Acute Kidney Injury and Chronic Kidney Disease

C H A P T E R

85

Acute Kidney Injury and Chronic Kidney Disease David Askenazi, Lorie B. Smith, Susan Furth, and Bradley A. Warady

ACUTE KIDNEY INJURY

ll

DEFINITION Previously referred to as acute renal failure, acute kidney injury (AKI) is characterized by a sudden impairment in kidney function, that results in the retention of nitrogenous waste products (e.g., urea) and alters the regulation of extracellular fluid volume, electrolytes, and acid-base homeostasis. The term acute kidney injury has replaced acute renal failure by most critical care and nephrology societies, primarily to highlight the importance of recognizing this process at the time of injury as opposed to waiting until failure has occurred (Mehta, 2007). Despite its limitations, serum creatinine (SCr) is the most commonly used measure to evaluate glomerular filtration in the clinical setting and is more specific than blood urea nitrogen (BUN). BUN is an insensitive measure of glomerular filtration rate (GFR) because it can be increased out of proportion to changes in GFR with high dietary protein intake, gastrointestinal bleeding, use of steroids, and hypercatabolic states. If the BUN-to-SCr ratio exceeds 20, increased urea production or increased renal urea reabsorption that occurs in prerenal azotemia should be suspected (Feld et al, 1986). In the neonatal population, the most common SCr ­cut-point used to define AKI has arbitrarily been set at 1.5 mg/dL or greater, independent of day of life and regardless of the rate of urine output. Although SCr is the most common method to diagnose AKI, it has significant shortcomings, including ll

ll

ll

ll

ll

SCr concentrations may not change until 25% to 50% of the kidney function has already been lost; therefore it may be days after an injury before a significant rise in SCr is seen (Brion, 1986). At a lower GFR, SCr will overestimate renal function because of tubular secretion of creatinine (Brion, 1986). SCr varies by muscle mass, hydration status, sex, age, and gender. Different methods (Jaffe reaction versus enzymatic) produce different values, and medications and bilirubin can affect SCr measured by the Jaffe method (Lolekha, 2001; Rajs and Mayer, 1992). Once a patient receives dialysis, SCr can no longer be used to assess kidney function because SCr is easily dialyzed.

Additional problems with using SCr as a measure of AKI specific to neonates include ll

SCr in the first few days of life reflects the mother’s SCr; thereafter, the normal distribution of SCr values demonstrates variation that is greatly dependent on the level of prematurity and age (Gallini, 2000) (Figure 85-1)

ll

Normal nephronogenesis in healthy term infants begins at 8 weeks’ gestation and continues until 34 weeks’ gestation, at which time the number of nephrons (1.6 to 2.4 million) approximates that of an adult (Abrahamson, 1991). Depending on the degree of prematurity, GFR steadily improves from 10 to 20 mL/min per 1.73 m2 during the first week of life to 30 to 40 mL/min per 1.73 m2 by 2 weeks after birth, concomitant with alterations in renal blood flow. GFR improves steadily over the first few months of life (Brion, 1986) (Table 85-1). Bilirubin levels in premature infants are normal at birth, rise in the first several days, and return to normal after a few weeks; this can have an impact if the Jaffee method of SCr determination is used (Lolekha, 2001).

Other problems with using a threshold cutoff to define AKI is that this approach fails to delineate the severity, timing, and cause of the injury. In the adult and pediatric populations, classification definitions of AKI based on SCr and urine output have gained acceptance. The two most common classification schemes that delineate different severities of AKI are the Risk, Injury, Failure, Loss, and End-Stage Renal Disease (RIFLE) (Bellomo et al, 2004) and the Acute Kidney Injury Network (AKIN) (Mehta, 2007) classifications. These AKIN classification definitions have created commonality in defining AKI, and they clearly show that incremental degrees of AKI independently affect survival after correcting for comorbidities, complications, and severity of illness in pediatric and adult studies (­Abosaif et al, 2005; Chertow et al, 1998; Hoste et al, 2006). These data suggest that patients may not only die with kidney failure, but that this dysfunction causes functional and transcriptional changes in the lungs and other organs that ultimately lead to poor outcomes (­Bellomo et  al, 2004; Elapavaluru and Kellum, 2007; Hoste et al, 2006). These studies have not been reproduced in critically ill neonatal populations. In children, Akcan-Arikan et al (2007) proposed a modified pediatric RIFLE (pRIFLE) classification that is similar to the adult RIFLE classification except for a lower cutoff of SCr to achieve the F category, thereby reflecting the fact that children have a lower baseline SCr. An SCr of 4.0 mg/ dL is not needed to have severe dysfunction. Similar classification definitions of AKI are greatly needed to better describe the incidence and outcomes of AKI in different populations of critically ill neonates. Delineation between the 1st week of life (when the infant has a high SCr level that will slowly fall) and changes in SCr level after the 1st week will be needed in a neonatal AKI classification system. Despite these working classification systems, the diagnosis of AKI is problematic, because current diagnosis relies on two functional abnormalities: functional changes in SCr (marker of GFR) and oliguria. Both of these 1205

1206

PART XVII  Renal and Genitourinary Systems 150.0 140.0 GA <27 wks GA 27-28 wks GA 29-30 wks GA 31-32 wks

Serum creatinine (M/L)

130.0 120.0 110.0 100.0 90.0 80.0 70.0 60.0 50.0 FIGURE 85-1  Serum creatinine concentrations (μM/L) during the first days of life, with values given as means and standard error for infants born at different gestational ages.

40.0 30.0

0

1

2

3

4

1-3 days

14.0 ± 5

1-7 days

18.7 ± 5.5

4-8 days

44.3 ± 9.3

3-13 days

47.8 ± 10.7

1.5-4 months

67.4 ± 16.6

8 years

103 ± 12

7 10 Days of life

17

24

31

Adapted from Schwartz GJ, Furth SL: Glomerular filtration rate measurement and estimation in chronic kidney disease, Pediatr Nephrol 22:1839-1848, 2007. GFR, Glomerular filtration rate.

measures are late consequences of injury and not markers of the injury itself. The ideal marker to detect AKI should be upregulated shortly after an injury and be independent of the GFR (Askenazi, 2009a). Current studies of urinary and serum biomarkers of AKI promise to improve our ability to diagnose AKI early in its disease process. For example, urine and serum neutrophil gelatinase-associated lipocalin, urine interleukin-18, kidney injury marker 1, and others have been shown to predict which neonates undergoing cardiopulmonary bypass will develop a rise in SCr level by greater than 0.5 mg/dL (Mishra et al, 2005; Parekh et al, 2006) (Figure 85-2). Creating AKI definitions using early injury biomarkers, which can ultimately predict morbidity and mortality, is of paramount importance. In addition, well-designed clinical research studies on early noninvasive biomarkers of AKI in different neonatal populations are needed to better characterize AKI and predict outcomes. Once we can reliably identify AKI early in the disease process, preventive and therapeutic interventions can be studied to improve outcomes in neonates with AKI.

EPIDEMIOLOGY Critically ill neonates are at risk of AKI because they are commonly exposed to nephrotoxic medications and have frequent infections that lead to multiple organ failure (Andreoli, 2004). The exact incidence of neonatal AKI is difficult to quantify because infants commonly

38

45

52

IL-18 (pg/ml) AKI

m2) 200 Mean biomarker level

GFR (mL/min per 1.73

6

250

TABLE 85-1  Insulin Clearance GFR in Healthy Premature Infants

Age

5

150

100

NGAL (ng/ml) AKI

50 IL-18 (pg/ml) no AKI NGAL (ng/ml) no AKI

0

2

4

6

12

24

48

Time after cardiopulmonary bypass (h) FIGURE 85-2  Mean values of urine interleukin-18 (pg/mL) and neutrophil gelatinase-associated lipocalin (ng/ml) over the first hours after cardiopulmonary bypass in infants who develop acute kidney injury (50% increase in serum creatinine) compared with those who did not develop acute kidney injury.

have nonoliguric renal failure and may therefore not be screened with SCr for AKI. Additionally, most reports define only severe cases of AKI (usually SCr >1.5 mg/dL or need for renal replacement therapy [RRT]) (Agras et al, 2004; Stapleton et al, 1987). Published studies estimate the incidence of neonatal AKI in critically ill neonates between 8% to 24% and mortality rates between 10% to 61% (Andreoli, 2004). Moghal et al (1998) suggested that relative to adult and pediatric critically ill populations, neonates have the highest incidence of AKI.

