ARF affects approximately 10% of severely ill neonates (17). The majority of ARF cases are prerenal, also called vasomotor nephropathy. Hypovolemia, hypotension, and hypoxemia are some of the main causes of prerenal ARF. In addition to the well-known risk factors (i.e., PDA, intracerebral hemorrhage, respiratory distress syndrome, necrotizing enterocolitis, pharmacotherapy), recent studies have shown that the genetic polymorphisms may contribute to ARF too. The major substances implicated in the pathogenesis of ARF in the neonate are vasoactive agents such as angiotensin II, adenosine or renal prostaglandins, and factors participating in the regulation of inflammatory pathways, called cytokines.

Angiotensin II is a potent pressor agent that increases intraglomerular pressure in the kidneys by constricting mainly postglomerular vessels (18). This effect plays a key role in the maintenance of neonatal glomerular filtration at very low perfusion pressure, as demonstrated by the deleterious effects of angiotensin-converting enzyme (ACE) inhibitors (19) and angiotensin II-receptor antagonists (20). Nobilis et al. looked at genetic variants in the ACE and the angiotensin II-receptor 1 (AT1-receptor) genes but could not detect an impact on the risk of ARF in preterm infants with a birth weight of less than 1,500 g (21). However, Harding et al. showed a relation between genetic variants in the ACE gene and the postnatal adaptation of preterm infants with GAs between 30 and 32 weeks. They found that patients with a certain genotype were at an increased risk for poor postnatal adaptation (22). Cytokines may affect renal function in several different ways (23). The body’s ability to produce cytokines varies greatly. Several studies have offered evidence supporting the contribution of genetic polymorphisms of cytokine-encoding genes to this individual variance and have postulated an association with risk for cytokine-mediated disorders in adults. Treszl et al. showed in 92 very low-birth-weight infants with severe systemic infection that high tumor necrosis factor α producer and low interleukin 6 producer genotypes were more prevalent (26%) in neonates with ARF as compared with neonates without ARF (6%) suggesting that preterm infants with systemic infection may be at increased risk for ARF if they possess the aforementioned haplotype (24).

The currently available data on the effect of genetic polymorphisms on the risk of ARF in preterm neonates are still very limited, but the investigation of polymorphisms of genes encoding for other receptors and peptides potentially involved in the complexity of the pathogenesis of ARF might increase our knowledge on the relevance of this variants in neonates who develop ARF.

Developmental changes in the GFR of preterm infants have been the subject of many studies (4,25,26). Despite
the fact that most studies included only a limited number of infants with a wide variation of postnatal age, almost all reports showed the presence of a GA-dependent increase in the GFR. Recently, the effects of GA and body weight on the GFR on day 3 of life were studied (27). GFR measurements were performed in 147 preterm infants with a GA between 23.4 and 37.0 weeks by means of a continuous inulin infusion technique. Mean GFR values increased significantly with GA and body weight. Multivariate analysis indicated that GA, but not body weight, was the major determinant for this increase in GFR. The clinical implications of this GA-dependent maturation of the GFR become apparent when one considers drugs that are primarily eliminated by glomerular filtration. Recent studies have investigated the pharmacokinetics of cephalosporins and penicillins in preterm infants (28,29,30) and showed that the clearance of ceftazidime and amoxicillin increased significantly with increasing GA (28,30). In infants with the lowest GAs, this resulted in drug accumulation, necessitating dosage adjustments base on GA (28). For aminoglycosides and glycopeptides, which are potentially more toxic than cephalosporins and other β-lactam antibiotics, the urge to adapt prescribing practices is even higher (31,32,33,34,35,36).

