Significance of the endogenous digoxin-like substance in infants and mothers

Role of endogenous digitalis-like factors in the clinical manifestations of severe preeclampsia: a sytematic review

Endogenous digitalis-like factor(s), originally proposed as a vasoconstrictor natriuretic hormone, was discovered in fetal and neonatal blood accidentally because it cross-reacts with antidigoxin antibodies (ADAs). Early studies using immunoassays with ADA identified the digoxin-like immuno-reactive factor(s) (EDLF) in maternal blood as well, and suggested it originated in the feto-placental unit. Mammalian digoxin-like factors have recently been identified as at least two classes of steroid compounds, plant derived ouabain (O), and several toad derived bufodienolides, most prominent being marinobufagenin (MBG). A synthetic pathway for MBG has been identified in mammalian placental tissue. Elevated maternal and fetal EDLF, O and MBG have been demonstrated in preeclampsia (PE), and inhibition of red cell membrane sodium, potassium ATPase (Na, K ATPase (NKA)) by EDLF is reversed by ADA fragments (ADA-FAB). Accordingly, maternal administration of a commercial ADA-antibody fragment (FAB) was tested in several anecdotal cases of PE, and two, small randomized, prospective, double-blind clinical trials. In the first randomized trial, ADA-FAB was administered post-partum, in the second antepartum. In the post-partum trial, ADA-FAB reduced use of antihypertensive drugs. In the second trial, there was no effect of ADA-FAB on blood pressure, but the fall in maternal creatinine clearance (CrCl) was prevented. In a secondary analysis using the pre-treatment maternal level of circulating Na, K ATPase (NKA) inhibitory activity (NKAI), ADA-FAB reduced the incidence of pulmonary edema and, unexpectedly, that of severe neonatal intraventricular hemorrhage (IVH). The fall in CrCl in patients given placebo was proportional to the circulating level of NKAI. The implications of these findings on the pathophysiology of the clinical manifestations PE are discussed, and a new model of the respective roles of placenta derived anti-angiogenic (AAG) factors (AAGFs) and EDLF is proposed.

Digoxin concentration in saliva and plasma in infants, children, and adolescents with heart disease

BACKGROUND

Because of its narrow therapeutic index, therapeutic monitoring of digoxin is important in the management of infants and children receiving the drug for cardiac failure or arrhythmias, or following accidental ingestion. Whether saliva can replace plasma in the therapeutic monitoring of digoxin therapy in children is unclear.

OBJECTIVE

This study assessed the value of determining saliva digoxin concentration in infants, children, and adolescents.

METHODS

Infants, children, and adolescents receiving digoxin for various indications, whose digoxin dosage had remained unchanged for ≥10 days, and whose compliance was good according to the parents were enrolled. Digoxin concentration was measured in paired specimens of citric acid-stimulated mixed saliva and plasma obtained simultaneously.

RESULTS

Eighteen children (10 boys, 8 girls; mean [SD] age, 42.3 [53.1] months [range, 2 months-14 years]) were included in the study. Digoxin therapy was administered for cardiac failure due to dilated cardiomyopathy in 9 patients (50.0%), ventricular septal defect in 4 (22.2%), supraventricular tachycardia in 3 (16.7%), and after cardiac surgery in 2 (11.1%). Digoxin concentration in the 20 paired specimens obtained varied from 0.0 to 0.92 ng/mL (mean [SD], 0.25 [0.26] ng/mL) in saliva and from 0.27 to 1.54 ng/mL (mean [SD], 0.77 [0.40] ng/mL) in plasma. The mean plasma/saliva digoxin concentration ratio was 2.8.

CONCLUSIONS

This study of infants, children, and adolescents receiving digoxin for a variety of indications and whose dose was unchanged for ≥10 days showed that marked individual variability in the saliva/plasma concentration ratio precludes the use of saliva in predicting the plasma digoxin concentration. The value of saliva digoxin (as opposed to plasma digoxin) measurements in the assessment of the cardiac effects of the drug in children remains to be determined.

