Accuracy of methemoglobin measurements: comparison of six different commercial devices and one manual method.

OBJECTIVE: During nitric oxide inhalation, methemoglobinemia needs to be monitored. We compared six commercially available instruments and one manual method for methemoglobin measurements. In addition, we studied whether and to what degree methylene blue interferes with methemoglobin measurements. DESIGN: In vitro methodologic study. SETTING: Research laboratory in a university hospital. PATIENTS: Five healthy volunteers from whom red blood cells were obtained. INTERVENTIONS: Methemoglobinemia was generated in a red blood cell suspension by nitric oxide; methemoglobin was measured with six commercial instruments and one manual photometric method to calculate variation coefficients and to determine the differences between the devices. Methemoglobin was measured with and without the addition of methylene blue with two instruments. Measurements were performed immediately after the addition of methylene blue. MEASUREMENTS AND MAIN RESULTS: All six commercially available instruments had variation coefficients of <0.1 at methemoglobin concentrations of 5%, whereas the manual photometric method did not reach a variation coefficient of <0.1 at 8% of methemoglobin. Apart from two devices that measured slightly but significantly higher methemoglobin levels, all instruments measured similar values of methemoglobin when the same samples were determined simultaneously. Higher concentrations of methylene blue (10, 40, 100 microM) reduced substantially the apparent concentrations of methemoglobin. Interference by methylene blue was most pronounced at low methemoglobin levels. CONCLUSIONS: With some limitations, all commercial instruments that were tested performed adequately for the monitoring of methemoglobinemia. Methylene blue interferes with the methemoglobin measurements in a dose-dependent manner.

Early versus late surfactant treatment in preterm infants of 27 to 32 weeks' gestational age: a multicenter controlled clinical trial.

OBJECTIVE: To investigate whether early (<1 hour after birth) surfactant administration would be superior to late treatment (2-6 hours after birth) in preterm infants. STUDY DESIGN: Randomized controlled multicenter clinical trial. PATIENTS AND METHODS: Prenatal randomization of all infants of 27 to 32 weeks' gestational age stratified by center after parental informed consent. Early treatment: 100 mg/kg body weight bovine surfactant (SF-RI1, Alveofact; Dr K. Thomae, Biberach, Germany) to infants requiring intubation after birth. Late treatment: identical dosage to infants requiring intubation up to 6 hours of age with the fraction of inspired oxygen >0.4 at 2 to 6 hours after birth. Primary endpoint: the time on mechanical ventilation. Main secondary endpoints: mortality, bronchopulmonary dysplasia, intraventricular hemorrhage >/=grade III, and periventricular leukomalacia. Sample size calculation: at least 280 infants to prove superiority of either approach (alpha = 0.05; beta = 0.90). RESULTS: Enrollment of 317 infants, 154 randomized to early surfactant treatment, 163 to late surfactant treatment. Study infants (all following data intent-to-treat groups: early versus late surfactant) were similar with respect to: gestational age, 29.5 +/- 1.6 weeks versus 29.7 +/- 1.6 weeks; birth weight, 1227 +/- 367 g versus 1269 +/- 334 g; and the rate of prenatal corticosteroids, 79.9% versus 72.8%. Duration of mechanical ventilation: 3 days (0-8) versus 2 days (0-6) (median, interquartile); further outcome variables: death or bronchopulmonary dysplasia (day 28) 25.9% versus 23.9%, mortality 3.2% versus 1.8%, intraventricular hemorrhage >/=grade III 6.5% versus 3.7%, and periventricular leukomalacia 5.2% versus 5.5% not differing statistically. CONCLUSION: In preterm infants with a high rate of prenatal glucocorticoids, early surfactant administration was not found to be superior to late treatment in terms of relevant outcome variables.

Reduction of NO-induced methemoglobinemia requires extremely high doses of ascorbic acid in vitro.

The objective of the present study was to investigate the treatment of nitric oxide (NO)-induced methemoglobinemia by ascorbate and its consequences on red blood cell (RBC) glutathione in vitro. RBC were obtained from five healthy volunteers. The following experiments were carried out: (1) After methemoglobin generation by NO, ascorbate was added (2) RBC were simultaneously exposed to NO and ascorbate (3) Methemoglobin was generated by NO, ascorbate was added and incubation with NO continued. (1) After discontinuation of NO, the mean half life for methemoglobin was reduced from 195 min (controls) to 60 min (10 mM ascorbate) in a dose-dependent manner. (2) Methemoglobin formation after 3 h of NO exposure was 2.7 +/- 0.3% in controls and 1.8 +/- 0.1% with 10 mM ascorbate (p < 0.01). (3) Further methemoglobin formation was inhibited only by 10 mM ascorbate (p < 0.001). NO incubation did not affect RBC glutathione (86.5 +/- 19.6 and 86.5 +/- 19.6 mg/l, respectively). Treatment with 10 mM ascorbate significantly decreased glutathione (p < 0.002). In vitro, NO-induced methemoglobin formation is significantly decreased only by a high (10 mM) ascorbate concentration. Glutathione, critical for ascorbate activity, is not influenced by NO.