Betamethasone is a synthetic glucocorticoid with a potency equivalent to dexamethasone. The drug is prescribed to pregnant women with an increased risk of preterm delivery before week 34 of gestation. The objective of this treatment is to accelerate maturation of the alveolar epithelium and stimulate synthesis of lipid and protein components of the pulmonary surfactant complex to prevent hyaline membrane disease. Much of the data regarding renal responses to antenatal glucocorticoid treatment are derived from animal studies. For example, prolonged fetal betamethasone infusions have been shown to increase GFR and urine flow in both near-term fetal and newborn lambs (37). Although fetal cortisol infusion may increase fetal renal blood flow, betamethasone-induced increases in GFR result primarily from an increase in filtration fraction (37). Thus, although glucocorticoids increase blood pressure and thus will indirectly alter renal perfusion pressure, glucocorticoid-induced increases in GFR are primarily related to changes in renal vasculature resistance. This phenomenon is of interest because an increase in filtration fraction, rather than total renal blood flow, appears to be the primary mechanism for the marked perinatal increase in GFR observed in term newborn lambs (38). Antenatal betamethasone treatment significantly increases GFR in preterm newborn lambs supported by mechanical ventilation (39). The effects of prenatal exposure to betamethasone on the GFR of preterm infants have been studied by several investigators (27,40,41,42). The majority of these studies did not show an increase of the GFR during the first week of life after prenatal exposure to glucocorticoids. However, in three of these studies, creatinine clearance was used as a less reliable marker for the GFR in preterm infants, and a small number of children were studied (40,41,42). This might have prevented the authors from demonstrating an increase in the GFR in the first week of life after prenatal exposure to glucocorticoids. The only study that showed an increase in GFR was hampered by the fact that most pregnant women who were treated with betamethasone were also treated with indomethacin, thereby minimizing the number of women who were treated with betamethasone alone (27). However, betamethasone reversed indomethacin-induced decreases in GFR (27). It was hypothesized that an increase in renal plasma flow due to betamethasone may overcome intrarenal vasoconstriction secondary to the decreased synthesis of intrarenal prostaglandins by indomethacin. More recently, Allegaert and Anderson reported no impact of prenatal exposure to betamethasone on postnatal amikacin clearance in preterm neonates indicating the need for a prospective study investigating the impact of prenatal exposure to betamethasone on GFR (43).

All components of the renin–angiotensin system (RAS) exist within the fetal kidney during the early stages of development and participate as promoting factors for the growth of this organ, more specifically its angiogenesis, and have an important role in controlling intrarenal hemodynamics (44,45,46). In the early fetal stage, renin-containing cells are present in the developing intrarenal branches of the renal artery. Renin is also distributed in other vascular parts including the arcuate, interlobar, and afferent arterioles. Renin mRNA gene expression markedly increases throughout the fetal life to peak in the perinatal period. This gene could be under the influence of adrenergic input, as its expression is abolished with renal denervation (45). In early life, renin is almost exclusively detected in the juxtaglomerular apparatus. Renin acts on plasma angiotensinogen to form angiotensin I.

Angiotensin I-converting enzyme is a dipeptidyl carboxypeptidase that releases the pressor peptide angiotensin II (ANG II) from angiotensin I and inactivates bradykinin as well (47). ACE is present in both vascular and extravascular tissues (brush border) of the kidney. Although the extravascular localization of ACE is not fully known, this enzyme is found on glomerular endothelial cells at the place where the capillary invades the inferior cleft of the S-shaped body. ACE may participate in the tubular handling process of ANG II, as it has been found on the apical and basolateral membranes already in the early nephron stage. This glomerular distribution in the fetal kidney looks different as compared with the more mature kidney, where ACE is essentially found in the peritubular endothelial cells. The switch of ACE from glomerular to peritubular vessel with maturation has been well documented and occurs progressively during infancy. In addition to the renal hemodynamic regulation, the ANG II locally generated in the glomerulus also stimulates angiogenesis through the stimulation of its receptors.

It has been demonstrated that ANG II acts as a growth factor for renal cells and therefore plays a crucial role in the development of the kidney through its two receptors: AT1-R and angiotensin II receptor 2 (AT2-R) (48). Both receptors are indeed independently present in mammalian fetal kidney tissues, and AT2-R seems to predominate (48,49,50,51,52). AT2-R mRNA has been found in almost all fetal tissues, including the metanephros and undifferentiated mesenchymal and connective tissues. AT1-R has been found more specifically in the adrenal glands, liver, and kidney. Within the kidney, both AT1-R and AT2-R mRNAs are expressed in the metanephros at 14 days of gestation, when branching of the ureter bud has already started. AT1-R expression in the immature glomeruli coincides with mesangial cell differentiation from the pericyte, and continues throughout adulthood in the glomeruli and in the tubulointerstitium, whereas AT2-R expression decreases after birth except in large cortical blood vessels. AT1-R mRNA is expressed in mature glomeruli, in maturing S-shaped bodies, and in the proximal and distal tubule as well. Early in the embryologic period, AT2-R mRNA is first expressed in mesenchymal cells adjacent to the stalk of the ureter epithelium. The expression is then extended in the mesenchyme cells of the nephrogenic area and in the collecting ducts. AT2-R is also invariantly expressed in the epithelial cells of the macula densa.