Using pharmacokinetics to predict the effects of pregnancy and maternal-infant transfer of drugs during lactation

Knowledge of pharmacokinetics and the use of a mechanistic-based approach can improve our ability to predict the effects of pregnancy for medications when data are limited. Despite the many physiological changes that occur during pregnancy that could theoretically affect absorption, bioavailability does not appear to be altered. Decreased albumin and alpha(1)-acid glycoprotein concentrations during pregnancy will result in decreased protein binding for highly bound drugs. For drugs metabolised by the liver, this can result in misinterpretation of total plasma concentrations of low extraction ratio drugs and overdosing of high extraction ratio drugs administered by non-oral routes. Renal clearance and the activity of the CYP isozymes, CYP3A4, 2D6 and 2C9, and uridine 5'-diphosphate glucuronosyltransferase are increased during pregnancy. In contrast, CYP1A2 and 2C19 activity is decreased. The dose of a drug an infant receives during breastfeeding is dependent on the amount excreted into the breast milk, the daily volume of milk ingested and the average plasma concentration of the mother. The lipophilicity, protein binding and ionisation properties of a drug will determine how much is excreted into the breast milk. The milk to plasma concentration ratio has large inter- and intrasubject variability and is often not known. In contrast, protein binding is usually known. An extensive literature review was done to identify case reports including infant concentrations from breast-fed infants exposed to maternal drugs. For drugs that were at least 85% protein bound, measurable concentrations of drug in the infant did not occur if there was no placental exposure immediately prior to or during delivery. Knowledge of the protein binding properties of a drug can provide a quick and easy tool to estimate exposure of an infant to medication from breastfeeding.

Pregnancy-induced changes in pharmacokinetics: a mechanistic-based approach

Observational studies have documented that women take a variety of medications during pregnancy. It is well known that pregnancy can induce changes in the plasma concentrations of some drugs. The use of mechanistic-based approaches to drug interactions has significantly increased our ability to predict clinically significant drug interactions and improve clinical care. This same method can also be used to improve our understanding regarding the effect of pregnancy on pharmacokinetics of drugs. Limited studies suggest bioavailability of drugs is not altered during pregnancy. Increased plasma volume and protein binding changes can alter the apparent volume of distribution (Vd) of drugs. Through changes in Vd and clearance, pregnancy can cause increases or decreases in the terminal elimination half-life of drugs. Depending on whether a drug is excreted unchanged by the kidneys or which metabolic isoenzyme is involved in the metabolism of a drug can determine whether or not a change in dosage is needed during pregnancy. The renal excretion of unchanged drugs is increased during pregnancy. The metabolism of drugs catalysed by select cytochrome P450 (CYP) isoenzymes (i.e. CYP3A4, CYP2D6 and CYP2C9) and uridine diphosphate glucuronosyltransferase (UGT) isoenzymes (i.e. UGT1A4 and UGT2B7) are increased during pregnancy. Dosages of drugs predominantly metabolised by these isoenzymes or excreted by the kidneys unchanged may need to be increased during pregnancy in order to avoid loss of efficacy. In contrast, CYP1A2 and CYP2C19 activity is decreased during pregnancy, suggesting that dosage reductions may be needed to minimise potential toxicity of their substrates. There are limitations to the available data. This analysis is based primarily on observational studies, many including small numbers of women. For some isoenzymes, the effect of pregnancy on only one drug has been evaluated. The full-time course of pharmacokinetic changes during pregnancy is often not studied. The effect of pregnancy on transport proteins is unknown. Drugs eliminated by non-CYP or non-UGT pathways or multiple pathways will need to be evaluated individually. In conclusion, by evaluating the pharmacokinetic data of a variety of drugs during pregnancy and using a mechanistic-based approach, we can start to predict the effect of pregnancy for a large number of clinically used drugs. However, because of the limitations, more clinical, evidence-based studies are needed to fully elucidate the effects of pregnancy on the pharmacokinetics of drugs.