Extracorporeal circulation increases nitric oxide-induced methemoglobinemia in vivo and in vitro.

OBJECTIVES: Methemoglobinemia is a well-known side effect of nitric oxide inhalation. We tested the hypothesis whether cardiopulmonary bypass increases methemoglobin formation by nitric oxide. DESIGN: A two-phase study: a) a controlled, prospective in vivo study of inhaled nitric oxide treatment followed by b) a second, prospective and controlled in vitro study. SETTING: Pediatric intensive care unit and research laboratory in a university hospital. PATIENTS: The in vivo study consisted of 25 patients following open-heart surgery and 19 children with acute respiratory distress syndrome (ARDS) or persistent pulmonary hypertension of the newborn. The in vitro study consisted of 20 patients with and 20 patients without cardiopulmonary bypass. INTERVENTIONS: For the in vivo study, methemoglobin measurements were taken before and after application of 20 parts per million (ppm) of nitric oxide. For the in vitro study, red blood cells of patients were incubated with 32 ppm nitric oxide before and after surgery. Methemoglobin, glutathione, adenosine triphosphate (ATP), and nicotinamide adenine dinucleotide/nicotinamide adenine dinucleotide phosphate (NADH/NADHP) concentrations were compared. MEASUREMENTS AND MAIN RESULTS: For the in vivo study, nitric oxide inhalation increased methemoglobin from 0.2 +/- 0.1% to 1.2 +/- 0.7% in patients receiving nitric oxide after open-heart surgery and from 0.2 +/- 0.1% to 0.5 +/- 0.4% in ARDS/persistent pulmonary hypertension of the newborn patients (p < .01). For the in vitro study, nitric oxide incubation of red blood cells increased methemoglobin concentration from 3.7 +/- 1.9% preoperatively to 7.4 +/- 2.4% after open-heart surgery. This increase was not observed in patients who did not undergo cardiopulmonary bypass (3.6 +/- 1.6% vs. 3.6 +/- 1.9%; p < .001). In erythrocytes of patients undergoing extracorporeal circulation, there was no difference between pre- and postoperative glutathione, ATP, and NADH/NADPH concentrations. CONCLUSIONS: Cardiopulmonary bypass is an important risk factor for methemoglobinemia when inhaled nitric oxide is applied. This risk is not secondary to diminished concentrations of energetic substrates.

Neonatal morbidity and mortality associated with maternal haemolysis elevated liver enzymes and low platelets syndrome.

UNLABELLED: To compare the impact of maternal haemolysis, elevated liver enzymes and low platelets (HELLP) syndrome, uncomplicated hypertension in pregnancy (HIP), and no hypertension (controls) on neonatal morbidity and mortality, 108 infants were matched with respect to gestational age, date of birth, and gender. The HELLP group infants had more grade 3 and 4 respiratory distress syndromes (36%) than the HIP group (19%) or controls (11%). Cardiovascular instability (arterial hypotension, volume resuscitation) was significantly more common in HELLP neonates (20% and 31%) than in HIP infants (9% and 6%) or controls (3% and 9%). Both, HELLP and HIP infants showed a higher incidence of growth retardation than the controls. After 32 weeks of gestation the incidence of severe neonatal morbidity was not different. CONCLUSION: : Before 32 weeks of gestation both respiratory and cardiovascular morbidity and intra-uterine growth retardation associated with HIP is further aggravated by a maternal HELLP syndrome.

Airway nitric oxide in asthmatic children and patients with cystic fibrosis.