Studies performed to determine the localization of AT1-R and AT2-R will help identify the specific and crucial role of ANG II for kidney development. Via AT1-R, ANG II stimulates proliferation, regulates nitric oxide synthase expression, has growth-promoting effects, and acts on glomerular mesangial and tubular cell differentiation during nephrogenesis (53). It further mediates biological actions such as maintenance of circulatory homeostasis and cell proliferation (54). Furthermore, ANG II participates in the downregulation of AT2-R and renin gene expression. In the growth-retarded fetus, AT2-R expression has been downregulated, and it has been postulated that this downregulation is associated with a higher risk of hypertension in adulthood.

Given the fundamental role of the RAS either in utero for general renal morphogenesis or during the first days of life to promote adequate glomerular filtration, administration of drugs with ACE inhibitory effects or acting as AT1-R or AT2-R inhibitors during pregnancy or during the first days of life is strictly contraindicated.

Two COX isoforms are known: COX-1, which is expressed constitutively in almost all organs, and COX-2, which is usually absent in most organs but can be induced by various stimuli (55). These enzymes have a key role in the biosynthesis of prostanoid derivatives (56). In adults, renal prostaglandin synthesis is thought to counterbalance vasoconstrictive agents (e.g., ANG II), and renal vasodilatory prostanoids are primarily derived from COX-2 (57).

In the fetus, prostaglandins are crucial in the early phase of nephrogenesis, more specifically in the glomerulogenesis and in the differentiation of the nephrons. For example, metabolites of arachidonic acid modulate the activity of Na⁺/K⁺/ATPase along the nephron, and this action is age dependent (58). Vasodilator prostanoids counteract the high vascular resistance in utero and during the first days of life. Prostaglandin E₂ (PGE₂) and prostaglandin I₂ could also act as potent and rapid stimulators of renin secretion through prostaglandin receptors located on renal juxtamedullar cells, as has been demonstrated in more mature animals (59). Experimental studies show that a constitutive cortical as well as a medullary COX-2 are overexpressed in fetal life and during the first days of life, and this accounts for the high excretion of vasodilator prostaglandins (60,61).

Numerous prostaglandin receptors have been identified, and their role in renal development has become increasingly clear (62). Four PGE₂-receptor subtypes have been identified in the kidney as well: EP1, EP2, EP3, and EP4. They are localized both on glomerular vessels (EP1, EP2, and EP4) and on different parts of the tubule (EP1, EP3, and EP4). Overexpression of some prostaglandin receptors (EP2 and EP4) has been demonstrated in the glomerular afferent vessels of the developmental kidney, allowing for increased activity of vasodilator prostanoids. The vasodilation of the afferent arteriole via these receptors is the way by which prostaglandins counteract the high vascular resistance generated by the ANG II-mediated vasoconstriction of efferent arteriola. It is the main mechanism for maintaining glomerular filtration in fetal and early postnatal life. Overexpression of the tubular EP3 receptor (located in the distal tubule and collecting duct) is needed for amniotic fluid formation and to excrete water during the first days of life (63,64,65). In addition, it has been postulated that the downregulation of the apical collecting duct water channel AQP2 also results in the excretion of hypotonic urine in utero and during the first days of life (66,67). Embryonic calcium-sensing receptor expression is another mechanism involved in the blockade of arginine vasopressin action, resulting in hypotonic urine, during antenatal life (68).