Comparison of microparticle enzyme and fluorescence polarization immunoassays in pediatric patients not receiving digoxin

Previously, investigators have measured endogenous digoxin-like immunoreactive substances (DLIS) in pediatric patients. Digoxin-like immunoreactive substances may cross-react with digoxin assays to produce false-positive digoxin concentrations; hence, the validity of digoxin concentrations in pediatric patients is questionable. The authors compared the presence and magnitude of apparent DLIS using the microparticle enzyme immunoassay (MEIA) AxSYM Digoxin II and the fluorescence polarization immunoassay (FPIA) TDx Digoxin II, in the serum of 80 pediatric patients who were hospitalized with normal serum creatinine but had not been administered digoxin. Patients ranged in age from 1 day to 16 years (mean age, 4.96 +/- 5.17 years). Serum creatinine and total bilirubin were 0.5 +/- 0.18 mg/dl and 1.3 +/- 0.17 mg/dl, respectively. Forty-eight percent of MEIA samples and 79% of FPIA samples had measurable DLIS values. Digoxin-like immunoreactive substance concentrations for the MEIA (0.07 +/- 0.09 ng/ml) and FPIA assays (0.1 +/- 0.1 ng/ml) were statistically different (p = 0.01); however, no sample had a DLIS value >0.38 ng/ml. A poor correlation was noted between patient age, serum creatinine, total bilirubin, and DLIS concentration. The MEIA and FPIA assays effectively minimized DLIS cross-reactivity making both technologies clinically acceptable for serum digoxin measurement in pediatric patients with normal serum creatinine and total bilirubin.

Therapeutic drug monitoring in special populations

Therapeutic drug monitoring (TDM) is commonly used to maintain “therapeutic” drug concentrations. Even in compliant patients, with “average” drug kinetics, TDM is useful to identify the causes of unwanted or unexpected responses, prevent unnecessary diagnostic testing, improve clinical outcomes, and even save lives. TDM has greatest promise in certain special populations who are: (a) prone to under- or overrespond to usual dosing regimens, (b) least able to tolerate, recognize, or communicate drug effects, or who are (c) intentionally or accidentally misdosed. TDM is especially useful in patients at the extremes of age, in adolescents, and in patients who are either taking multiple drugs or expressing unusual pharmacokinetics as a result of physiological, environmental, or genetic causes. Less-well-appreciated uses of TDM include prevention of dangerousunderdosing of patients, investigation of adverse drug reactions, and identification of serious medication errors, even for a number of drugs that are not traditionally monitored. TDM can be useful for some drugs in any patient and for most drugs in some special populations.

The rate of rise of corticotrophin releasing factor and endogenous digoxin-like immunoreactivity in normal and abnormal pregnancy

Maternal plasma concentrations of corticotrophin releasing factor (CRF) and endogenous digoxin-like immunoreactivity (EDLI) were estimated in 80 normal and 88 abnormal pregnancies which were sampled sequentially from 24 weeks gestation to delivery. A slope was fitted for each woman's antenatal EDLI and CRF values, both of which rose significantly during gestation, and the mean of the slopes for the normal and abnormal groups for each value compared. There was no evidence of significant mean differences between groups for EDLI but there was evidence of a significant mean difference for CRF (P less than 0.05). After adjustment for other variables which may affect pregnancy outcome, the slopes for CRF were found not to be significantly related to outcome.