Infants With Perinatal Hypoxia Most studies of neonatal AKI describe term infants with asphyxia at the time of birth. Hypoxic-ischemic encephalopathy accounts for 23% of the 4 million neonatal deaths globally and high rates of disability (Lawn et al, 2005).

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

Three different observational studies describe the incidence of AKI (defined as SCr >1.5 mg/dL) in asphyxiated critically ill newborns. Karlowicz and Adelman (1995) compared term infants with severe asphyxia (according to Portman’s asphyxia morbidity scoring system) (Portman et al, 1990) to similar infants with moderate asphyxia scores. They found that AKI occurred in 20 of 33 (66%) of infants with severe asphyxia compared to 0 of 33 (0%) in those with moderate asphyxia. In a case control analysis (matching for gestational age and birthweight in otherwise healthy newborns), Aggarwal et al (2005) observed the incidence of AKI in infants with 5-minute Apgar scores of 6 or less to be 56% versus 4% in controls. Similarly, Gupta et al (2005) found a 47% incidence of AKI and 14.1% mortality in infants with Apgar scores of 6 or less. All these studies report more than 50% of AKI cases to be nonoliguric, which highlights the insensitivity of oliguria in predicting AKI in neonates.

General Neonatal Population Three recent studies have explored AKI in the general critically ill neonatal population and confirmed the high incidence of AKI and mortality in these infants. In a retrospective analysis, Agras et al (2004) found 25% hospital mortality in neonates with AKI. Many (47%) of their patients had nonoliguric renal failure, and premature infants constituted 31% of the cases. Mathur et al (2006) prospectively studied mostly term neonates with sepsis and found a 26% incidence of AKI. The mortality rate was significantly higher in those with AKI versus those without AKI (70.2% versus 25%; p <0.001). Although this study gives insight into the incidence of AKI in neonates, this study is limited by their choice of the definition of AKI (BUN >20) and their inability to control for gestational age, birthweight, comorbidity, and severity of illness. To better ascertain the independent role of AKI on survival in premature infants, a recent case-control study ­(Askenazi et al, 2009b) matching premature infants by gestational age and birthweight found that for every 1 mg/dL increase in SCr, the odds ratio for death increased by a factor of almost 2 (odds ratio [OR], 1.94; 95% confidence interval [CI], 1.13 to 3.32). The OR for death increased even when confounding variables were adjusted (adjusted OR, 3.44; 95% CI, 1.23 to 9.61). Because of study design limitations, caution must be exercised in making inferences about the true incidence and outcomes of neonatal AKI based on these studies. Additional prospective, multicenter studies are needed to determine whether this association exists.

Infants Requiring Cardiopulmonary Bypass The incidence and outcome data for neonates who require cardiopulmonary bypass are based on single-center experiences, most of which define AKI as a requirement of RRT. Reported incidence of AKI in this population is between 2.7% and 24.6% (Bailey et al, 2007; Picca et  al, 1995), with survival rates ranging from 21% to 80% (Picca et al, 1995; Sorof et al, 1999). The causes of AKI in this group are based on preoperative factors, intraoperative changes, and postoperative events. Preoperative risk factors for

1207

AKI include hypotension, use of diuretics, angiotensinconverting enzyme inhibitors [ACE-I], indomethacin, hypoxemia, hypothermia, infection, positive pressure ventilation, nephrotoxic medications, and intravenous contrast. Intraoperative changes including hemodynamic changes, aortic clamp time, and inflammatory response to cardiopulmonary bypass influence postoperative AKI. The most important factors that influence the development of postoperative AKI are cardiac performance and sepsis (Picca et al, 2008).

Infants Requiring Extracorporeal Membrane Oxygenation Several studies of infants and children (Cavagnaro et al, 2007; Meyer et al, 2001; Sell et al, 1987; Shaheen et al, 2007; Weber et al, 1990) who receive extracorporeal membrane oxygenation suggest both AKI and RRT are associated with mortality. To explore the independent role of AKI and receipt of RRT on outcomes, Askenazi et al (2011) performed a retrospective cohort study of 7941 neonates enrolled between 1998 and 2008 in the extracorporeal lifesupport organization registry. AKI was defined as infants in the registry who had an SCr level greater than 1.5 mg/ dL or an International Classification of Diseases code for acute renal failure. Neonatal mortality was 2175 in 7941 (27.4%). Nonsurvivors experienced more AKI (413/2175 [19%] versus 225/5766 [3.9%]; p <0.0001) and more of them received RRT (863/2175 [39.7%] versus 923/5766 [16.0%]; p <0.0001) than survivors. After adjusting for confounding variables, the adjusted OR for the neonatal group was 3.2 (p <0.0001) after AKI and 1.9 (p <0.0001) with RRT. Additional studies to ascertain AKI risk factors, test novel therapies, and optimizing the timing and delivery of RRT can positively affect survival. One of the most common morbidities of prematurity is the propensity to develop bronchopulmonary dysplasia; it affects 10% and 40% of surviving very low-birthweight and extremely low-birthweight infants, respectively (Eichenwald and Stark, 2008). The pathophysiology of this chronic lung condition involves elevated levels of proinflammatory interleukins, tumor necrosis factor-α, leukotrienes, and increased pulmonary vasculature permeability, which culminate in abnormal lung development and fibrosis. Not only does AKI cause pulmonary edema secondary to volume overload, but evidence in ischemic, nephrotoxic, and bilaterally nephrectomized animal models shows that AKI induces a proinflammatory process highlighted by increase in neutrophils, tumor necrosis factor-α, interleukins, free radicals, endothelial growth factors, and granulocyte colony-stimulating factor (Faubel, 2008; Hoke et al, 2007; Kim do et al, 2006). Clinically it has been recognized that critically ill adults receiving mechanical ventilation with AKI have a dismal prognosis (80% mortality), and they have an impaired ability to wean from mechanical ventilation (Vieira et al, 2007). To date, little is known about the lung-kidney interactions in premature infants, nor the role of AKI on chronic lung disease. A large prospective cohort study with classification definitions of AKI is greatly needed to better understand the incidence and independent effect of AKI in asphyxiated infants, premature infants, infants undergoing cardiopulmonary

1208

PART XVII  Renal and Genitourinary Systems

bypass, and the general critically ill newborn population. The role of the kidney in acute and chronic pulmonary disease in premature infants needs to be explored.

PATHOPHYSIOLOGY Prerenal Azotemia Prerenal azotemia (sometimes referred to as pre–kidney failure, but perhaps better termed acute kidney success) occurs in response to decreased renal blood flow (RBF). Causes of prerenal azotemia in neonates include loss of effective circulating blood volume (perinatal blood loss, hemorrhage), dehydration (diarrhea, transepidermal free water losses, poor intake, gastric or chest tube losses), capillary leak (hydrops, infection, or hypoalbuminemia), increased abdominal pressure (necrotizing enterocolitis, ascites, repair or reduction of gastroschisis, omphalocele, or congenital diaphragmatic hernia), and decreased cardiac output (cardiac surgery, heart failure, or the use of extracorporeal membrane oxygenation, which results in a lack of pulsatile flow) (Liem et al, 1995a, 1995b). Nonsteroidal antiinflammatory drugs (NSAIDs), such as indomethacin and angiotensin-converting enzyme inhibitors (ACE-Is), can decrease RBF (Box 85-1). When low RBF occurs, renal autoregulation preserves GFR by increasing renal sympathetic tone, activating the renin-angiotensin-aldosterone system, and increasing activation of hormones such as vasopressin and endothelin. An increase in filtration fraction (GFR/RBF × 100) increases peritubular oncotic pressure, resulting in enhanced proximal tubular sodium and water reabsorption (Feld et al, 1986) in those with intact tubular function. These renal hemodynamic changes lead to decreased water and sodium losses, so as to maintain systemic volume expansion and blood pressure. Oliguria does not develop in some newborns because of poor vasopressin secretion, weak renal responsiveness to vasopressin (Dixon and Anderson, 1985), poor tubular function in underdeveloped tubular cells, or after prolonged or severe hypoperfusion. In the context of renal hypoperfusion, correction of the underlying condition restores normal renal function unless renal hypoperfusion has been sufficiently severe or prolonged that renal parenchymal damage has already developed. Once parenchymal damage occurs, renal tubular cell damage (acute tubular necrosis) occurs even if renal perfusion is restored.