NSAIDs inhibit the enzymatic activity of both COX-1 and COX-2 and thereby block the formation of prostanoids (56). COX-2-selective NSAIDs as well as the conventional nonselective NSAIDs such as indomethacin may cause a reversible decline in GFR and renal perfusion (69). Experimental data have also shown that endogenous PGE₂ downregulates inducible nitric oxide synthase (iNOS) induction and that the decrease of PGE₂ production by indomethacin COX inhibition results in enhancement of interleukin-1β-induced steady-state iNOS mRNA levels and NO production in mesangial cells (70). Although it has not yet been demonstrated in glomerular vessels, this mechanism highlights a possible feedback mechanism that could exist between prostaglandins and NO, as has been shown for ANG II.

In the perinatal period, NSAIDs are used (a) as a tocolytic agent, (b) for closure of a PDA, and (c) to reduce polyuria in patients with congenital salt-losing tubulopathies.
Numerous case reports have shown transient fetal/neonatal oliguria following exposure to nonselective NSAIDs (13,71,72,73). In addition, Butler-O’Hara et al. (74) reported a significant prolonged rise in plasma creatinine in infants exposed prenatally to indomethacin. Moreover, Allegaert et al. showed that prophylactic administration of ibuprofen or acetylsalicylic acid had the same impact (20% reduction) on the clearance of aminoglycosides in the preterm neonate (75). NSAIDs may even cause fatal renal failure in the neonate (73). The renal pathology associated with this antenatal NSAIDs exposure is characterized by small and immature glomeruli and cystic dilations in the renal cortex (13,73). Whether these functional and histologic changes can also be attributed to an imbalance of vasodilatory prostanoids and vasoconstrictive agents needs to be demonstrated.

The availability of COX-2-selective inhibitors made some investigators believe that the detrimental effects of the nonselective NSAIDs possibly could be related to inhibition of COX-1, and that administration of COX-2-selective inhibitors would not result in perinatal renal impairment (76). However, following this initial enthusiasm about the fact that COX-2 inhibitors could be renal sparing in the perinatal period, data from recent studies have shown severe fetal oliguria (77) and even fatal renal failure in neonates antenatally exposed to the COX-2-selective inhibitor nimesulide (14,78). This might even indicate that COX-2 is more essential for normal renal development and function than COX-1. In rodents, the essential role of COX-2, but not COX-1, for proper renal development during the perinatal period has been well established, indicating that COX-2 might be more essential for normal renal development and function than COX-1 (79,80,81,82).

Based on the current literature, NSAIDs should not be used during renal development. However, indomethacin is still frequently prescribed to inhibit preterm uterine contractions before week 34 of gestation. Short-term exposure to indomethacin leads to a reduction of the GFR, whereas conflicting data exist about the effect on the GFR after long-term exposure (72,83,84,85,86,87). Animal studies have indicated that the inhibition of prostaglandin synthesis by indomethacin increases renal vascular resistance (88). This subsequently results in an impaired renal blood flow and a concomitant reduction in the GFR (88). To investigate the impact of prenatal exposure to betamethasone and indomethacin on the pharmacokinetics of ceftazidime, 136 preterm infants were studied (28). Twenty-five of these infants were treated with indomethacin alone, and 21 infants were treated with both indomethacin and betamethasone. The results of this study clearly demonstrated that prenatal exposure to indomethacin alone significantly decreases ceftazidime clearance and increases serum half-life of ceftazidime. The coadministration of betamethasone prevented these changes. These results indicate that after prenatal exposure to indomethacin alone, additional dosage adjustments are indicated (28).

NSAID-induced reduction of neonatal GFR is an important example that in utero exposure to drugs can have a profound effect on the pharmacokinetics of drugs administered to the newborn infant. Prescribing clinicians should be aware of possible fetal drug exposure and its potential consequences on neonatal renal clearing capacity.

It is recommended not to adjust dosage regimens for therapeutic agents during the first 4 weeks of life (89). Despite the fact that these recommendations are derived from studies that did not stratify infants according to postnatal age, this recommendation was made on the assumption that no significant postnatal increment in GFR has been documented in preterm infants. Previous studies on the postnatal development of GFR in preterm infants indeed show conflicting data (4,6,25,90,91). However, several investigators reported the presence of a significant increase in the GFR in the first 10 days after birth (4,6,90). This postnatal age-dependent increase in GFR could be used to predict the pharmacokinetics of drugs that are mainly eliminated by glomerular filtration.

