Understanding the sodium pump and its relevance to disease

Na,K-ATPase (sodium pump; EC 3.6.1.37) is present in the membrane of most eukaryotic cells and controls directly or indirectly many essential cellular functions. Regulation of this enzyme (ion transporter) and its individual isoforms is believed to play a key role in the etiology of some pathological processes. The sodium pump is the only known receptor for the cardiac glycosides. However, endogenous ligands structurally similar to digoxin or ouabain may control the activity of this important molecular complex. Here we review the structure and function of Na,K-ATPase, its expression and distribution in tissues, and its interaction with known ligands such as the cardiac glycosides and other suspected endogenous regulators. Also reviewed are various disorders, including cardiovascular, neurological, renal, and metabolic diseases, purported to involve dysfunction of Na,K-ATPase activity. The escalation in knowledge at the molecular level concerning sodium pump function foreshadows application of this knowledge in the clinical laboratory to identify individuals at risk for Na,K-ATPase-associated diseases.

Mechanism and elimination of aspirin-induced interference in Emit II d.a.u. assays

The presence of salicylates in urine reduces the signal in Emit assays (Syva), potentially yielding false-negative drugs-of-abuse screening results. We demonstrate that the principal urinary metabolite of salicylate, salicyluric acid (SUA; 2-hydroxybenzoylaminoacetic acid), interferes with the measurement of NADH formed in the assay by reducing the molar absorptivity of NADH at 340 nm. Thus, for a given concentration of d.a.u. analyte the change in absorbance over the assay time interval is less in the presence of SUA. With the Emit cocaine assay on the Hitachi 704 analyzer, the rate of absorbance change (delta AR) monitored at 340 nm for a specimen containing approximately 270 micrograms/L benzoylecgonine (BE) was 57 +/- 1.9 mA/min without SUA and 29 +/- 2.7 mA/min with 5 g/L SUA (n = 20). In contrast, delta AR determined at 376 nm was 18.6 +/- 0.5 mA/min with and 17.9 +/- 0.8 mA/min without 5 g/L SUA (n = 20). Measuring the Emit assay signal at wavelengths where SUA has no absorbance (376 nm) eliminates the interference due to SUA while maintaining the precision of the assay near the cutoff concentration for BE (300 micrograms/L).

Mechanism and elimination of aspirin-induced interference in Emit II d.a.u. assays

The presence of salicylates in urine reduces the signal in Emit assays (Syva), potentially yielding false-negative drugs-of-abuse screening results. We demonstrate that the principal urinary metabolite of salicylate, salicyluric acid (SUA; 2-hydroxybenzoylaminoacetic acid), interferes with the measurement of NADH formed in the assay by reducing the molar absorptivity of NADH at 340 nm. Thus, for a given concentration of d.a.u. analyte the change in absorbance over the assay time interval is less in the presence of SUA. With the Emit cocaine assay on the Hitachi 704 analyzer, the rate of absorbance change (delta AR) monitored at 340 nm for a specimen containing approximately 270 micrograms/L benzoylecgonine (BE) was 57 +/- 1.9 mA/min without SUA and 29 +/- 2.7 mA/min with 5 g/L SUA (n = 20). In contrast, delta AR determined at 376 nm was 18.6 +/- 0.5 mA/min with and 17.9 +/- 0.8 mA/min without 5 g/L SUA (n = 20). Measuring the Emit assay signal at wavelengths where SUA has no absorbance (376 nm) eliminates the interference due to SUA while maintaining the precision of the assay near the cutoff concentration for BE (300 micrograms/L).