We evaluated the effects of asthma and cystic fibrosis on nitric oxide (NO) concentrations in the respiratory tract. NO levels in orally exhaled air and nasal gas samples were studied in 90 asthmatic patients (4-14 yrs), 67 patients with cystic fibrosis (CF) (5-32 yrs), and 68 controls (4-34 yrs). NO concentrations measured by chemiluminescene were correlated with the patient's vital capacity, forced expiratory volume in one second (FEV1) and specific airway resistance. In all groups, NO concentrations in orally-exhaled air correlated with the inhaled ambient NO (r = 0.85-0.91). At an ambient NO concentration of 0 parts per billion (ppb), asthmatic patients exhaled air with higher NO concentrations than cystic fibrosis patients and controls (8.0 +/- 6.1 ppb (n = 33); 4.9 +/- 2.6 ppb (n = 23); and 3.0 +/- 2.5 ppb (n = 37); respectively; p < 0.001). Similar results were obtained for ventilation-adjusted orally-exhaled NO. Nasal NO concentrations were lower in patients with CF (23 +/- 17 ppb) than in controls and asthmatics (96 +/- 47 and 103 +/- 64 ppb; p < 0.001). There was no relationship between nasal or oral NO and pulmonary function tests. Our results suggest that ambient NO levels influence NO concentrations in orally-exhaled air. Like adults, asthmatic children exhale more NO than their controls. Reduced nasal NO concentrations in patients with cystic fibrosis may reflect chronic epithelial cell damage or an increased mucosal barrier impeding NO diffusion into the airway.

Biosynthesis of prostacyclin and thromboxane A2 during chronic hypoxaemia in children with cyanotic congenital heart disease.

The high risk of vaso-occlusive events in children younger than 4 years with cyanotic congenital heart disease and polycythaemia has been attributed to increased thromboxane (Tx) A2 formation. In older children with cyanotic congenital heart disease, however, the risk of vaso-occlusive events is much lower. We therefore hypothesized that the formation of TxA2 and prostacyclin is not disturbed in this age group. We measured urinary excretion of stable index metabolites of in vivo TxA2 and prostacyclin formation by gas chromatography-mass spectrometry in nine children (age 5.9-14.4, median 8.7 years) with cyanotic congenital heart disease, and in nine healthy, age-matched control subjects. The patients excreted less 2,3-dinor-TxB2 (systemic TxA2 formation, P = 0.03), 2,3-dinor-6-keto-PGF1 alpha (systemic prostacyclin formation. P = 0.03) and TxB2 (renal TxA2 formation, P = 0.01) than the control subjects. We conclude that in children older than 5 years with cyanotic congenital heart disease, endogenous synthesis of TxA2 and prostacyclin is not stimulated. This result may explain the lower risk of vaso-occlusive events in this age group as compared with younger children. In addition, our results suggest that chronic hypoxaemia may affect the in vivo formation of TxA2 and prostacyclin and the metabolic disposition of TxB2.

Increased thromboxane biosynthesis in childhood hemolytic uremic syndrome.

Vascular endothelial cell damage plays a central role in the pathogenesis of the hemolytic uremic syndrome (HUS), resulting in intravascular platelet activation and thrombotic microangiopathy. A deficiency of the antiaggregatory prostacyclin (PGI2) has been postulated by experiments under ex vivo conditions. However, this observation has not been confirmed in vivo. The pathophysiological contribution of thromboxane (Tx)A2, a potent vasoconstrictor and platelet-aggregating prostanoid which is predominantly produced by platelets, has not been elucidated so far. In order to quantitate endogenous formation of TxA2 in children with HUS, plasma concentrations of the enzymatic metabolite 11-dehydro-TxB2 of TxA2 and urinary excretion rates of three major TxA2 metabolites, TxB2, 11-dehydro-TxB2 and 2,3-dinor-TxB2 were analyzed using gas chromatography/mass spectrometry. PGI2 biosynthesis was assessed by measuring urinary excretion of an index metabolite of its systemic production, 2,3-dinor-6-keto-prostaglandin (PG) F1 alpha, and an index of its renal production, 6-keto-PGF1 alpha. TxA2 biosynthesis was markedly elevated in the acute phase of HUS. This activation could be detected for a longer period of time than the presence of thrombocytopenia. Concomitantly in the acute phase, renal PGI2 formation was significantly elevated and systemic PGI2 formation was elevated in 50% of the patients. These data indicate that TxA2 formation is increased in the acute phase in patients with HUS. This enhanced biosynthesis is consistent with increased platelet activation, whereas the increased PGI2 biosynthesis reflects predominantly renal endothelial cell damage.

Increased renal biosynthesis of prostaglandin E2 and thromboxane B2 in human congenital obstructive uropathy.