Recent data indicate that there is a significant increase in GFR postnatally (92). GFR values in infants increase with a mean of 0.19 mL per min during the 7-day period between day 3 and day 10 after birth. In utero the weekly increase is 0.035 mL per min (27). This indicates that the postnatal increase of the GFR between days 3 and 10 after birth is 5.4 times higher compared to the intrauterine changes. Therefore, postnatal age seems to be associated with an acceleration of the maturation of the GFR. Our studies showed that this increase in the GFR resulted in a significant increase in the clearance of ceftazidime (92). These findings are consistent with the results of some investigators (93,94), whereas other studies could not find any relation with postnatal age (95,96). Kenyon et al. (96) reported that postnatal renal function maturation exerts a significant influence on the developmental pharmacokinetics of amikacin. These authors speculated, however, that the rapid maturation of the renal function in the first week of life is not present in the extremely preterm population. We showed that the rapid postnatal change in the GFR is also present in very young preterm infants and is primarily responsible for the increase in the clearance of ceftazidime (92). Dosage adjustments seem therefore already indicated during the first weeks of life despite the current recommendation not to adjust dosage regimens for therapeutic agents during the first 4 weeks of life.

PDA is a common clinical problem among preterm infants (97). The ductus arteriosus is a normal fetal vascular connection between the left pulmonary artery and the descending aorta. In utero, the ductus serves to allow the majority of blood flow leaving the right ventricle to circumvent the high-resistance pulmonary circulation and flow directly into the descending aorta. This directs oxygen-deprived blood to flow toward the placenta, the fetal source for reoxygenation. After birth, the elimination of
the low-resistance placenta results in an increase in systemic vascular resistance, and the exchange of air for fluid in the lungs creates decreased pulmonary resistance. Constriction of the ductus arteriosus and functional closure generally are spontaneous after birth, redirecting blood flow toward the lungs, which then assumes oxygenation. Factors crucial to the closure of this vessel appear to be oxygen tension, concentrations of circulating prostaglandins, and available muscle mass in the ductus. In preterm infants, higher circulating concentrations of prostaglandins, an immature ductus, and/or an immature respiratory system contribute to continued patency of the ductus (97). After birth, the increased systemic vascular resistance combined with the fall in pulmonary vascular resistance result in a shift of blood flow across the ductus from what previously was a right-to-left shunt (before birth) to a left-to-right shunt (after birth) if the ductus remains patent. This hemodynamic change can lead to a left ventricular overload, increased left-ventricular end-diastolic pressure and volume, increased left atrial pressure, and congestive heart failure (97). The physiologic consequences of PDA relate primarily to hypoxia, hypoperfusion, fluid overload, and acidosis (97,98). Although pharmacokinetic studies directly examining the differences between newborns with and without PDAs are sparse, potential pharmacokinetic changes can be easily predicted. A complicating factor, however, is the varying degree to which these changes occur among neonates with PDA, obviously leading to quite variable drug disposition. Our data showed that a PDA and postnatal exposure to indomethacin altered the aforementioned rapid postnatal increase in GFR (92). This phenomenon will probably delay the need for dosage adjustment during the first 2 postnatal weeks in these preterm infants.

The volume of drug distribution may be altered in neonates with a PDA. Drugs that distribute primarily into body water may demonstrate an increased volume of distribution, as fluid overload is common in newborns with PDA. The presence of acidosis may decrease protein binding of some drugs (e.g., theophylline) and consequently increase volume of distribution (99). Concurrent acidosis may alter the ionization of agents with a pK_a close to 7.4 (e.g., phenobarbital), permitting increased concentrations of unionized molecules that are available to cross biological membranes more freely and potentially distribute more extensively into tissue (100). Comparisons of neonatal pharmacokinetic data reveal that volume of distribution apparently is increased in the presence of a PDA for several drugs (101,102,103).