Cardiac troponin-T immunoassay for diagnosis of acute myocardial infarction

We evaluated the analytical and clinical performance of an immunoassay for cardiac troponin T (cTnT). Within-run and total imprecision ranged from 1.6 to 11.3%. The sensitivity and linear range was 0.015 and 13 micrograms/L, respectively. Frozen samples were stable for at least 8 weeks. No interferences were seen with lipids or bilirubin (total and conjugated). Hemoglobin caused a negative bias at concentrations > 4 g/L. Heparinized plasma showed a 6% negative bias compared with serum. The clinical utility of cTnT was compared with that of creatine kinase (CK)-MB (mass assay). The sensitivity of cTnT measurements from 63 patients with acute myocardial infarction (AMI) (cTnT cutoff 0.1 microgram/L) was 60% at 0-3 h, 59% at 3-6 h, 94% at 6-9 h, 90% at 9-12 h, 99% at 12-24 h, 92% at 24-48 h, 83% at 48-72 h, and 100% at 72-96 h. Corresponding results for CK-MB (cutoff 5.0 micrograms/L and 2.5% relative index) were 45%, 64%, 82%, 97%, 87%, 81%, 54%, and 59%. The specificity of the markers from 49 non-AMI patients was 46% and 79% for cTnT and CK-MB, respectively. We show that CK-MB is more specific for diagnosis of AMI, and propose that cTnT is more sensitive to myocardial injury.

Cardiac troponin-T immunoassay for diagnosis of acute myocardial infarction

We evaluated the analytical and clinical performance of an immunoassay for cardiac troponin T (cTnT). Within-run and total imprecision ranged from 1.6 to 11.3%. The sensitivity and linear range was 0.015 and 13 micrograms/L, respectively. Frozen samples were stable for at least 8 weeks. No interferences were seen with lipids or bilirubin (total and conjugated). Hemoglobin caused a negative bias at concentrations > 4 g/L. Heparinized plasma showed a 6% negative bias compared with serum. The clinical utility of cTnT was compared with that of creatine kinase (CK)-MB (mass assay). The sensitivity of cTnT measurements from 63 patients with acute myocardial infarction (AMI) (cTnT cutoff 0.1 microgram/L) was 60% at 0-3 h, 59% at 3-6 h, 94% at 6-9 h, 90% at 9-12 h, 99% at 12-24 h, 92% at 24-48 h, 83% at 48-72 h, and 100% at 72-96 h. Corresponding results for CK-MB (cutoff 5.0 micrograms/L and 2.5% relative index) were 45%, 64%, 82%, 97%, 87%, 81%, 54%, and 59%. The specificity of the markers from 49 non-AMI patients was 46% and 79% for cTnT and CK-MB, respectively. We show that CK-MB is more specific for diagnosis of AMI, and propose that cTnT is more sensitive to myocardial injury.

Decreased signal in Emit assays of drugs of abuse in urine after ingestion of aspirin: potential for false-negative results

During routine drug analysis with the Syva d.a.u. Emit immunoassays we observed a high frequency of urines with lower rates of changes in absorbance (delta A R) than the rate for a drug-free urine calibrator. Many of these urines contained salicylates. Among 40 urines with apparent salicylate concentrations between 15 and 420 mg/dL tested for benzoylecgonine (BE), 20 had delta A R < -4 (range +2 to -28 mA/min). The rates decreased with increasing salicylate: delta A R = -0.057 x (salicylate, mg/dL) -0.22 mA/min (r = 0.85, n = 40, P < 0.01). Urines from 100 control subjects (no salicylate) had mean +/- SD delta A R values of -1.05 +/- 2.2 mA/min (range +3 to -7; only two were < -4 mA/min). Although direct addition of salicylic acid (200 mg/dL) to urine specimens did not reproduce the negative bias, ingestion of aspirin (acetylsalicylic acid) did by -0.09 mA/min per 1 mg/dL (72.4 mumol/L) salicylate. Negative biases observed for other Emit d.a.u. assays after salicylate ingestion lead us to conclude that ingestion of therapeutic doses of aspirin may cause false-negative results for drug screens in urines by this technology.