Intrinsic Acute Kidney Injury Prolonged or severe hypoperfusion is the most common cause of intrinsic AKI. Other causes of intrinsic AKI include nephrotoxic medications and sepsis, which can cause AKI in both hypodynamic and hyperdynamic blood flow. Approximately 6% to 8% of newborns admitted to NICUs have intrinsic AKI, with severe perinatal asphyxia being the most common cause (Stapleton et al, 1987). Other rare causes of AKI include renal vein thrombosis, renal artery thrombosis, uric acid nephropathy, hemoglobinuria, and myoglobinuria (see Box 85-1). Congenital abnormalities of the kidneys and urinary tract (CAKUT) are discussed further in the section under chronic kidney disease (CKD) of the newborn.

BOX 85-1  Causes of Acute Kidney Injury in the Newborn* Prerenal ­Azotemia Loss of effective blood volume Absolute loss Hemorrhage Dehydration Relative loss ↑ Capillary leak Sepsis NEC RDS ECMO Hypoalbuminemia Renal hypoperfusion Congestive heart failure Pharmacologic agents Indomethacin Tolazoline ACE inhibitors

Intrinsic Acute Kidney Injury

Obstructive Renal Failure

Acute tubular necrosis Severe renal ischemia Nephrotoxins Infections Congenital infections Pyelonephritis Bacterial endocarditis Renal vascular causes Renal artery thrombosis Renal vein thrombosis DIC Nephrotoxins Aminoglycosides Indomethacin Amphotericin B Radiocontrast dyes Acyclovir Intrarenal obstruction Uric acid nephropathy Myoglobinuria Hemoglobinuria Congenital malformations Bilateral renal agenesis Renal dysplasia Polycystic kidneys

Congenital ­malformations Imperforate prepuce Urethral stricture PUV Urethral diverticulum Ureterocele Megaureter UPJ obstruction Extrinsic compression Sacrococcygeal teratoma Hematocolpos Intrinsic obstruction Renal calculi Fungus balls Neurogenic bladder

ACE, Angiotensin-converting enzyme; DIC, disseminated intravascular coagulation; ECMO, extracorporeal membrane oxygenation; NEC, necrotizing enterocolitis; PUV, posterior urethral valve; UPJ, ureteropelvic junction. *In many cases, a combination of several causative factors contributes to the development of acute renal failure. For example, absolute hypovolemia, increased capillary leak–induced loss of effective circulating blood volume, and reflex renal vasoconstriction—all can contribute to renal hypoperfusion and ensuing renal injury in newborns with severe forms of shock.

Ischemic Acute Kidney Injury Despite being the best-oxygenated organ, the kidney is susceptible to hypoxic-ischemic injury because of the redistribution of its blood flow under pathologic circumstances to the vital organs, and because of the unique vascular supply of the renal medulla. The presentation and course of renal damage depend on the severity and duration of the insult. In contrast to prerenal azotemia, renal function abnormalities in intrinsic AKI are not immediately reversible. The severity of intrinsic AKI ranges from mild tubular dysfunction, to acute tubular necrosis, to renal infarction and corticomedullary necrosis with irreversible renal damage (Feld et al, 1986). Prerenal azotemia and ischemic AKI represent extremes on a continuum of physiologic responses. The main difference between prerenal AKI and ischemic AKI is that in the latter, hypoperfusion induces injury to renal parenchymal cells, particularly to tubular epithelium of the terminal medullary portion of the proximal tubule (S3 segment) and of the medullary portion of the thick ascending limb of the loop of Henle. The course of ischemic AKI can be subdivided into prerenal, initiation, extension, maintenance, and recovery phases (Sutton et al, 2002) (Figure 85-3). If during prerenal azotemia, restoration of renal blood flow occurs, GFR can return promptly to normal. The initiation phase includes the original insult and the associated events resulting in a

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

100

Pr

er

en

al

A

GFR (%)

Initiation B n sio ten Ex

ry

ve

o ec

C Maintenance

R

0 0

1

2

3

4

5

6

7

Days FIGURE 85-3  Schematic representation of stages of the progression in acute kidney injury.

drop in GFR. Tubular dysfunction with low GFR represents the maintenance phase. The duration of the maintenance phase depends, at least in part, on the severity and duration of the initial insult. The recovery phase is characterized by the gradual restoration of GFR and tubular function, which can take months to occur. During the maintenance and recovery phases of AKI, the kidney is susceptible to further damage from additional insults. Recognition of the different phases of intrinsic AKI is helpful in the diagnosis, clinical management, and prognostication of the disorder. The histologic hallmark of severe ischemic AKI is damage to epithelial tubular cells with a characteristic bleb formation and loss of brush border in the apical portion of the cell, cytoskeleton disruption, and loss of tight junctions between cells. If injury is severe enough, apoptosis and necrosis will occur with resultant desquamation of cells, which lead to tubular obstruction. Tubular epithelial cells are critical in the pathophysiology of ischemic AKI, and damage to the innermost lining of the renal vascular system, endothelial cells, has a critical role in the initiation, extension, maintenance, and recovery phases of ischemic AKI (Basile, 2007; Basile et al, 2001; Molitoris and Sutton, 2004). When endothelial cell damage occurs, activation of vasoconstriction, impaired vasodilation, and impaired leukocyte adhesion result in capillary obstruction and distorted peritubular capillary morphology. Capillary obstruction and impaired morphology lead to a cycle of increasing ischemia and vascular inflammation. The loss of endothelial cell function may represent an important therapeutic target in which vascular support or endothelial regeneration by progenitor cells can affect the short- and long-term consequences of AKI (Liu and Brakeman, 2008). Damaged endothelial and tubular cells do not only lead to dysfunction within the kidney; they produce a systemic inflammatory response that has been shown to lead to significant distant organ dysfunction. The inflammatory dysregulation is due, at least in part, to dysfunctional immune, inflammatory, and soluble mediator metabolism. AKI has also been shown to directly affect the brain, lung, heart, liver, bone marrow, and gastrointestinal tract (Awad and Okusa, 2007; Sorof et al, 1999). Mice with AKI (induced by bilateral renal ischemia for 60 minutes) had increased levels

1209

of proinflammatory chemokines, keratinocyte-derived chemoattractant, and granulocyte colony-stimulating factor in the cerebral cortex and hippocampus, which results in increased neuronal pyknosis and microgliosis in brain (Liu and Brakeman, 2008). In the lung, mechanistic studies demonstrate that AKI induces increased pulmonary vascular permeability, soluble and cellular inflammation, and dysregulated salt and water channels. Because neurologic and pulmonary morbidity is high in the critically ill neonatal population, the potentially deleterious effects of AKI on these organs need to be explored.

Nephrotoxic Acute Kidney Injury Pharmacologic agents are the most common cause of nephrotoxic AKI in neonates, although endogenous substances (hemoglobin and myoglobin) may also be toxic to the kidney. These toxins can cause neonatal AKI by decreasing renal perfusion (NSAIDs, diuretics, ACE-Is), direct tubular injury (aminglycosides, cephalosporins, amphotericin B, rifampin, vancomycin, NSAIDs, contrast media, myoglobin–hemoglobin), interstitial nephritis, and tubular obstruction (acyclovir). Although the following discussion is not a comprehensive review, some of the most common nephrotoxic medications in neonates are described. Indomethacin, a prostaglandin inhibitor used to treat patent ductus arteriosus in premature infants, is one of the most commonly used medications in the neonatal intensive care unit. Severe, although usually transient, nephrotoxicity can occur with indomethacin administration. The primary mechanism of renal action of these drugs is the potentiation of vasoconstrictive and sodium- and water-retaining effects of angiotensin II, norepinephrine, and vasopressin by the indomethacin-induced inhibition of renal prostaglandin production. Because neonatal renal function is more dependent on local prostaglandin production than that of the euvolemic adult—especially when intravascular volume is decreased as a result of fluid restriction, increased capillary leak, and transepidermal water losses in the preterm infant with patent ductus arteriosus—indomethacin administration is commonly associated with elevated SCr concentrations, decreased urine output, and hyponatremia (Cifuentes et al, 1979). Amphotericin B alters renal function by directly affecting tubular function, resulting in renal tubular acidosis and increasing urinary potassium excretion. Although these nephrotoxic effects are most often reversible, cases of fatal neonatal renal failure caused by amphotericin B toxicity have been reported (Baley et al, 1984). Amphotericin B lipid complex (ABLC), liposomal amphotericin and other lipid formulations of this drug consist of nonliposomal lipid bilayers complexed with amphotericin B. This lipid bilayer causes a higher affinity to fungal rather than mammalian cellular membranes and therefore is less nephrotoxic (Wurthwein et al, 2005). Auron et al (2007) recently showed no differences in blood urea nitrogen or SCr, serum sodium, and serum potassium in 35 premature infants (average birthweight, 764 ± 196 g) compared with similar infants (controlling for gestational age and birthweight). Therefore a 2-week course of amphotericin B complex is likely to be safe in premature infants, although studies to explore longer use of ABLC are needed.