Effect of endogenous digoxin-like substances on the interpretation of high concentrations of digoxin in children

The objective of this study was to assess the effect of endogenous digoxin-like substances on the interpretation of excessive concentrations of digoxin in children. After the development of a high-pressure liquid chromatography (HPLC) method for digoxin in our laboratories, we analyzed sera of children in whom the fluorescence polarization immunoassay identified potentially toxic concentrations of the glycoside (greater than 3 nmol/L; 2.3 ng/ml). Sixteen of them were receiving long-term digoxin therapy, and one had an accidental overdose. The immunoassay yielded significantly higher concentrations (4.1 +/- 1.2 nmol/L; 3.2 +/- 0.9 ng/ml) than the HPLC method (3.3 +/- 1.6 nmol/L; 2.6 +/- 1.2 ng/ml; p less than 0.01). In five cases (30%) these differences were clinically significant because administration of digoxin had been discontinued in the presence of true digoxin concentrations within the therapeutic range and the lack of clinical toxic effects. These data suggest that therapeutic drug monitoring using immunoassays of digoxin may be too inaccurate to detect potential toxic effects, and that much more weight should be focused on clinical monitoring. The HPLC method for assay of digoxin is extremely meticulous and will not become clinically available; therefore the development of better immunoassays should be encouraged.

[Preeclampsia: physiopathology and prospects for early detection]

Preeclampsia is a complication of pregnancy characterized by hypertension, edema and proteinuria, beginning after 20 weeks of gestation. Six percent of the pregnant women in North America develop this disease, which is associated with increased morbidity and mortality for the mother and her baby. The physiopathology remains uncertain despite many research efforts. Actual hypotheses seek to explain the vasospasm that characterizes the disease. Among the many factors influencing vascular reactivity and possibly implicated are: the renin-angiotensin system, prostaglandins, progesterone and its metabolites, calcium, magnesium, digoxin-like immunoreactive substance(s), auricular natriuretic factor, substances secreted by platelets and leukotrienes. Prevention of the disease is limited by the absence of a biological or clinical marker with good sensitivity and appropriate specificity. Many biochemical or hematological parameters have been reported: uric acid, calcium, magnesium, proteinuria, blood iron, hematocrit, platelet count, antithrombin III, estrogen and progesterone. The combination of several tests could be superior to the use of each test individually, providing a better sensitivity and improving the positive predictive value. With early detection, new therapies for the prevention of the disease could be experimented on the higher risk women before the apparition of clinical symptoms or signs. Furthermore, those tests could be used in the study of the pathophysiology and in the choice of the best therapy.

Digoxin clearance and serum beta-2-microglobulin in neonates

To find reliable indicators of digoxin clearance (CLdig) in the neonatal period, we investigated the linear correlation of CLdig and the reciprocal of CLdig (1/CLdig) with serum beta 2-microglobulin (S beta 2-MG), serum creatinine, blood urea nitrogen, age, and weight on 25 occasions in 21 neonates with congenital heart disease. The S beta 2-MG value showed a significantly closer correlation to 1/CLdig (r = 0.84, p less than 0.0001) than to the other values. The regression equation was (1/CLdig) = 0.15 X (s beta 2-MG) + 0.08. Creatinine and blood urea nitrogen values correlated less closely with 1/CLdig (r = 0.67 and 0.71, respectively). Age and weight had no significant linear correlation with CLdig and 1/CLdig. Determination of s beta 2-MG values allowed an estimate of CLdig by means of the regression equation between s beta 2-MG and 1/CLdig, and permitted a prediction of the required maintenance dose of digoxin. We conclude that s beta 2-MG is a good indicator of CLdig in neonates, and that the determination of s beta 2-MG values may facilitate the advance individualization of digoxin therapy in the neonatal period.