Decreased signal in Emit assays of drugs of abuse in urine after ingestion of aspirin: potential for false-negative results

During routine drug analysis with the Syva d.a.u. Emit immunoassays we observed a high frequency of urines with lower rates of changes in absorbance (delta A R) than the rate for a drug-free urine calibrator. Many of these urines contained salicylates. Among 40 urines with apparent salicylate concentrations between 15 and 420 mg/dL tested for benzoylecgonine (BE), 20 had delta A R &lt; -4 (range +2 to -28 mA/min). The rates decreased with increasing salicylate: delta A R = -0.057 x (salicylate, mg/dL) -0.22 mA/min (r = 0.85, n = 40, P &lt; 0.01). Urines from 100 control subjects (no salicylate) had mean +/- SD delta A R values of -1.05 +/- 2.2 mA/min (range +3 to -7; only two were &lt; -4 mA/min). Although direct addition of salicylic acid (200 mg/dL) to urine specimens did not reproduce the negative bias, ingestion of aspirin (acetylsalicylic acid) did by -0.09 mA/min per 1 mg/dL (72.4 mumol/L) salicylate. Negative biases observed for other Emit d.a.u. assays after salicylate ingestion lead us to conclude that ingestion of therapeutic doses of aspirin may cause false-negative results for drug screens in urines by this technology.

Demonstrating instrument-reagent flexibility: a carbamazepine enzyme immunoassay reagent system

Versatility of immunoassay reagents is beneficial to laboratories seeking cost-effective combinations of tests and automated instrumentation. In such cases, both immunoassay analytical performance and instrument independence must be assessed. Considering this, we determined the compatibility of a new carbamazepine EMIT 2000 reagent system with two fully automated but different kinetic rate analyzers (Hitachi 704 and Cobas MIRA), comparing results to a reagent-dedicated fluorescent polarization automated device (TDx) as reference. In order to more stringently assess reagent antibody specificity, we tested recovery of purified carbamazepine spiked into sera pooled from different hospital groups (normal, renal failure, hepatic failure, term pregnancy, cord blood). Cross-reactivity was additionally tested using patient sera containing various amounts of tricyclic antidepressants, compounds structurally but not functionally related to carbamazepine. Despite distinct operational differences between analyzers, precision (< 5.5% CV) and accuracy (> 95% recovery) compared well to the TDx method. However, when carbamazepine was spiked into sera from patients with hepatic failure or at term pregnancy, all three methods measured a negative bias in recovery of 16-20%. No significant cross-reactivity was observed at normal therapeutic concentration of certain tricyclic compounds, though measurable cross-reactivity was detected when present at toxic serum concentrations. We conclude that the EMIT carbamazepine immunoassay is adaptable to the different kinetic rate analyzers studied. Analytical specificity should, furthermore, be assessed in the context of interferences likely to be clinically encountered.

A Multicenter Evaluation of Lipid Profiling with a Compact Analyzer (Miles Clinistat)

We evaluated the Clinistat Analyzer (Miles Inc., Diagnostics Division, Elkhart, IN) for measuring cholesterol, triglycerides, and high-density lipoprotein (HDL) cholesterol at three medical centers. The system, based on multilayer film technology, uses precalibrated, dry film reagent disks. Ten microliters of serum is applied to the dry film reagent disk in the test procedure. For HDL-cholesterol measurement, serum is pretreated by precipitation with phosphotungstic acid and magnesium chloride. Total precision (CVs) of each of the three assays was less than or equal to 5%. The assay ranges were linear and satisfactory for clinical use. Patients' results compared well with established methods. No significant interferences were found with hemolysis, icterus, and lipemia.

A multicenter evaluation of lipid profiling with a compact analyzer (Miles Clinistat)

We evaluated the Clinistat Analyzer (Miles Inc., Diagnostics Division, Elkhart, IN) for measuring cholesterol, triglycerides, and high-density lipoprotein (HDL) cholesterol at three medical centers. The system, based on multilayer film technology, uses precalibrated, dry film reagent disks. Ten microliters of serum is applied to the dry film reagent disk in the test procedure. For HDL-cholesterol measurement, serum is pretreated by precipitation with phosphotungstic acid and magnesium chloride. Total precision (CVs) of each of the three assays was less than or equal to 5%. The assay ranges were linear and satisfactory for clinical use. Patients' results compared well with established methods. No significant interferences were found with hemolysis, icterus, and lipemia.