Animal experiments have shown that after ureter obstruction hydronephrotic kidneys release increased amounts of prostaglandin E2 (PGE2) and thromboxane A2 (TxA2), suggesting that these prostanoids modify renal blood flow and excretory function in this model. To test the hypothesis that these mechanisms are also operative in congenital obstructive uropathy, we measured prostanoid excretion rates in 12 neonates and infants with congenital unilateral or bilateral hydronephrosis. Prostanoid determinations were performed by gas chromatography mass spectrometry. PGE2 and thromboxane B2 (TxB2) (non-enzymatic metabolite of TxA2) excretion exceeded the normal range in eight and 11 of 12 patients, respectively. Median PGE2 excretion was 22, range 4-572 ng/h/1.73 m2 (normal 3-16). Median TxB2 excretion was 22, range 3-188 ng/h/1.73 m2 (normal 3-7). No other renal prostanoids (prostaglandin F2 alpha, 6-keto-prostaglandin F1 alpha) or systemic prostanoid metabolites (PGE-M, 2,3-dinorthromboxane B2, 11-dehydro-thromboxane B2, 2,3-dinor-6-keto-prostaglandin F1 alpha) were consistently elevated. A second group of 12 neonates with congenital obstructive uropathy was followed sequentially. PGE2 and thromboxane B2 excretion rates increased even further after surgical decompression and gradually normalized during follow-up. There was a significant relationship between elevated FeNa and enhanced PGE2 and TxB2 excretion. These data suggest that endogenous renal formation of PGE2 and TxA2 is selectively stimulated in hydronephrotic kidneys in neonates and infants. PGE2 and TxA2 may be involved in modulating renal function in these infants.

Unusual metabolism of prostacyclin in infants with persistent septic pulmonary hypertension.

In urine of healthy man, the major metabolite of prostacyclin is 2,3-dinor-6-oxo-prostaglandin F1 alpha. The excretion rates of this compound as well as of 2,3-dinor-thromboxane B2, a major metabolite of thromboxane A2, in two newborns with septic persistent pulmonary hypertension were about 30- to 50-fold higher than the normal range (2,3-dinor-6-oxo-prostaglandin F1 alpha: 3-15 ng/h/1.73 m2; 2,3-dinor-thromboxane B2: 8-25 ng/h/1.73 m2). The ratios of 2,3-dinor-6-oxo-13,14-dihydro-prostaglandin F1 alpha/2,3-dinor-6-oxo-prostaglandin F1 alpha in these two infants were about 100% and 800%, respectively whereas in controls the excretion of the 13,14-dihydro metabolite was found to be about 10-25% of 2,3-dinor-6-oxo-prostaglandin F1 alpha. Thus in patients with septic persistent pulmonary hypertension and extremely high excretion rates of prostacyclin and thromboxane A2 metabolites, the pattern of metabolites differs from those of healthy man.

Pharmacologically active levels of PGE1 in neonates with congenital heart disease.

In general, prostanoids act as local mediators, not as circulating hormones. A specific exception to this rule is the infusion of prostaglandin E1 in patients with ductus arteriosus-dependent pulmonary or systemic blood flow associated with congenital heart disease. We therefore measured prostaglandin E1 plasma levels by gas chromatography-mass spectrometry during effective infusion of prostaglandin E1 in 10 neonates. Prostaglandin E1 plasma levels ranged from 22 to 530 (median 56) pg/ml in these patients. Since prostaglandin E1 is not synthesized endogenously to any significant extent, these plasma concentrations constitute genuine circulating levels not confounded by the common problem of e vivo artifacts. If endogenous prostanoids (e.g. prostaglandin E2) are suspected as circulating mediators, plasma levels detected by reliable methods ought to be in the same range as prostaglandin E1 plasma levels in the present investigation.

Endogenous formation of prostanoids in neonates with persistent pulmonary hypertension.

Endogenous formation of thromboxane A2 and prostacyclin were evaluated in seven neonatates with persistent pulmonary hypertension by serial gas chromatographic mass spectrometric determination of their urinary metabolites dinor-thromboxane B2 and dinor-6-keto-prostaglandin F1 alpha, respectively. The patients were studied until their hypertension had resolved on clinical criteria. Urinary excretion of dinor-thromboxane B2 and dinor-6-keto-prostaglandin F1 alpha was increased when the persistent pulmonary hypertension was associated with group B streptococcal (n = 2) and pneumococcal (n = 1) sepsis. Based on urinary metabolite excretion, endogenous formation of thromboxane A2 and prostacyclin did not consistently differ from normal neonates in four patients with non-septic persistent pulmonary hypertension (hyaline membrane disease (n = 2), asphyxia, and meconium aspiration). These data suggest that thromboxane A2 is not a universal mediator of persistent pulmonary hypertension. It may, however, have a role in the pathophysiology of early onset group B streptococcal disease, and persistent pulmonary hypertension of other infectious aetiology. If these findings are confirmed by further studies, thromboxane synthetase inhibition or receptor antagonism may offer a potential therapeutic approach in neonates with persistent pulmonary hypertension associated with sepsis.