1210

PART XVII  Renal and Genitourinary Systems

Aminoglycosides are one of the most commonly used medications in the treatment of suspected or proven neonatal sepsis. Aminoglycosides inhibit lysosomal phospholipases, leading primarily to proximal tubule cell damage (Giuliano et al, 1984), although changes in the ultrastructure of the glomerulus also occur (Ojala et al, 2001). A metaanalysis of 11 studies in septic neonates demonstrated that both once-a-day and multiple-dose regimens led to adequate clearance of sepsis. Although rates of ototoxicity or nephrotoxicity were not different among the two groups, pharmacokinetic studies reveal that once-a-day dosing causes less drug accumulation in the kidney’s ­proximal cells, and it usually achieved adequate peak concentrations (>5.0 μg/dL) while avoiding toxic trough levels (<2 μg/dL) (Rao et al, 2006). Aminoglycosides should be used with caution in any patient with renal dysfunction, concomitant nephrotoxic medication use, or poor renal perfusion owing to volume, hypoalbuminuria, or heart failure. In patients with poor renal dysfunction or risk for impaired aminoglycoside clearance, serial monitoring to assure proper clearance of medication is needed to prevent AKI. Because aminoglycoside toxicity is usually nonoliguric, serial monitoring of SCr values is necessary, especially during prolonged administration of these antibiotics, to detect their potential nephrotoxicity in the newborn. Acyclovir has replaced vidarabine for treatment of herpes simplex disease because of its ease of use and more favorable side effect profile. High-dose acyclovir (60 mg/kg/day for 21 days) decreased the mortality and CNS disease from sepsis and meningitis associated with herpes simplex virus to 29% and 4%, respectively (Kimberlin et al, 2001). Acyclovir is an antiviral agent that is eliminated rapidly in urine through glomerular filtration and tubular secretion. It is nearly insoluble in urine and may precipitate, particularly in the distal tubular lumen. Intravenous high-dose acyclovir treatment can lead to intratubular crystal precipitation and renal failure. Acyclovir-related nephrotoxicity can be limited by avoiding its use in those with renal insufficiency or intravascular volume depletion, infusing the drug slowly (over several hours), and by assuring adeqaute hydration to maintain high urinary flow rate, which will reduce the likelihood of crystal deposition in tubules (Izzedine et al, 2005). Finally, medications given to pregnant women can cause combined ischemic and nephrotoxic renal injury in the fetus, resulting in the clinical presentation of AKI after birth. A frequently used class of medications that fall into this category are the NSAIDs prescribed for tocolysis— both nonselective and selective cyclooxygenase inhibitors such as indomethacin or ketoprofen (Bannwarth et al, 1999; Peruzzi et al, 1999)—which can lead to severe AKI in the newborn.

Acute Obstruction The most common cause for obstruction induced kidney dysfunction in the newborn is congenital malformations, including imperforate prepuce, urethral stricture, prunebelly syndrome, and posterior urethral valves. Other causes of acute obstruction include neurogenic bladder, extrinsic compression (e.g., hematocolpos, sacrococcygeal teratoma), and intrinsic obstruction from renal calculi or

fungal balls. Depending on the cause and associated damage to the kidneys, relieving the obstruction will markedly improve renal function.

EVALUATION The pregnancy history, findings on prenatal tests, vital signs, changes in neonatal body weight, physical examination, interventions, and medications prescribed provide important clues about the cause of neonatal AKI. Serum laboratory values to be monitored in the infant with AKI include serum sodium, potassium, chloride, bicarbonate, calcium, phosphorus, magnesium, urea, creatinine, uric acid, glucose, blood gases, hemoglobin, and platelets. As discussed in the section defining AKI, SCr often does not rise for days after an injury, thus monitoring these values for several days after the inciting event is necessary to determine if AKI occurred. If urine is available, then urinalysis, urine culture, and a spot urine sample for sodium, creatinine, and osmolality can help to differentiate the cause. One of the major goals in the initial evaluation of neonatal AKI is to determine whether the kidney is hypoperfused. Several laboratory, clinical, and therapeutic interventions can help to delineate prerenal azotemia from intrinsic AKI (Table 85-2). Decreased body weight, tachycardia, dry mucous membranes, poor skin turgor, flattened anterior fontanel, and elevated serum sodium levels can be seen in infants with low intravascular volumes. Measurement of serum albumin will alert the physician if appropriate oncotic pressure is present (serum albumin >2.0, but preferable at 2.5 mg/dL). When the kidney is hypoperfused, it will avidly retain sodium and water to preserve overall intravascular volume. Preservation of urine sodium

TABLE 85-2  Diagnostic Indices Suggestive of Prerenal Azotemia versus Intrinsic Acute Kidney Injury in the Newborn

Prerenal ­Azotemia

Intrinsic Acute ­Kidney Injury

Urine flow rate (mL/kg/h)

Variable

Variable

Urine osmolality (mOsm/L)

>400

≤400

Urine-to-plasma osmolal ratio

>1.3

≤1.0

Urine-to-plasma creatinine ratio

29.2 ± 1.6*

9.7 ± 3.6*

Urine [Na+] (mEq/L)

10-50

30-90

FENa† (%)

<0.3 (0.9 ± 0.6)*

>3.0 (4.3 ± 2.2)*

Renal failure index‡

<3.0 (1.3 ± 0.8)*

>3.0 (11.6 ± 9.5)*

Response to fluid challenge

Improved tachycardia, increased UOP

No effect on tachycardia or UOP

Data from Feld LG, Springate JE, Fildes RD: Acute renal failure. I. Pathophysiology and diagnosis. J Pediatr 109:401-408, 1986; Karlowicz MG, Adelman RD: Acute renal failure in the neonate, Clin Perinatol 19:139-158, 1992; and Mathew OP, Jones AS, James E, et al: Neonatal renal failure: usefulness of diagnostic indices, Pediatrics 65: 57-60, 1980. UOP, Urine output. *Mean ± SD. †Fractional excretion of sodium (FENa) = (Urine [Na+]/Serum [Na+])/(Urine [Cr]/ Serum [Cr]) × 100. ‡Renal failure index (RFI) = Urine [Na+]/(Urine [Cr]/Serum [Cr]).

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

and water is dependent on intact tubular function. Disturbances of tubular function can occur with diuretic use, ischemic or nephrotoxic tubular injury, or primary tubular diseases. Laboratory markers of prerenal azotemia include low urinary sodium excretion, low fractional excretion of sodium, low renal failure index, and high BUN:SCr ratio. These laboratory studies have important limitations in premature infants. Unfortunately, laboratory tests to determine whether elevated SCr is from prerenal azotemia versus intrinsic AKI are insensitive and nonspecific in premature infants because of underdeveloped tubular function. Normal fractional sodium excretion in preterm infants born before 32 weeks’ gestation is usually higher than 3% (Ellis and Arnold, 1982). In addition, because of the developmentally regulated limitation of their concentrating capacity and the effects of low protein intake and urea excretion on urine osmolality, the urine-to-plasma creatinine ratio instead of the urine-to-plasma osmolality ratio should be used in newborns to evaluate their renal tubular reabsorptive capacity (Feld et al, 1986). If suspicion of renal hypoperfusion is high, an appropriate fluid challenge with 10 to 20 mL/kg of isotonic fluids (usually normal saline) over 30 minutes should be given. Close observation to vital signs and urine output may serve to delineate the presence of intravascular hypoperfusion. Several boluses may be necessary with careful prescription of fluid volume for the next 24 hours. Care to avoid fluid challenges is advised in those with suspected urinary outlet obstruction or congestive heart failure. A second major goal of AKI evaluation is to detect anatomic causes of AKI, if present. A renal and bladder ultrasound examination should be performed without delay if an obstructive process is suspected and to detect congenital renal abnormalities if present. If hematuria, hypertension, or both are present, the possibility of renal vascular disease should also be considered. Doppler ultrasound examination of renal vessels can be performed if renal vascular thrombosis is suggested.