Isolation of digoxin-like immunoreactive factors from mammalian adrenal cortex

Endogenous digoxin-like immunoreactive factors (DLIF) are present in serum and tissues of humans and animals. To date, a tissue source for these factors has not been rigorously defined nor have these factors been isolated to identifiable homogeneity. In this study, we define the distribution of DLIF in mammalian tissues, demonstrate the adrenal cortex to be the principal source of this factor in bovine, and isolate DLIF to chromatographic homogeneity using high performance liquid chromatography (HPLC). DLIF concentrations in tissue extracts from rats measured as follows: adrenal glands, 44.3; serum, 6.3; liver, 5.2; kidney, 1.2; heart, brain, or lungs, less than 1.4 ng of digoxin-equivalent per g of protein. Human tissues showed similar results. In dogs, the ratio of the DLIF concentration in lumbar vein serum to that in infrarenal inferior vena cava serum was 3.3 +/- 0.4 (mean +/- S.E., n = 4). Bovine adrenal cortex contained 7 times more DLIF per g of tissue than the adrenal medulla. 70 +/- 4% (n = 7) of the total bovine cortical DLIF activity (6,159 pg of digoxin-equivalent) applied to a reverse phase HPLC column eluted as one definitive fraction. 60% of the digoxin-like immunoreactivity extracted from bovine serum also co-eluted with DLIF from adrenal. None of the 14 steroid molecules or 7 cardiac glycoside congeners co-eluted with the major DLIF activity. Our data indicate that 947 pmol of DLIF is equivalent to 1 pmol of digoxin-equivalent immunoreactivity. Preliminary mass spectral analysis suggests that purified DLIF has a molecular mass of 780 daltons comprised of one 390-dalton aglycone component plus several sugar moieties. This study establishes a definitive link between DLIF in serum and the adrenal cortex as a source tissue. We also demonstrate a method for purifying DLIF to chromatographic homogeneity with an extraction capacity of 1.2 nmol of DLIF per g of adrenal cortex.

Radial partition fluorescent immunoassay of thyrotropin. Analytic evaluation and clinical correlation

The authors evaluated the analytic and clinical performance of a sensitive radial partition fluorescent enzyme immunoassay for thyrotropin (TSH) performed on Stratus and compared it with a nonsensitive radioimmunoassay (RIA) method. Sensitivity of 0.15 mIU/L was obtained, and precision, specificity, and linearity were acceptable. A good correlation was observed between the two assays in samples from 311 hospitalized patients (r = 0.976). Stratus TSH results were outside the reference range for 20% of clinically euthyroid patients (n = 126), and 2.4% had undetectable levels. The clinically hyperthyroid group (n = 11) with the exception of one patient had TSH values below 0.2 mIU/L. Only 39% of hypothyroid patients on thyroid hormone replacement (n = 74) had TSH values in the reference range, with 38% and 23% exhibiting low and high values, respectively. All untreated primary hypothyroid patients (n = 8) had elevated TSH concentrations. The authors conclude that this sensitive TSH assay is useful for diagnosing hyperthyroidism when there is a clinical suspicion but cannot be recommended for thyroid screening in hospitalized patients.

A physician's office-based digoxin assay (Seralyzer) evaluated for interference by endogenous digoxin-like immunoreactive factors