The role of eicosanoids in paediatrics.

Prostanoids are unsaturated cyclic fatty acids, are synthesized primarily from arachidonic acid, and, like the leukotrienes, belong to the growing family of eicosanoids. As tissue hormones, prostanoids act on specific receptors near their site of synthesis and degradation. Prostanoids operate as modulators and mediators in a large spectrum of physiological processes. They are involved in the regulation of maternal and fetal circulation, patency of the ductus arteriosus, platelet-vessel wall interaction and kidney function. Besides their physiological function in protecting organ perfusion under stress conditions, they are also involved in diseases as described in the hyperprostaglandin E2-syndrome or--together with leukotrienes--in inflammatory processes. More specific pharmacological tools than the nonsteroidal antiinflammatory drugs, such as receptor antagonists, selective synthesis inhibitors, and eicosanoid analogues offer the prospect of enriching our arsenal of pharmacotherapeutic interventions in a variety of diseases. Before active intervention, however, more and specific biochemical analyses are required to identify the pathophysiological role of eicosanoid.

Excretion of primary prostanoids and their metabolites during acute volume expansion.

Simultaneous determination of urinary excretion rates of primary unmetabolized prostanoids and their enzymatic metabolites were performed by gas chromatography-mass spectrometry (GC/MS) or tandem mass spectrometry (GC/MS/MS). Changes in kidney function were induced by acute (4 h) volume expansion. Despite marked changes in urine flow, GFR, urinary pH, osmolality, sodium and potassium excretion, only a insignificant or transient rise in the enzymatic prostanoid metabolites (2,3-dinor-6-keto-PGF1 alpha, PGE-M, 2,3-dinor-TxB2 and 11-dehydro-TxB2) was observed. The excretion rates of the primary prostanoids were elevated in parallel with the rise in urine flow: PGE2 rose (p less than 0.05) from 14.2 +/- 4.0 to 86.2 +/- 20.7, PGF2 alpha from 60.0 +/- 4.9 to 119.8 +/- 24.0, 6-keto-PGF2 alpha from 7.2 +/- 1.3 to 51.5 +/- 17.0, and TxB2 from 11.2 +/- 3.3 to 13.6 +/- 3.6 ng/h/1.73 m2 (means +/- SEM) at the maximal urine flow. Except for 6-keto-PGF1 alpha and TxB2, this rise in urinary prostanoid levels was only transient despite a sustained fourfold elevated urine flow. We conclude that urine flow rate acutely affect urine prostanoid excretion rates, however, over a prolonged period of time these effects are not maintained. The present data support the concept that urinary levels of primary prostanoids mainly reflect renal concentrations whereas those of enzymatic metabolites reflect systemic prostanoid activity. From the excretion pattern of TxB2 one can assume that this prostanoid represents renal as well as systemic TxA2 activity.

Thromboxane receptor-mediated bronchial and hemodynamic responses in ovine endotoxemia.

The role of thromboxane A2 in sheep endotoxemia, an animal model of the adult respiratory distress syndrome, was investigated by a combined biochemical and pharmacological approach. Endogenous thromboxane biosynthesis was assessed by gas chromatographic-mass spectrometric analysis of urinary (thromboxane B2, 2,3-dinor-thromboxane B2) and plasma (11-dehydrothromboxane B2) metabolites that demonstrated a significant stimulation by endotoxin. The functional relevance of thromboxane A2 was probed with a specific thromboxane-prostaglandin endoperoxide receptor antagonist, SQ 29548. The antagonist significantly blunted the increase in pulmonary arterial pressure, pulmonary vascular resistance, lung lymph flow, and lymph protein clearance induced by endotoxin. Whereas the reduction in lung compliance caused by endotoxin was abolished, the augmented airway resistance was unaffected. From the simultaneous increase in thromboxane biosynthesis and effects of receptor blockade, it was concluded that thromboxane A2 mediates the early pathophysiological changes of sheep endotoxemia. Thromboxane receptor antagonism may offer a potential therapeutic approach to patients at risk of the adult respiratory distress syndrome.