MANAGEMENT The approach using fluid boluses (if appropriate) as part of the evaluation of prerenal azotemia also serves as the initial management of this condition. If obstruction of the urinary outflow is discovered, then interventions to eliminate the obstruction should be undertaken followed by plans for surgical correction. Polyuria with electrolyte losses can occur after relief of the obstruction; therefore close monitoring of serum electrolytes, especially bicarbonate, and appropriate replacement of these losses are necessary. Besides these specific management options, there are currently no specific medical therapies to treat AKI. To maximize the chance for survival, the clinician must support the cardiorespiratory system, maintain maximal nutrition, balance homeostasis, and manage the consequences of AKI. Dialysis can provide renal suppport to achieve goaloriented therapies. Despite numerous promising therapies in animal studies and many clinical pharmacologic trials, no approved therapies are available to treat AKI in neonates, children, or adults. Only supportive care of the patient (with medications, renal support therapies, or both) is currently

1211

available for the care of the infant with AKI (Walters et al, 2009). Several therapies are commonly used in patients with AKI; however, few data are available to support the use of low-dose dopamine, fenoldopam (a selective dopamine-1 receptor agonist), or diuretics for the treatment or prevention of AKI. Dopamine can increase renal perfusion in the sick preterm and term infant with prerenal azotemia caused by hypoxemia, acidosis, or indomethacin administration (Seri, 1995; Seri et al, 1998, 2002). Although low dose dopamine increases renal perfusion, well-powered randomized controlled studies in adults with AKI have reached the same conclusion (Bellomo et al, 2000; Friedrich et al, 2005; Hoste et al, 2006; Marik, 2002). Compared with placebo, low-dose dopamine does not improve survival, shorten hospital stay, or limit dialysis use. These studies have not been performed in children or neonates. Diuretics are commonly used to induce diuresis in critically ill neonates; however, no studies in neonates, children, or adults have shown that diuretics are effective in preventing AKI or improving outcomes once AKI occurs (Bellomo et al, 2000). If loop diuretics are to be used in neonates, continuous doses of furosemide may be superior to larger intermittent doses. In neonates with cardiac surgery, continuous doses were shown to be as effective despite smaller total quantities of medications. The authors conclude that those with continuous dosing may have less risk for nephrotoxicity or ototoxicity than occurs with large intermittent doses of this drug (Luciani et al, 1997). The potential toxicity of long-term and aggressive furosemide therapy—including ototoxicity, interstitial nephritis, osteopenia, nephrocalcinosis, hypotension, and persistence of patent ductus arteriosus—should be considered, especially in the preterm newborn (Karlowicz and Adelman, 1992). The mannitol test (used to test whether a patient has renal hypoperfusion) is contraindicated in newborns with a predisposition for intraventricular hemorrhage or periventricular leukomalacia because of a druginduced, sudden increase in serum osmolality. Fenoldopam is a selective dopamine-1 receptor agonist whose effects include vasodilation of renal and splanchnic vasculature, increased renal blood flow, and increased GFR. Fenoldopam is approved to treat severe hypertension in adults, but is not clinically approved for the treatment of AKI. Nonetheless, its use in neonates with AKI has been explored in several single-center analyses. Although two separate retrospective single-center analyses (Moffett et al, 2008; Yoder and Yoder, 2009) found increased urine output in a select group of neonates with oliguria, Ricci et al (2008) performed a prospective controlled trial of lowdose fenoldopam (0.1 μg/kg/min) in infants undergoing cardiac surgery with cardiopulmonary bypass. Compared with placebo, low-dose fenoldopam did not show beneficial effects on AKI incidence, fluid balance control, time to sternal closure, time to extubation, or time to intensive care discharge. Because low-dose dopamine, diuretics, and fenoldopam are unlikely to positively affect the outcomes of infants with AKI, efforts should be maximized to support the infant’s cardiorespiratory system and to manage and prevent the ill effects of AKI. If systemic hypotension develops despite adequate volume administration, early initiation of blood pressure

1212

PART XVII  Renal and Genitourinary Systems

support often establishes appropriate renal perfusion (Seri et al, 1993, 1998). In cases of pressor-inotrope–resistant hypotension and shock, a brief course of low-dose hydrocortisone has been demonstrated to be effective in restoring systemic perfusion and renal function in preterm neonates (Seri, 2001). Other management goals include maintaining blood oxygen content, providing blood products for specific indices, limiting severe acidosis, and maintaining normal serum albuminemia (at least 2.0 mg/dL, but preferably 2.5 mg/dL). Hypertension is common in neonates with AKI. It can be caused by increased renin release in malformed or damaged kidneys or secondary to increased intravascular volume from a lack of free water clearance. If hypertension is due to fluid overload, inducing free water clearance with diuretics or fluid removal with dialysis will address its cause. Calcium-channel blockers work by selectively causing vasodilatation of the venous system. Short-acting calcium-channel blockers (e.g., isradipine) are reliable, have a quick onset of response, and are well tolerated. Longacting calcium-channel blockers (e.g., amlodipine) take longer to take effect, but they provide less lability with longer dosing intervals. β-Blockers (propranolol or labetolol) are also commonly used to treat hypertension in neonates. Use of ACE-I in children with ischemic AKI should be avoided, because it can produce further renal hypoperfusion and can alter intrarenal hemodynamics in an already injured kidney. See Chapter 88 for more information on neonatal hypertension. Managing fluids in the critically ill neonate with AKI can be difficult. These infants may require large volumes of fluid to maintain adequate nutrition and hematologic indices and to provide appropriate medications. However, these fluids can be detrimental in a child with oliguria or anuria, because they can cause congestive heart failure, chest wall edema, and pulmonary failure. Therefore, once adequate intravascular volume has been restored, prevent severe fluid overload (by limiting crystalloid infusions) and maximize nutritional supplements concentration. Severe fluid restriction limiting intake to insensible and gastrointestinal and renal losses is sometimes required, but at a heavy price (inadequate nutrition). Decisions regarding placement of dialysis access should be made early in the course of AKI before severe fluid overload has occurred, because once severe fluid overload occurs, placement of a peritoneal dialysis or a hemodialysis catheter can be significantly more difficult, as is support of the infant with severe pulmonary edema. Electrolyte abnormalities can vary depending on the cause of AKI. For example, aminoglycoside toxicity is commonly nonoliguric with ongoing potassium and magnesium losses. Alternatively, ischemic AKI causes oliguria– anuria, hyponatremia, hyperkalemia, hyperphosphatemia, and hypocalcemia. Management of electrolyte disorders can usually be managed by attention to electrolyte intake during the initial course of AKI with frequent evalution and specific therapies. Most cases of hyponatremia are due to water overload and less commonly due to low total body sodium content. Attention to fluid status is critical to determine the cause and proper therapy of hyponatremia. In cases of non­ symptomatic hypervolemic hyponatremia (serum sodium

concentrations usually between 120 and 130 mEq/L), restriction of free water intake is recommended. If hyponatremia at this level results in clinical signs and symptoms (e.g., lethargy, seizures) or serum sodium concentration falls to less than 120 mEq/L, use of 3% sodium chloride over 2 hours according to the following formula should be considered: ( ) Na + Required (mEq) = [Na + ] Desired − [Na + ] Actual × Body weight (kg) × 0.7