A digoxin test for a physician's office based-chemistry analyzer (Ames Seralyzer) was evaluated for possible interference by digoxin-like immunoreactive factors (DLIF). Sera from patients likely to have high concentrations of DLIF (renal and hepatic patients, pregnant women, and neonates) as well as from normal patients and umbilical cord blood were analysed by the Seralyzer digoxin immunoassay and by a fluorescence polarization digoxin immunoassay (Abbott TDx) known to detect DLIF. For all patients who were not taking digoxin (n = 85) only four patients (4.7 percent) measured apparent digoxin values greater than 0.2 ng per mL by the Seralyzer compared to 64 (75 percent) by the TDx analyzer. Measurements of DLIF from adrenal extracts demonstrated a 17-fold greater potency for detection of DLIF by the TDx (2.9 ng per mL) compared to the Seralyzer technique (0.18 ng per mL). However, recovery data suggest that the presence of digoxin reduces the potency of DLIF interference as a function of increasing digoxin concentrations especially for the TDx assay. This diminished DLIF crossreactivity in the presence of digoxin is one explanation for the comparable correlation observed for both non-renal and renal failure patients taking digoxin when measured by these two immunoassays.

[Retinal detachment following posterior capsulotomy using a YAG laser]

This work analyzes 9 cases of retinal detachment occurred after posterior capsulotomy with Yag laser. No relationship has been found between the timing of the Yag capsulotomy after cataract extraction, and seriousness or precocity of the retinal detachment. The size of the capsulotomy does not seem to play a main part either. The study shows the part of the posterior capsule and the importance of associated risks (myopia, peripheral degenerative lesions, previous retinal detachment in the fellow eye). The role of the Yag laser treatment seems doubtful.

Criteria for identifying endogenous compounds as digoxin-like immunoreactive factors in humans

Endogenous digoxin-like immunoreactive factors (DLIF) are factors in plasma that interact with anti-digoxin antibodies. In this report we propose specific empirical criteria that must be satisfied by any group of endogenous compounds purported to account for DLIF activity in human plasma. These criteria include immunoreactive potency relative to existing physiologic concentrations as well as the biochemical and protein binding properties of these compounds. Recent studies have identified several congeners of fatty acids and phospholipids, hydrocortisone, and dehydroepiandrosterone-sulfate as compounds likely to account for DLIF activity in plasma. Using the above criteria we demonstrate that the highest reported plasma concentrations of these compounds combined account for less than 25% of DLIF reported in healthy adult subjects, less than 11% in newborns, less than 27% in pregnant women, and less than 39% in patients with renal failure. Human serum albumin at a concentration of 40 g/l completely abolished any detectable interaction of these compounds with both anti-digoxin antibodies or canine kidney Na/K-ATPase. The immunoreactive and physical properties of these compounds are also not consistent with those reported for DLIF. We conclude that these compounds do not account for the plasma DLIF concentrations measured in human subjects nor are they likely to play a role as specific endogenous regulators of Na/K-ATPase.

Endogenous digoxin-like immunoreactivity in elderly patients with normal serum creatinine concentrations

The effect of digoxin-like immunoreactive substance (DLIS) on serum digoxin determinations in elderly patients with normal serum creatinine concentrations was studied. Patients in the study group were over 70 years of age; the control population was under 55 years of age. All patients had serum creatinine concentrations of less than or equal to 1.5 mg/dL. Apparent digoxin content of serum samples from patients receiving digoxin and from patients in each age group who were not receiving digoxin was determined in duplicate by each of two radioimmunoassays: RIANEN (New England Nuclear), which detects DLIS, and Immophase (Corning Medical and Scientific), which is far less cross-reactive with DLIS. Results from the patients under 55 years of age were analyzed for inherent bias between the assay methods. In patients over 70 years of age, concentrations of DLIS (differences between serum digoxin concentrations measured by RIANEN and Immophase) were compared with creatinine clearance values. No DLIS was detected in sera of patients who had not received digoxin. For patients who received digoxin, nearly all apparent digoxin concentrations were within the usual therapeutic range. For patients without liver or renal failure who received digoxin, no significant difference in digoxin concentrations was observed between the RIANEN and Immophase assays in either the young (n = 17) or elderly (n = 26) patients. There was no significant increase in the difference between the RIANEN and Immophase results with decreasing creatinine clearance. In the elderly patients with normal serum creatinine concentrations, there was no evidence that measurement of serum digoxin concentration using the RIANEN and Immophase assays was compromised by DLIS.