Digoxin-like immunoreactive substances in chronic liver disease

Digoxin-like immunoreactive substances, which cross-react with digoxin antibody, have been found to have natriuretic effect and Na+,K+-ATPase inhibitory effect. The role of digoxin-like immunoreactive substances in chronic liver disease was studied by radioimmunoassay in 63 serum and 60 urine samples from 58 patients with chronic liver disease and compared with 16 controls. Although the mean serum digoxin-like immunoreactive substances level of compensated chronic liver disease patients (0.06 +/- 0.05 ng per ml, p less than 0.01) was higher than that of controls (0.02 +/- 0.03 ng per ml), only four patients had serum digoxin-like immunoreactive substances higher than 0.10 ng per ml. Mean serum digoxin-like immunoreactive substances level was much higher in patients with decompensated chronic liver disease who had ascites (0.32 +/- 0.17 ng per ml, p less than 0.001), hepatorenal syndrome (0.57 +/- 0.20 ng per ml, p less than 0.001) and hepatic encephalopathy (0.43 +/- 0.20 ng per ml, p less than 0.001). Five patients with recent variceal hemorrhage requiring transfusions and saline infusion had significantly increased serum digoxin-like immunoreactive substances (mean: 0.16 +/- 0.06 ng per ml, p less than 0.001) before the development of clinically detectable ascites.(ABSTRACT TRUNCATED AT 250 WORDS)

A comparative trial of digoxin and digitoxin in the treatment of congestive heart failure

A randomized, crossover, single-blind study compared the efficacy and dosing accuracy of digoxin and digitoxin in 15 ambulatory patients wth congestive heart failure. Loading doses and maintenance doses were calculated according to published equations that adjust for sex, height, and lean body weight (for digitoxin), plus estimated creatinine clearance (for digoxin). At each 2-week visit, serum drug concentrations were measured and compliance with the prescribed regimen was assessed by tablet count. At the end of each study period, a congestive heart failure (CHF) score was determined in a blinded fashion by the same physician. Patient compliance was unusually high (greater than or equal to 80%) at every visit. Therapeutic concentrations were achieved with digoxin and digitoxin in 5 and 14 patients, respectively (p less than 0.05). During digitoxin therapy, CHF scores were lower than pretreatment values (p less than 0.05). The difference between CHF scores during the digoxin and digitoxin periods did not achieve significance (0.05 less than p less than 0.06). Therapeutic serum concentrations can be achieved more easily and frequently with digitoxin than digoxin without compromising the patient's CHF status.

Digoxin-like immunoreactive substances in severe acute liver disease due to viral hepatitis and paracetamol overdose

The levels of endogenous serum digoxin-like immunoreactive substances were investigated during development of encephalopathy in patients with fulminant hepatic failure. The 67 patients studied had varying degrees of hepatic failure as a result of viral hepatitis or paracetamol overdose. Serum levels of digoxin-like immunoreactive substances were significantly increased in both viral hepatitis and paracetamol overdose, with mean values of 0.42 +/- S.D. 0.25 ng per ml (n = 36) and 0.53 +/- 0.19 ng per ml (n = 31), respectively, as compared to normal control subjects with mean values of 0.01 +/- 0.02 ng per ml (n = 21, p less than 0.001). A statistically significant correlation was found between serum digoxin-like immunoreactive substances and the degree of encephalopathy in the viral hepatitis patients and with the serum creatinine in the paracetamol overdose patients where renal failure was more severe. No correlation was found with liver damage as assessed by the prolongation of the prothrombin time, serum AST or bilirubin values. Experiments with ultrafiltration and heating showed that both free nonprotein-bound digoxin-like immunoreactive substances and the total digoxin-like immunoreactive substances measured were increased. Column chromatography of ultrafiltrates of fulminant hepatic failure serum on Sephadex G-25 demonstrated at least two peaks with digoxin-like immunoreactive activity. Reduced renal function is an important factor in the increased serum level of digoxin-like digoxin-like immunoreactive substances, but their presence due to liver failure, where there is increased permeability of the blood-brain barrier, could be relevant to the development of hepatic encephalopathy.