Possible complications of hypertonic saline administration include congestive heart failure, pulmonary edema, hypertension, intraventricular hemorrhage, and periventricular leukomalacia. Care should be taken not to increase serum sodium concentration more rapidly than 0.5 mEq/h. Severe hyperkalemia is a life-threatening medical emergency. Hyperkalemia that is unresponsive to medical management is one of the most common indications for peritoneal and hemodialysis in the newborn (Coulthard and Vernon, 1995; Karlowicz and Adelman, 1992). Signs of progressive hyperkalemia on the electrocardiogram, in order of severity, consist of tall peaked T waves, heart block with widened QRS complexes, U wave formation, the development of sine waves, and finally cardiac arrest. Measures to remove potassium from the body include oral or rectal sodium polystyrene (Kayexalate), loop diuretics to enhance potassium excretion (if not anuric), and dialysis. Several methods to move potassium from the extracellular to the intracellular compartment are available, including albuterol inhalation, sodium bicarbonate, and insulin plus glucose. Adequate ionized calcium levels for cardioprotection should be sought in the context of hyperkalemia (Table 85-3). Vemgal and Olhson (2007) performed a metaanalysis of studies on the management of hyperkalemia in premature infants. Given the limited data available, no firm clinical practice recommendations on which treatment modality was best to treat infants with hyperkalemia are available, except that insulin plus glucose may be better in premature infants. Hyperphosphatemia is common in AKI and should be treated with low phosphorus intake. Breastmilk and Similac 60/40 both contain low phosphorous and low potassium compared with other neonatal infant formulas. Significant elevations in serum phosphate represent a risk of development of extraskeletal calcifications of the heart, blood vessels, and kidneys in the newborn, especially when the calcium-phosphorus product exceeds 70 (Sell et al, 1987). Calcium carbonate may be used as a phosphate-binding agent in infants whose phosphorous intake exceeds excretion. Although rarely an indication for dialysis without fluid overload or hyperkalemia, severe hyperphosphatemia is best treated with dialysis. The incidence of hypocalcemia is low in neonates with severe and prolonged AKI, especially in those who develop an inability to convert 25-hydroxy–vitamin D to 1,25-hydroxy–vitamin D. Ionized calcium should be measured when low total calcium levels and concomitant hypoalbuminemia are encountered, because the latter can affect total calcium levels. If ionized calcium is decreased and the newborn is symptomatic, 100 to 200 mg/kg of calcium gluconate should be infused over 10 to 20 minutes and repeated every 4 to 8 hours as necessary. If hypocalcemia

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

1213

TABLE 85-3  Medical Management of Hyperkalemia in the Newborn

Drug

Dose

Onset of Action

Duration of Action

Calcium gluconate (10%)

0.5-1 mL/kg (IV over 10 min)

1-5 min

15-60 min

Sodium bicarbonate (3.75% solution)

1-2 mEq/kg (IV over 10 min)

5-10 min

2-6 h

Insulin

1 IU/5 g glucose (IV bolus or continuous infusion)

15-30 min

4-6 h

Glucose

≤14 mg/kg/min (IV bolus or continuous infusion)

15-30 min

4-6 h

Furosemide

1 mg/kg dose or as continuous infusion

5-10 min

2-3 h

Sodium polystyrene sulfonate

1 g/kg dose every 6 h as needed (orally/rectally)

1-2 h*

4-6 h

Dialysis

Nephrology will prescribe dialysis treatment

Immediate

Duration of therapy

*Onset of action may take up to 6 hours, and the drug may be ineffective in preterm infants born at less than 29 weeks’ gestation. IV, Intravenous.

is severe, oral or intravenous calcitriol can be administered to increase intestinal reabsorption of calcium. Normal acid–base homeostasis depends on the kidneys’ ability to reabsorb bicarbonate; therefore infants with AKI commonly have a non–anion gap metabolic acidosis. Replacement with bicarbonate or acetate as a base is indicated in infants with AKI to avoid or treat metabolic acidosis. In infants with severe respiratory failure, large doses of bicarbonate should be avoided because they can culminate in increased carbon dioxide retention. Metabolic acidosis should be treated aggressively in infants with severe pulmonary hypertension, because an acidic environment can worsen this condition. Nutritional goals in infants with AKI are similar to those of infants without AKI. Commonly parenteral nutrition, feeds, or both will need to be concentrated to avoid excessive fluid gains. If nutritional goals are unable to be achieved because of oliguria or ongoing fluid overload, the potential risks of dialysis therapy versus the potential risks associated with inadequate calorie and protein administration should be discussed with the parents. If a neonate is receiving continuous peritoneal dialysis or hemodialysis, an additional 1 g/kg/day of protein is needed to supplement the protein losses that occur with these forms of dialysis (Zappitelli et al, 2008, 2009). In a neonate with AKI, careful assessment of medication dosing is imperative. Because many drugs are excreted in the urine, impaired metabolism or clearance from the kidneys can cause drug accumulation and adverse side effects. In infants receiving dialysis, pharmacokinetic properties of drugs (e.g., volume of distribution, protein binding, size, charge), dialysis modality (peritoneal ­dialysis ­versus hemodialysis), interval of dialysis (intermittent versus continuous) will affect drug availability (Churchwell and Mueller, 2009). Consultation with pharmacists and a nephrologist familiar with drug dosing in renal failure is invaluable.

Dialysis Decisions to initiate dialysis (especially in infants with severe congenital malformations of the kidney and urinary tract) is a complex decision that requires a multidisciplinary approach to guide the family as they consider very difficult decisions. Access placement and some technical challenges make infant dialysis more difficult than in older children, but this therapy is feasible in experienced

programs with dedicated pediatric neonatologists, dialysis nurses, and surgeons.

Indications for Dialysis Initiation Absolute indications to initiate dialysis include: severe electrolyte abnormalities that are not correctable with medical interventions, life-threatening intoxication by medications that can be cleared with dialysis, inborn errors of metabolism, fluid overload, inability to provide adequate nutritional requirements, and uremia. If renal dysfunction, fluid overload, or both occur, then discussions about dialysis initiation should occur early in the disease process, because prolonged fluid overload or uremia can create worse pulmonary edema and cardiopulmonary instability and make placement of access for dialysis difficult. The timing of dialysis initiation in infants with AKI is controversial. Several observational studies show a clear advantage in adults receiving dialysis early versus late (Liu et al, 2006; Ronco et al, 1986). In addition, multicenter data show that the degree of fluid overload at the initiation of dialysis is an independent risk factor for survival in critically ill children (Gillespie et al, 2004; Goldstein et al, 2001; Symons et al, 2007) and adults (Gibney et al, 2008; Mehta, 2009). Similar studies in neonates need to be performed. Over the last decade, advocates for early initiation of renal support argue that critically ill patients benefit from early dialysis. Metabolic control is gained faster because of earlier removal of excess fluid, and provision of renal support allows maximal nutrition without progressive fluid overload. Because technical access and machine advances have made neonatal dialysis safer and technically possible, early initiation of dialysis can improve outcomes in critically ill neonates with AKI. Further study is needed before recommendations on the timing of dialysis can be made.

Access The limiting factor in performing dialysis in the smallest of babies is access to the peritoneal or vascular space for dialysis. Peritoneal dialysis access can be performed with either a straight uncuffed catheter or a curved and tunneled cuffed catheter. The advantage of the straight uncuffed catheter is that it can be placed at the bedside and can be used soon after insertion. However, these catheters are more likely to become infected or leak fluids around the insertion site. The ideal peritoneal dialy­sis catheter is one with two subcutaneous cuffs and

1214

PART XVII  Renal and Genitourinary Systems

a downward-facing exit site away from the diaper area and away from a gastrostomy tube (Auron et al, 2007). As with all pediatric surgical procedures, the exact catheter, timing, and location of catheter insertion needs to be tailored to the individual patient (Shaheen et al, 2007). If a hernia is present, it should be repaired at the time of catheter insertion. Vascular access for hemodialysis requires a large (at least 7F, but preferably 8F) double-lumen catheter that can be placed in the femoral or internal jugular vein. Double-lumen catheters that are smaller than 7F have a much higher chance of developing problems during the dialysis procedure (Symons et al, 2007). Catheterization of the umbilical vessels does not provide sufficient blood flow necessary for conducting hemodialysis. If 7F catheter cannot be placed, two 5F catheters in different sites can be life-saving. If the need for dialysis is likely to last more than 1 week, a cuffed catheter is preferred to decrease the likelihood of infection. The length of the catheter should be chosen such that the tip of the catheter resides in the superior vena cava–right atrium (for internal jugular catheters) and the inferior vena cava for femoral catheters. Unless no other choice is available, use of the subclavian artery should be avoided in infants who are likely to require longterm renal replacement therapy, because in the future the forearm fistula of the ipsilateral arm can fail with mild stenosis of the subclavian vein.

Peritoneal Dialysis Once a peritoneal dialysis access catheter is placed and the decision to start dialysis has occurred, small-volume continuous cycles (10 mL/kg) are performed. Fluid is filled with dialysate solution, left in the peritoneal cavity to dwell, and drained. Continuous cycles are performed, with each cycle lasting approximately 1 hour. The dextrose concentration in the fluid will determine the amount of net water losses (ultrafiltration). Complications associated with peritoneal dialysis include peritonitis, leakage around the catheter exit site rendering the dialysis virtually impossible, tunnel infection, catheter malfunction, and obstruction by omentum (Coulthard and Vernon, 1995). Fluid leakage into other compartments (including the chest in patients without an intact diaphragm) can occur and if suspected, the fluid composition will reveal high glucose levels if a leak is present. Absolute or relative contraindications to peritoneal dialysis include necrotizing enterocolitis, abdominal wall defects, and the presence of an intraabdominal foreign body, such as a ventriculoperitoneal shunt or diaphragmatic patch.