Excretion of endogenous digoxin-like immunoreactive factors in human urine is a function of urine flow rate

We studied the effect of varying water and salt intake on the renal excretion of endogenous digoxin-like immunoreactive factors (DLIF). DLIF were measured in human urine and serum by competitive displacement of 125I-labeled digoxin from anti-digoxin antibodies. Diuresis was selectively induced in normal healthy humans by acute water ingestion, and natriuresis was preferentially induced by acute saline ingestion. We found the amount of endogenous immunoreactivity excreted in urine to be correlated with urine flow rate but not with urinary sodium excretion. Urinary excretion of DLIF, normalized to creatinine, was 3.6-fold greater at a urine flow rate of 5.5 mL/min than at 0.5 mL/min. On the other hand, saline intake increased urine flow rate 1.9-fold and increased sodium excretion threefold, but did not affect urinary excretion of DLIF. Fractional excretion of DLIF was linearly related to fractional excretion of water. This study demonstrates that normalization of DLIF values to urinary creatinine does not make DLIF excretion independent of urine flow rate and underscores the need for information on urine flow rate when DLIF measurements in urine are being interpreted.

Excretion of endogenous digoxin-like immunoreactive factors in human urine is a function of urine flow rate

We studied the effect of varying water and salt intake on the renal excretion of endogenous digoxin-like immunoreactive factors (DLIF). DLIF were measured in human urine and serum by competitive displacement of 125I-labeled digoxin from anti-digoxin antibodies. Diuresis was selectively induced in normal healthy humans by acute water ingestion, and natriuresis was preferentially induced by acute saline ingestion. We found the amount of endogenous immunoreactivity excreted in urine to be correlated with urine flow rate but not with urinary sodium excretion. Urinary excretion of DLIF, normalized to creatinine, was 3.6-fold greater at a urine flow rate of 5.5 mL/min than at 0.5 mL/min. On the other hand, saline intake increased urine flow rate 1.9-fold and increased sodium excretion threefold, but did not affect urinary excretion of DLIF. Fractional excretion of DLIF was linearly related to fractional excretion of water. This study demonstrates that normalization of DLIF values to urinary creatinine does not make DLIF excretion independent of urine flow rate and underscores the need for information on urine flow rate when DLIF measurements in urine are being interpreted.

Endogenous digoxin-like immunoreactivity in blood is increased during prolonged strenuous exercise

Digoxin-like immunoreactive factors (DLIFs) in serum may represent endogenous cardiotropic agents. We determined if blood levels of these endogenous factors changed during prolonged strenuous exercise. Total and loosely protein-bound (LPB) DLIF were measured by radioimmunoassay in the serum of nine healthy subjects during prolonged exercise to exhaustion. Mean total and LPB serum levels of DLIF increased by 72% (580 to 945 pg/mL) and 63% (53 to 91 pg/mL) over baseline values in digoxin equivalents (p less than 0.01), respectively, after three hours of exercise at 70% of VO2max. The prevalent serum nonesterified fatty acids (arachidonic, linoleic, oleic, palmitic, and stearic acids) as well as hydrocortisone did not account for the observed elevations in DLIF. Percent left ventricular fractional shortening (%FS) and mean velocity of left ventricular circumferential fiber shortening (mVCF) measured echocardiographically were lower (-18.0% and -16.4%, respectively, p less than 0.05) after exercise as compared to prior to exercise. Cardiac left ventricular dysfunction as measured by %FS did correlate with blood levels of DLIF (r = -0.680, p less than 0.02). These observations may suggest a relationship between serum levels of DLIF and cardiac fatigue.