Digoxin-like immunoreactive substance in fetuses with and without cardiac pathology

The concentration of digoxin-like immunoreactive substance was measured by immunoassay in umbilical venous blood from six normal and 19 fetuses with a variety of cardiac and noncardiac disorders. Fetal blood was obtained either by percutaneous umbilical blood sampling (n = 13) or at delivery (n = 12). Three women received digoxin for fetal indications. Healthy control fetuses had significantly less digoxin-like immunoreactive substance measured (mean concentration below the limit of the assay sensitivity) than was found in ill fetuses whose mothers received digoxin (p less than 0.005). However, the fetal concentrations of immunoreactive digoxin in fetuses with a cardiac abnormality were similar whether the mother had (0.93 +/- 0.4 ng/ml) or had not (1.27 +/- 0.4 ng/ml) received digoxin (p = 0.1452). Although there was a significant negative correlation between digoxin concentration and gestational age (R = -0.5079, p less than 0.01), the youngest fetuses examined were generally the sickest. The correlation with gestational age was not significant if the normal control fetuses were excluded. One fetus with a cardiac tachyarrhythmia was examined during and after transplacental therapy. There was no change. It is possible that previously measured "digoxin" in "treated" fetuses represents digoxin-like immunoreactive substance and that only small amounts of maternally administered digoxin actually reach the ill fetus. Our findings suggest that a randomized trial of maternal digitalization for the treatment of fetal supraventricular tachycardia is essential prior to its acceptance as effective therapy.

Endogenous digitalis-like compounds in essential and experimental hypertension

The hypothesis that endogenous digitalis-like compounds might participate in body sodium and water homeostasis have led us to investigate the presence in plasma of compounds interacting with digoxin antibodies in man and rats. The apparent levels of digoxin-equivalents in plasma of control subjects (n = 21) and patients with essential hypertension (n = 48) or end-stage renal failure (n = 13) were 24.7 +/- 3.2, 34.4 +/- 4.4 and 98.7 +/- 17.4 pg/ml, p less than 0.05 and p less than 0.01 respectively. Positive correlations were observed between systolic and diastolic blood pressure and the apparent immunoreactivity of plasma. No relationship was found with the renal Na+ excretion or the plasma renin activity. The apparent digoxin-like immunoreactivity of the plasma was correlated with its ability to inhibit ouabain binding to the erythrocyte Na+ pump and to reduce the renal Na+,K+-ATPase activity. In rats with experimental hypertension, the plasma cross-reactivity with antidigoxin antibodies was also enhanced when compared to control rats (71.6 +/- 10.2 pg/ml, n = 12 and 57.3 +/- 5.0 pg/ml, n = 33 in Na+ loaded rats and in rats with reduced renal mass respectively compared to 43.4 +/- 3.7 pg/ml, n = 36, p less than 0.05). In spontaneously hypertensive rats (SHR), the apparent levels of digoxin- equivalents were higher than that of age-matched WKY normotensive rats. This increase was already present in prehypertensive SHR (3 week-old) (105.8 +/- 12.4 vs 40.0 +/- 6.5 pg/ml, n = 9 and 8, p less than 0.001) and persisted after hypertension has developed (134 +/- 12.6 vs 85 +/- 7.9 pg/ml, n = 7 and 8, p less than 0.005 in 30 week-old rats). The apparent affinity of the erythrocyte Na+,K+ cotransport for intracellular Na+ and the maximal rate of the Na+ pump were correlated with the plasma digoxin-like levels. These results confirm the presence in plasma of compounds possessing some of the functional and structural properties of cardioactive steroids, associated with a rise in blood pressure.