Hemodialysis Once reliable access to the vascular space is achieved, the hemodialysis procedure can be performed in neonates. The two types of hemodialysis, intermittent hemodialysis and continuous renal replacement therapy (CRRT), differ mainly on the duration of the procedure. Intermittent hemodialysis is significantly more efficient than CRRT. Because much larger volumes of dialysis are used, the blood flow becomes the limiting factor in the amount of clearance that can be achieved. Even with the

smallest dialyzers and neonatal tubing, most infants need blood priming of the extracorporeal circuit for therapy. Skilled pediatric hemodialysis nurses are required at the bedside during the entire procedure, which typically lasts 3 to 4 hours. Achieving adequate fluid removal is sometimes difficult, especially in hemodynamically unstable infants. This approach usually requires systemic heparinization, with activated clotting time usually maintained at 180 to 200 seconds, rendering this approach risky in preterm newborns and others at high risk for intracranial bleeding. Ronco et al (1986) described the use of CRRT in a critically ill newborn. Before the recent roller pump technology, CRRT used the patient’s arterial blood pressure as a pump for the dialysis machine. The advent of newer roller pump technology improved the accuracy of small flows as low as 10 mL/min and provided the option of using the machine to pump blood via a double lumen catheter placed in a major vein. The main advantage of a continuous modality is that lower blood flow and fluid removal rates can be used to accomplish the desired ultrafiltration and clearance goals. Anticoagulation with CRRT is achieved with either systemic heparin or citrate regional anticoagulation. The advantage of regional citrate anticoagulation is that the patient is not anticoagulated; however, this method of anticoagulation has the added risk of hypocalcemia caused from citrate excess (especially in those infants with impaired liver metabolism) and metabolic alkalosis (Tolwani and Willie, 2009). Outcome data in neonates who require CRRT are scarce. Symons et al (2003) reported a survival rate of 32 in 85 (37.6%) in neonates who received CRRT in five large children’s hospitals in the United States. Between 2001 and 2007, the prospective pediatric CRRT group, a multicenter registry of 14 pediatric CRRT programs, reported outcomes for infants in whom CRRT was initiated before 1 month of age (Symons et al, 2007). Approximately 8% (35 neonates) in the registry were dialyzed in the 1st month of life. In this group, the median age was 8 days; the median weight was 3.2 kg, with the smallest infant weighing 1.3 kg. Of the 35 infants in the registry, 24 received dialysis for fluid overload, electrolyte imbalance, or both, and 11 of 35 received dialysis for inborn errors of metabolism. Overall survival was 43%. Infants receiving dialysis for inborn errors had a better survival rate (73%) compared with others (30%). Several technical issues specific to infants arise when using CRRT for dialysis. The extracorporeal volume can incorporate greater than 50% of the infant’s blood volume. For example, a 2.3-kg infant has a blood volume of approximately 184 mL; the Prisma circuit requires 90 mL, which is approximately 50% of the infant’s blood volume. Priming the blood circuit with blood will effectively limit this problem. In addition, because standard protocols on citrate regional anticoagulation anticipate normal liver metabolism of citrate, caution must be exercised when performing dialysis in premature infants or newborns with multiple organ failure who may have impaired liver function. Risk for bradykinin reaction that can occur at the initiation of CRRT with AN69 dialyzer membranes can be reduced using several techniques (Brophy et al, 2001; Hackbarth et al, 2005).

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

ACUTE KIDNEY INJURY AS A CAUSE OF LONG-TERM CHRONIC KIDNEY DISEASE Total GFR is determined by the filtration rate of a single nephron and the number of nephrons present. When the number of nephrons is diminished, single nephron GFR increases as the kidney works to compensate for low nephron numbers. This compensatory hypertrophy causes glomeruli to function under increased intracapillary hydraulic pressure, which over time causes damage to capillary walls. This abnormal process leads to progressive glomerulosclerosis, proteinuria, hypertension, and CKD (Brenner et al, 1996). The hyperfiltration hypothesis has been applied and confirmed in autopsy data of patients with hypertension (Keller et al, 2003; Ohishi et al, 1995) and has been reported at length in infants with intrauterine growth retardation (Barker and Osmond, 1988; Barker et al, 1989; Manalich et al, 2000; Wadsworth et al, 1985; White et al, 2009). A systematic review and metaanalysis in 2009 concluded that low-birthweight infants (5.5  lb [2.5  kg]) were 70% more likely to develop CKD later in life compared to individuals with normal birthweight (White et al, 2009). Premature infants (even those born appropriate for gestational age) are born with low nephron numbers. Using computer-assisted morphometry, Rodriguez et al (2004) showed that premature infants have lower numbers of nephrons compared with term infants. Premature infants who had long survival (to at least 36 weeks after conception) had nephron numbers similar to those in premature infants with short survival, suggesting that the extrauterine environment does not allow for proper neoglomerulogenesis. In addition, preterm infants with AKI had fewer nephrons than did similar infants without AKI (Rodriguez et al, 2004). Animal and epidemiology data suggest that AKI leads to CKD. As discussed in the section on ischemic AKI, tubular and vascular endothelial cellular damage occurs with prolonged hypoperfusion. Animal models suggest that although tubular recovery occurs, damage to vascular endothelial cells remains and leads to interstitial fibrosis and progressive kidney dysfunction (Basile et al, 2001). Studies on children with AKI show that more than 50% of those with AKI have at least one sign of CKD 3 to 5 years after the inciting event. Large adult studies suggest that after AKI, rates of CKD (low GFR) are 5% to 10%, with approximately 3% to 5% developing end-stage renal disease (ESRD). The exact prevalence of CKD after neonatal AKI is not known. Stapleton (1987) reviewed the published singlecenter data and reported a 40% to 88% prevalence of long-term CKD after oliguric renal failure. Since then, other retrospective, small, single-center retrospective studies describe similar tubular and glomerular dysfunction and hypertension in survivors of neonatal AKI (Abitbol et al, 2003; Chevalier et al, 1984; Polito et al, 1998). Data on outcomes of premature infants after AKI are scarce. Rodriguez et al (2004) performed a cross-sectional study on premature infants (birthweight <1000  g) during childhood and found that estimated GFR and tubular function were lower than in similar term children. Despite the limitations of these single-center studies,

1215

these data suggest that prematurity, intrauterine growth retardation and AKI lead to a lower number of nephrons, endothelium dysfunction, or both, and an increased risk of long-term renal dysfunction. To further delineate the likelihood and extent of AKI causing CKD, a prospective study is greatly needed to provide guidelines for longterm follow-up. Future studies on interventions (e.g., angiotensin-converting enzyme inhibitors) to decrease the rate of CKD progression in this growing population should be explored.

CHRONIC KIDNEY DISEASE Neonatal CKD is diagnosed when sustained derangements of glomerular filtration or tubular function occur with minimal to no resolution over time. In many cases CKD follows AKI, and in others the acute phase of the renal compromise has not been detected or has occurred in utero often as a result of anatomic abnormalities (e.g., hypoplasia, dysplasia, malformations), and the diagnosis of CKD is established without documented evidence of preexisting AKI. According to guidelines published by the Kidney Disease Outcomes Quality Initiative, CKD is present if there is evidence of kidney damage for more than 3 months, as defined by structural or functional abnormalities, with or without decreased GFR, or a GFR less than 60 mL/min per 1.73 m2 for more than 3 months in children older than 2 years with or without kidney damage. GFR increases with maturation from infancy and approaches the adult mean value by 2 years of age. Thus it is important to note that the ranges of GFR that define the stages of CKD apply only to children 2 years and older. The GFR ranges defining CKD do not apply to infants, because they will normally have a lower GFR even when corrected for body surface area. ESRD, the point at which dialysis or kidney transplantation is necessary to ameliorate the physiologic complications of uremia owing to kidney failure, represents the most severe stage of CKD. The pathophysiologic mechanisms leading to the progression of AKI to CKD and ESRD are discussed earlier in this chapter.