Improved interassay correlation of digoxin results in patients with and without renal failure by elimination of digoxin-like immunoreactive factors

Use of immunoassays that do not detect endogenous digoxin-like immunoreactive factors (DLIF) in serum significantly improves the between-assay correlation of digoxin results for patients. We investigated five different immunoassay methods (Abbott, Clinical Assays, Corning, Du Pont, and Syva), measuring digoxin by all five assays in sera from 38 patients in renal failure and in 40 patients with normal renal function, all taking digoxin. The mean standard error of the estimate (Sy X x) of digoxin results (compared for all five assays) were significantly lower for patients with normal renal function than for patients in renal failure (0.148 vs 0.293 microgram/L, P less than 0.001). Assays previously shown (Clin Chem 1987;33:401) to be the least sensitive to DLIF (Syva and Corning) gave the lowest mean scatter about the regression (Sy X x = 0.192 microgram/L, renal failure; 0.114 microgram/L, normal renal function) for all 10 assay correlations. Evidently, discrepancies between digoxin values as measured by different immunoassay kits for patients with renal disease can be attributed to DLIF. Moreover, because inaccurate digoxin results attributed to DLIF may not be limited exclusively to groups of patients with known increased concentrations of DLIF, the possibility of "latent" DLIF interference may be a problem in many other human subjects.

Effect of tetrasodium EDTA on enzymatic determinations of urinary oxalate

We studied the effects of pretreating urine samples with tetrasodium EDTA (TEDTA) before measuring urinary oxalate with an enzymatic kit (Sigma). Mean analytical recovery of added oxalic acid was only 49% (SD +/- 13%) when the assay was performed as recommended by the manufacturer, but treating samples with TEDTA improved recoveries (96 +/- 10%). In 20 unselected 24-h urine samples assayed with and without TEDTA treatment, the mean oxalate concentrations were significantly (P less than 0.001) different: 15.6 +/- 8.7 and 12.2 +/- 7.9 mg/L, respectively. TEDTA-treated urine samples stored for 14 days at -20 degrees C lost 20% of their oxalate concentration. Use of TEDTA simplifies sample preparation by eliminating the alkalinizing step needed to dissolve EDTA or disodium EDTA.

Effect of tetrasodium EDTA on enzymatic determinations of urinary oxalate

We studied the effects of pretreating urine samples with tetrasodium EDTA (TEDTA) before measuring urinary oxalate with an enzymatic kit (Sigma). Mean analytical recovery of added oxalic acid was only 49% (SD +/- 13%) when the assay was performed as recommended by the manufacturer, but treating samples with TEDTA improved recoveries (96 +/- 10%). In 20 unselected 24-h urine samples assayed with and without TEDTA treatment, the mean oxalate concentrations were significantly (P less than 0.001) different: 15.6 +/- 8.7 and 12.2 +/- 7.9 mg/L, respectively. TEDTA-treated urine samples stored for 14 days at -20 degrees C lost 20% of their oxalate concentration. Use of TEDTA simplifies sample preparation by eliminating the alkalinizing step needed to dissolve EDTA or disodium EDTA.

Improved interassay correlation of digoxin results in patients with and without renal failure by elimination of digoxin-like immunoreactive factors

Use of immunoassays that do not detect endogenous digoxin-like immunoreactive factors (DLIF) in serum significantly improves the between-assay correlation of digoxin results for patients. We investigated five different immunoassay methods (Abbott, Clinical Assays, Corning, Du Pont, and Syva), measuring digoxin by all five assays in sera from 38 patients in renal failure and in 40 patients with normal renal function, all taking digoxin. The mean standard error of the estimate (Sy X x) of digoxin results (compared for all five assays) were significantly lower for patients with normal renal function than for patients in renal failure (0.148 vs 0.293 microgram/L, P less than 0.001). Assays previously shown (Clin Chem 1987;33:401) to be the least sensitive to DLIF (Syva and Corning) gave the lowest mean scatter about the regression (Sy X x = 0.192 microgram/L, renal failure; 0.114 microgram/L, normal renal function) for all 10 assay correlations. Evidently, discrepancies between digoxin values as measured by different immunoassay kits for patients with renal disease can be attributed to DLIF. Moreover, because inaccurate digoxin results attributed to DLIF may not be limited exclusively to groups of patients with known increased concentrations of DLIF, the possibility of "latent" DLIF interference may be a problem in many other human subjects.