Tandem mass spectrometric determination of 11-dehydrothromboxane B2, an index metabolite of thromboxane B2 in plasma and urine.

11-Dehydrothromboxane B2 is one of the major enzymatic metabolites of thromboxane B2 (TXB2), a biologically inactive product of thromboxane A2. The short half-life of thromboxane A2 and ex vivo production of thromboxane B2 by platelet activation make these prostanoid metabolites inappropriate as indices of systemic thromboxane biosynthesis, whereas 11-dehydro-TXB2 has been shown to reflect the release of thromboxane A2 in the human blood circulation. Analysis of 11-dehydro-TXB2 in plasma and urine was performed by gas chromatography-mass spectrometry-mass spectrometry using the chemically synthesized tetradeuterated compound as an internal standard. The high selectivity of triple-stage quadrupole mass spectrometry (tandem mass spectrometry) considerably facilitates sample purification as compared to single quadrupole mass spectrometric determination. Plasma concentrations in five healthy male volunteers were in the range 0.8-2.5 pg/ml. Urinary excretion of 11-dehydro-TXB2 was higher than that of 2,3-dinor-TXB2: 1.2 +/- 0.36 micrograms/24 h vs 0.53 +/- 0.33 micrograms/24 h (n = 5). Thus 11-dehydro-TXB2 appears at present to be the best index metabolite of systemic TXA2 activity in plasma as well as in urine.

Role of prostaglandins in hyperprostaglandin E syndrome and in selected renal tubular disorders.

Renal and systemic prostanoid activity was assessed in various renal tubular disorders, using mass spectrometric determination of urinary excretion rates of primary prostaglandins (PGE2, PGF2 alpha, PGI2, and TXA2) and their systemically produced index metabolites. Only PGE2 levels (normal range: 2.0-16.4 ng/h per 1.73 m2) are elevated in Bartter syndrome (median: 43.4, range: 6.7-166.3), nephrogenic diabetes insipidus (46.2, 12.1-1290), Fanconi syndrome (96.6, 19.3-135.5), and in a complex tubular disorder in premature infants (40.7, 22.3-132.1), for which the term hyperprostaglandin E syndrome has been introduced. In this disorder with a Bartter-syndrome-like tubulopathy, the systemic features of the disease such as fever, diarrhoea and osteopenia with hypercalciuria were associated with increased systemic PGE2 activity. In most patients the urinary excretion rate of the systemic index metabolite of PGE2 (PGE-M) was markedly elevated (1028, 285-4709; normal range: 104-664 ng/h per 1.73 m2). Hypercalciuria per se was associated neither with increased renal nor with systemic PGE2 hyperactivity. Most problems in infants with hyperprostaglandin E syndrome could be controlled by long-term indomethacin treatment in contrast to the moderate and partial effect of this treatment in patients with Fanconi syndrome. Thus increased PGE2 synthesis plays a major role in the pathogenesis of hyperprostaglandin E syndrome, while in Fanconi syndrome PGE2 hyperactivity in the kidney is a secondary event and only aggravates the water and electrolyte wastage.

Prolonged indomethacin treatment in preterm infants with symptomatic patent ductus arteriosus: efficacy, drug level monitoring, and patient selection.

Indomethacin treatment for 1 week monitored by drug level determinations was evaluated in 32 preterm infants with symptomatic patent ductus arteriosus (sPDA). Inter- and intra-individual indomethacin dispositions varied considerably with the need for marked dosage adjustments to maintain the drug level within the proposed therapeutic range. The overall success rate of this prolonged treatment was 63%. There were no significant differences between the groups of responders (n = 20), relapsers (n = 5) and non-responders (n = 7) with respect to postnatal age, sex, total indomethacin dose, and indomethacin serum concentrations. The responders, however, had significantly higher birth weights. Eighty-five percent of infants weighing more than 1000 g (n = 20) were treated successfully. Only four of these children experienced adverse reactions. The benefit-to-risk ratio was lowest in the group of infants weighing 1000 g or less (n = 12) with a success rate of only 25% and, potentially, severe adverse reactions in ten infants. In conclusion, prolonged indomethacin treatment is an alternative to conventional short-term treatment and appears to be particularly efficacious and safe in infants weighing more than 1000 g. In infants weighing 1000 g or less and suffering from severe pulmonary diseases, this treatment cannot generally be recommended. The advantage of on-line drug level monitoring during indomethacin treatment deserves further investigation.