In healthy newborns, elimination of most drugs is usually diminished because of immature excretory functions. In the presence of a hemodynamically significant PDA, decreased renal and hepatic blood flow can be anticipated, potentially leading to further reductions in drug elimination capacity (98). The interpretation of drug clearance data in neonates with PDA often is confounded by the effects of mechanical ventilation, indomethacin or ibuprofen therapy, or surgical ligation, which may also influence blood flow to the liver and the kidneys (92,101,102,103,104).

The pharmacologic treatment of PDA with indomethacin further confounds pharmacokinetic interpretations because drug disposition changes related to drug interactions may be difficult to separate from those altered by the underlying disease state (92,104). The apparent accumulation of digoxin, gentamicin, amikacin, and vancomycin with concurrent use of indomethacin appears to be the consequence of the dual effect of decreasing renal elimination secondary to indomethacin and decreased volume of distribution once the PDA closes. As either or both interactions may play a role, drug concentrations should be monitored closely when indomethacin therapy is started and after it is discontinued.

Other NSAIDs have also been used to treat PDA in preterm infants but were associated with adverse effects, as in the case of sulindac (105) and mefenamic acid (106), or were less effective than indomethacin at closing the duct, as was shown for acetylsalicylic acid (107).

A relevant number of studies with ibuprofen for the treatment of PDA in preterm infants have been performed in the meanwhile. However, it was concluded from studies in neonates that ibuprofen as compared with indomethacin is effective at closing the duct and is associated with fewer cerebral and renal adverse effects (12,108,109,110). These conclusions are under debate based on animal studies that did not show any difference in renal side effects between animals treated with ibuprofen and those treated with indomethacin (10). Another large-scale randomized, controlled trial in preterm infants is being conducted aiming at several unresolved issues including very limited information on the pharmacokinetics of ibuprofen in the preterm neonate (104,111,112).

Drug disposition may be altered by the presence of a PDA and/or concomitant use of indomethacin. Close therapeutic drug monitoring is indicated because the changes in drug disposition may be abrupt with PDA closure or initiation of indomethacin therapy (92,104). More recent data have shown that the impact of prophylactic or therapeutic administration of ibuprofen on aminoglycoside or glycopeptides clearance is of a similar magnitude (113).

Dopamine has been widely used in neonatal intensive care for the treatment of hypotension or oliguria in sick preterm infants (114,115,116,117). Through stimulation of the adrenergic and dopaminergic receptors, dopamine exerts dose-dependent cardiovascular, renal, and endocrine effects that combat the clinical manifestations of shock. This discussion will be focused on the renal effects of dopamine therapy.

As long as renal perfusion pressure is within the autoregulatory range, the direct tubular rather than the renal hemodynamic action of dopamine is mostly responsible for the diuretic and saluretic effects of the drug (116). These effects are brought about by activation of the renal tubular dopamine receptors along the nephron. Data indicate the presence and functional integrity of these receptors and postreceptor mechanisms in the human kidney from as early as week 24 of gestation (114,115,116). Through selective activation of the renal vascular DA₁ and DA₂ receptors, low doses of dopamine induce an approximately 20% to 40% increase in renal blood flow without significantly influencing systemic blood pressure
(118). Whole-kidney GFR shows a variable, approximately 5% to 20% rise (118). Findings of micropuncture studies indicate that the mechanism of the drug-induced increase in the GFR is the enhancement of glomerular ultrafiltration pressure caused by a more pronounced vasodilation of the afferent than the efferent arteriole (119).

In cases of aminoglycoside toxicity, dopamine, by increasing renal blood flow and GFR, may be useful in facilitating renal excretion of these antibiotics. Dopamine also modifies the renal effects of furosemide and indomethacin (120,121). The recently discovered mechanisms of the renal actions of dopamine enable us to better understand the cellular basis and nature of these interactions. Independent of age and maturation, dopamine enhances the diuretic effect of furosemide in the anuric–oliguric patient (120). This interaction is thought to be the consequence of the dopamine-induced selective augmentation of medullary circulation and thus the enhancement of the delivery of furosemide to its site of action in the kidney (120). Inhibition of prostaglandin synthesis by indomethacin may cause severe, although usually transient, renal side effects in preterm infants and volume-depleted children. Although these findings need to be confirmed, dopamine attenuates the indomethacin-induced decrease in urine output and sodium excretion in sick preterm infants with PDA (121). However, the renal vasoconstrictive actions of indomethacin have not been reported to be influenced by dopamine (121). Dopamine interacts with the renal prostaglandin system mainly at the tubular level (122), and it is not surprising that the renal tubular but not the vascular actions of indomethacin are attenuated by the drug. This is further supported by a recent Cochrane analysis showing that there is no evidence from randomized trials to support the use of dopamine to prevent renal dysfunction (primarily GFR) in indomethacin-treated preterm infants and the lack of effect of dopamine administration on amikacin clearance (123,124).

