional external liquid controls with software, Process Control (PC) Solutions and Calibration Validation components that continually assess the function of the GEM Premier 3000 (GEM) analyzer and automatically initiate and document corrective actions. METHODS: We validated the performance claims of iQM by monitoring quality control (QC) materials at 4 clinical sites while analyzing approximately 10,550 patient samples. We compared iQM-measured QC values to traditional QC results, evaluating the number and type of error flags for patient samples, and used data from control results to calculate the average time to detect an error (ADT) for each analyte. RESULTS: The calculated ADT was approximately 3 min for all analytes except for sodium (17 min), glucose (11 min), and lactate (5.9 min). Precision of control materials in iQM cartridges was better than from external controls run on traditional analyzers. The iQM system detected errors in 0.46% of actual clinical samples. CONCLUSIONS: The findings from our study confirm that (a) iQM precision in a clinical setting is comparable to that found in previous studies done in a research setting, (b) the improved precision of control material on the iQM is likely because the internal control fluids are sealed and not susceptible to exposure from handling, and (c) the system detects and often corrects errors in specific samples that might not be reported by traditional analytical systems.

Changing systems for measuring activated clotting times: impact on the clinical practice of heparin anticoagulation during cardiac surgery.

BACKGROUND: The activated clotting time (ACT) is a standard monitor for heparin anticoagulation during cardiopulmonary bypass (CPB). This study determines the effect of upgrading our ACT system on our clinical practice with regards to the conduct and safety of heparin anticoagulation during cardiopulmonary bypass. METHODS: We compared the intraoperative heparin doses required for all adult cardiac surgery patients (n=1240) and postoperative bleeding for a subset of primary aortocoronary bypass (CABG) surgery procedures (n=285) from cohorts before and after the change in ACT systems. RESULTS: The heparin dose needed to exceed our target ACT of 480 sec for the duration of CPB was higher (45000 vs. 40000 units; p<0.0001), and the mean ACT during CPB was lower (557 vs. 618 sec; p<0.05) using the new ACT system. Furthermore, this coincided with decreased postoperative bleeding in the CABG subset (median value of 417 vs. 575 ml over 12 h; p<0.0005). CONCLUSIONS: We demonstrated that the introduction of the Actalyke ACT system significantly altered our clinical practice by increasing the heparin dose required to exceed our target ACT during CPB. Prospective study to determine the effect of Actalyke ACT system monitoring on hemostasis after cardiac surgery is merited.

The relationship between ionized and total serum magnesium concentrations during abdominal surgery.

STUDY OBJECTIVE: To determine the necessity for ionized magnesium (iMg) assay by evaluating the effect of abdominal surgery without massive transfusion on total magnesium (Mg) and iMg concentrations. DESIGN: Prospective, observational study. SETTING: Operating rooms at a university teaching hospital. PATIENTS: 31 patients without electrolyte abnormalities who were scheduled for abdominal surgery. INTERVENTIONS: None. MEASUREMENTS: Serum electrolytes, iMg, ionized calcium, proteins before and after surgery, and urine Mg were recorded. MAIN RESULTS: Total Mg and iMg decreased: Mg 23%, iMg 15%; p < 0.0001. Total and iMg were highly correlated (r = 0.728, p < 0.0001). Urine Mg output remained within normal limits (mean 0.24 mmol/hr). CONCLUSIONS: During abdominal surgery without massive transfusion, both total and ionized hypomagnesemia occur. Changes in iMg and total serum Mg concentrations are closely correlated. Total serum Mg assay, while overestimating the prevalence of ionized hypomagnesemia, is sufficient to screen for hypomagnesemia.

pH effects on measurements of ionized calcium and ionized magnesium in blood.

CONTEXT: It is well known that the concentration of ionized calcium in blood is affected by the pH of the specimen, since hydrogen ions compete with calcium for binding sites on albumin and other proteins. However, the relationship between pH and ionized magnesium concentration is not as well characterized. OBJECTIVE: To determine the effects of pH on ionized magnesium concentration over a wide range of pH values in serum or plasma. DESIGN: Both ionized calcium and ionized magnesium concentrations were measured in 3 sets of samples. (1) Pools of serum or whole blood at different pH values (7.20-7.60) achieved by adding a constant volume of acid or base (diluted solutions of either hydrochloric acid or sodium hydroxide) plus saline. These pools consisted of 2 serum and 3 heparinized whole blood pools collected from leftover blood remaining in clinical specimens in the Clinical Chemistry and Blood Gas Laboratories, respectively, at Duke University Medical Center. (2) Five whole blood specimens obtained from apparently healthy individual donors. (3) Twenty-six whole blood specimens obtained from individual patients (leftover blood from the Blood Gas Laboratory) in which pH was varied by in vitro loss or gain of carbon dioxide. RESULTS: Both ionized calcium and ionized magnesium concentrations decreased as the pH in the specimen increased, indicating the stronger binding of these ions with proteins in the more alkaline environment. CONCLUSION: We conclude that the rate of change of ionized magnesium concentration with pH change (0.12 mmol/L per pH unit) is significantly less than that of ionized calcium (0.36 mmol/L per pH unit). Furthermore, our findings indicate that if adjustment to pH 7.40 is necessary, the ionized magnesium test results need to be adjusted when pH is markedly abnormal, as is sometimes done for ionized calcium.