Effect of age and serum creatinine on endogenous digoxin-like substances in infants and children

An apparent digoxin-like immunoreactive substance(s) (DLIS) was evaluated in 374 pediatric patients, 0 to 72 months of age, not receiving digoxin. The relationship between DLIS presence or concentration and age, weight, gender, race, and serum creatinine was investigated. Twenty-seven percent of patients had a positive apparent DLIS concentration (greater than or equal to 0.2 ng/ml). The mean +/- SD concentration of DLIS in the positive group was 0.39 +/- 0.18 ng/ml (range 0.2 to 1.37 ng/ml). Patients in the DLIS-positive group were younger than those in the negative group (P less than 0.01). Although a greater percentage of infants younger than 6 months of age had measurable DLIS, the mean DLIS concentrations, when present, were not significantly different for all age groups (P greater than 0.05). No significant relationship was found between race or gender and DLIS. A weakly positive correlation between serum creatinine concentration and DLIS was noted (r = 0.22, P less than 0.03), but elevated serum creatinine measurements (greater than 0.6 mg/dL) did not correlate with DLIS concentration. These results support the hypothesis that the presence of DLIS is age related, but DLIS appears to be present in a much larger and older population of pediatric patients than previously reported.

Influence of assay methodologies and interferences on the interpretation of digoxin concentrations

The interpretation of serum digoxin concentrations is an area where pharmacists can have substantial clinical input. Important in this interpretation is the knowledge of various factors that may interfere with the reported concentration. These include drugs, metabolites of digoxin, protein concentration, and endogenous digoxin-like substances. The patients for whom these interferences are most significant include neonates, patients with renal failure and liver disease, and pregnant women. Pharmacist knowledge of these interferences can have an important impact on the quality of patient care.

Determinants of drug disposition in man

This article reviews some of the important determinants of variation in drug disposition such as age, gender, body weight, diet, environmental influences, drug - protein interactions, compliance, drug - drug interactions, endogenous substances, disease states, circadian variation and genetics.

Endogenous digitalis-like factors

Recent research has demonstrated the presence of endogenous compounds in blood and urine that crossreact with antibodies raised against digoxin. Given the widespread therapeutic use of digoxin and its being monitored clinically by immunoassay, such digoxin-like immunoreactive compounds pose significant diagnostic and interpretive problems. Serum levels of this factor(s) approaching therapeutic digoxin levels have been found in digoxin-free patients in renal failure, pregnant women, and newborns. The compound is incompletely characterized; however, existing data suggest that it is a small, neutral, nonpeptidic compound. In serum it is highly protein bound, and alterations in this binding appear to give rise to the false-positive assay results. The urinary form is probably conjugated. Digoxin-like immunoreactive substances may play a role in volume homeostasis and appear associated with essential and pregnancy-induced hypertension. If such roles are primary, measurement of digoxin-like immunoreactive substances may prove to be of value in and of itself.

Interpretation of excessive serum concentrations of digoxin in children

Between January 1981 and April 1984, excessive serum concentrations of digoxin (5 ng/ml or higher) were recorded in 47 children, aged 2 days to 16 years. In 10 patients, the high concentrations were measured 9.25 to 48 hours after death and were significantly higher than antemortem levels in all cases (8.3 +/- 2.4 (+/- standard deviation) postmortem vs 3.3 +/- 1.5 antemortem, less than 0.0001). In 15 patients (40.5% of the living patients) serum concentrations of 5 ng/ml or higher reflected sampling errors; drug levels were monitored too closely to the administration of a dose. None of these children had toxic manifestations of digoxin. In 10 patients, the excessive concentrations were associated with renal failure and a prolonged elimination half-life (T1/2) of digoxin; in 3 of these patients, there were signs of digoxin toxicity. Six cases were caused by digoxin overdose (accidental ingestions, pharmacy error and a suicide attempt). In 6 additional cases, the existence of an endogenous digoxin-like substance (EDLS) was shown to contribute to the excessive levels of the drug. One case could be attributed to digoxin-amiodarone interaction. In 10 of 37 living patients, digoxin toxicity was diagnosed. After excluding the 15 sampling errors and 6 cases with EDLS, this represents 63% of the cases. There was a good correlation between digoxin elimination T1/2 and serum creatine concentrations (r = 0.71, p less than 0.01). The above observations suggest that excessive serum concentrations of digoxin may not necessarily reflect potentially toxic levels.(ABSTRACT TRUNCATED AT 250 WORDS)