In conclusion, the administration of dopamine will enhance the renal clearing capacity of the preterm infant and can attenuate the detrimental side effects of high levels of potentially nephrotoxic drugs (i.e., aminoglycosides, indomethacin). Whether it should be administered prophylactically as a protective agent during, for example, indomethacin treatment needs to be prospectively validated. Moreover, the effect of dopamine administration on recommended dosing schemes of, for example, antibacterial agents have not been studied. The clinician should be aware that the use of dopamine might lead to subtherapeutic drug concentrations due to the increase in GFR.

Furosemide, a loop diuretic, is frequently administered to critically ill newborn infants to augment urine output and relieve pulmonary edema. Furosemide acts on the luminal side of the renal tubule at the thick ascending limb or the loop of Henle (125). It inhibits chloride reabsorption, thereby inhibiting passive reabsorption of sodium, and must be cleared by the kidney into the tubular fluid to exert its diuretic effect (126). Because the diuretic effect of furosemide is directly related to the renal tubular drug concentration, its effectiveness is correlated with the degree of renal function (126). In addition to the distal action, free water clearance is increased by an inhibition of carbonic anhydrase activity in the proximal tubule (127).

A number of hemodynamic responses contribute to the diuretic action of furosemide. Total renal vascular blood flow is increased (128), renal cortical blood flow is redistributed, and renin secretion is stimulated from the juxtaglomerular cells to the kidney (129). The mechanisms controlling each of these responses to furosemide are only partially defined but appear to be mediated, at least in part, through the prostaglandin system, which is activated almost immediately after furosemide administration (130). In a very recent study in critically ill pediatric patients, furosemide was observed to induce a prompt and generalized increase of the arachidonic acid-derived prostaglandins concomitant with increased renin production (131). Furosemide-induced increased prostaglandins may have played an integral role in the causation of the ensuing hemodynamic, diuretic, and neurohormonal changes. There is evidence to support the theory that furosemide increases renin release through an increase in renal prostaglandins (132). Diuresis may occur in response to increased renal blood flow induced by the increase in renal prostaglandin production. Furosemide causes a rapid diuresis in newborns of all GAs after parenteral administration. In premature infants, the onset of action is evident within 1 hour, but peak diuresis does not occur until 1 to 3 hours after dosing, with a duration of diuresis of approximately 6 hours (133). In very low-birth-weight infants, the plasma half-life exceeds 24 hours, and accumulation of furosemide to potentially ototoxic levels can occur when the drug is administered every 12 hours (134). By stimulating prostaglandin synthesis at both vascular and tubular sites in the kidney, furosemide prevents the occurrence of both the renal hemodynamic and the tubular side effects of indomethacin (135). However, the long-term furosemide administration also increases the occurrence of PDA in preterm infants (136), so extensive use of this drug may not be prudent in preterm infants treated with indomethacin to close their duct. In addition, it was recently shown in a study of critically ill pediatric infants that administration of furosemide can induce a decrease in cardiac output and an increase in systemic vascular resistance, potentially increasing the risk for paradoxical pulmonary edema (131). The clinician needs to consider such hemodynamic alterations when administering furosemide in critically ill preterm infants. As an alternative, a continuous infusion with furosemide may perhaps lead to more controlled diuresis with fewer hemodynamic alterations. Very recently this suggestion has been investigated in near-term neonates on extracorporeal membrane oxygenation (137). Clearly, this way of administering furosemide needs clinical investigation in preterm neonates.