EPIDEMIOLOGY Reports of the regional incidence of neonatal ESRD differ greatly, and the worldwide incidence is unknown. In one study, the estimated incidence of neonatal ESRD in the United States and Canada was 0.32 in 100,000 live births (Carey et al, 2007). Rees (2008) reported that six new infants per 1 million population initiate dialysis per year compared to three per 1 million in the United Kingdom. In a German study, Wedekin et al (2008) estimated the incidence of CKD in infants to be 1 in 10,000 live births and found a male-to-female ratio of 2.8:1. They also found that 53% of these infants were premature, a figure significantly higher than in the total infant population of Germany. The most common causes of renal failure in most studies of neonates are renal dysplasia or obstructive uropathy (Carey et al, 2007; Ledermann et al, 1999; Rees, 2008; Shroff et al, 2003; Wedekin et  al, 2008)—specifically, posterior urethral valves (Ledermann et al, 1999).

1216

PART XVII  Renal and Genitourinary Systems

SEQUELAE AND TREATMENT Anemia Anemia is a frequent complication of CKD in infants and children, and there is evidence to suggest that it is an important predictor of patient morbidity and mortality. The anemia of CKD is associated with a number of physiologic abnormalities, including decreased tissue oxygen delivery, increased cardiac output, cardiac enlargement, ventricular hypertrophy, congestive heart failure, and impaired immune responsiveness (Gafter et al, 1994; Mitsnefes et al, 2000). There is also evidence to suggest that the presence of anemia is associated with an increased risk of hospitalization in children with CKD (Staples et al, 2009). The prevalence of anemia increases with worsening stages of CKD. In a study of the North American Pediatric Renal Trials and Collaborative Studies (NAPRTCS), Staples et al (2009) found that the prevalence of anemia increased from 18.5% in CKD stage II to 68% in CKD stage V (Staples et al, 2009). Guidelines for the evaluation and treatment of anemia in the pediatric patient with CKD have been published as the KDOQI Clinical Practice Guidelines and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease by the National Kidney Foundation (National Kidney Foundation, 2006, 2007). According to the guidelines, anemia is defined as a hemoglobin (Hgb) concentration less than the 5th percentile of normal for age and sex (Table 85-4). The normative values used to define anemia in children older than 1 year are taken from the third National Health and Nutrition Examination Survey (Astor et al, 2002) database, whereas the norms for infants younger than 1 year are derived from other reference sources (Nathan and Orkin, 2003). The pathophysiology of anemia in infants and young children with CKD is primarily the result of a decrease in renal production of erythropoietin, iron deficiency, or both (Koshy and Geary, 2008). Other potential contributing factors include a shortened red blood cell life span, secondary hyperparathyroidism, hypothyroidism, folate and vitamin B12 deficiency, chronic inflammation and hemoglobinopathies. The etiology of absolute iron deficiency is multifactorial and can be related to poor intake, gastrointestinal blood loss, and repeated phlebotomies for laboratory tests. Functional iron deficiency, defined as occurring when there is a greater need for iron to support Hgb synthesis than can be released from iron stores, may in part be related to the presence of elevated levels of the liver-derived peptide hepcidin—a subject about which further study is needed in the pediatric CKD population (Zaritsky et al, 2009). The KDOQI guidelines recommend that the initial workup for anemia in children with CKD should include red blood cell indices, reticulocyte count, white blood cell count with differential and platelet count, and iron parameters (serum iron, total iron binding capacity, and serum ferritin) (National Kidney Foundation, 2006, 2007). Erythropoiesis-stimulating agents (ESAs) such as erythropoietin-alfa (EPO), along with iron supplements, are the key elements of anemia management in CKD. Whereas the average dose of EPO for children with CKD is 150 to 200 units/kg/week given by the subcutaneous route, younger children (<1 year) often require doses as high as

TABLE 85-4  Hemoglobin Levels (g/dL) in Children between Birth and 24 Months Old for Initiation of Anemia Workup

Age

Mean Hemoglobin

–2 SD

Term (cord blood)

16.5*

13.5†

1-3 d

18.5

14.5

1 wk

17.5

13.5

2 wk

16.5

12.5

1 mo

14.0

10.0

2 mo

11.5

9.0

3-6 mo

11.5

9.5

6-24 mo

12.0

10.5

From National Kidney Foundation: KDOQI clinical practice guidelines and clinical practice recommendations for anemia in chronic kidney disease, Am J Kidney Dis 47(Suppl 3):S88, 2006. *Data taken from normal reference values. †Values 2 standard deviations (SD) below the mean are equivalent to <2.5th percentile.

350 units/kg/week. The difference is possibly related to an increased presence of nonhematopoietic binding sites for EPO in younger children that lead to increased clearance (NAPRTCS, 2008; Port et al, 2004). Children receiving peritoneal dialysis (PD) typically require less EPO than patients receiving hemodialysis (HD) as a result of the blood loss that occurs with HD. A recent report on the use of darbepoetin-alfa, an analogue of erythropoietin with a longer half-life, has been conducted in infants with CKD and revealed that the therapy was effective when dosing regimens were individualized (12). Iron therapy typically consists of the provision of oral elemental iron in doses ranging from 2 to 3 mg/kg/day up to 6 mg/kg/day in two to three divided doses (National Kidney Foundation, 2006). Iron should be taken 2 hours before or 1 hour after all calcium containing phosphate binders to maximize gastrointestinal absorption. Levels of serum ferritin greater than 100 ng/mL and transferrin saturation greater than 20% are believed to reflect adequate iron stores in patients with CKD (National Kidney Foundation, 2006). In the absence of definitive evidence in pediatrics to ­support the association of benefit or harm to any given level of Hgb for an individual child, the target Hgb is 10 to 12 g/dL for patients receiving ESAs and iron therapy (National Kidney Foundation, 2006, 2007; Staples et al, 2009). In turn, the rate of rise of the Hgb level should be no more than 1 to 2 g/dL per month. At the initiation of ESA therapy, monitoring of the Hgb level should occur every 1 to 2 weeks; the frequency of monitoring can be decreased once the target Hgb level has been achieved and the patient is receiving a stable dose of ESA.

Acid-Base and Electrolytes Fluid and electrolyte requirements of individual children with CKD vary according to their primary kidney disease and the degree of residual kidney function. Infants and children normally have a relatively larger endogenous hydrogen ion load (2 to 3 mEq/kg) than do adults, resulting in metabolic acidosis as a common manifestation of CKD in

1217

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

TABLE 85-5  Recommended Parameters and Frequency of Nutritional Assessment for Children With CKD Stages 2 to 5 and 5D

Minimum Interval (mo) Age 0 to <1 y Measure

CKD 2-3

CKD 4-5

Age 1-3 yr CKD 5D

CKD 2-3

CKD 4-5

CKD 5D

Dietary intake

0.5-3

0.5-3

0.5-2

1-3

1-3

1-3

Height or length-for-age percentile or SDS

0.5-1.5

0.5-1.5

0.5-1

1-3

1-2

1

Height or length velocity-for-age percentile or SDS

0.5-2

0.5-2

0.5-1

1-6

1-3

1-2

Estimated dry weight and weight-for-age percentile or SDS

0.5-1.5

0.5-1.5

0.25-1

1-3

1-2

0.5-1

BMI-for-height-age percentile or SDS

0.5-1.5

0.5-1.5

0.5-1

1-3

1-2

1

Head circumference–for–age percentile or SDS

0.5-1.5

0.5-1.5

0.5-1

1-3

1-2

1-2

Modified from National Kidney Foundation: KDOQI clinical practice guideline for nutrition in children with CKD: 2008 update, Am J Kidney Dis 53(Suppl 2):S16, 2009. BMI, Body mass index.

children and an important negative influence on growth. Metabolic acidosis leads to both endogenous growth hormone and recombinant growth hormone resistance. Based on the experience of successfully enhancing the growth of infants and children with isolated renal tubular acidosis with alkalai therapy, it is recommended that children with CKD be treated to achieve a serum bicarbonate level of at least 22 mmol/L (National Kidney Foundation, 2009).

Nutrition Protein energy malnutrition is a common problem in patients with CKD and is one of the major contributors to poor growth in the first few years of life. The origin of malnutrition in children with CKD is multifactorial; however, an inadequate voluntary intake is considered a major contributing factor, especially in infants. Nausea and vomiting are common in infants and children with CKD, with delayed gastric emptying and gastroesophageal reflux being detected in as many as 75% of patients with these problems (Ruely et al, 1989). In view of the importance of the nutritional status to the outcome of the infant and young child with CKD, frequent monitoring of the patient is mandatory. An age-related schema for parameters and frequency of nutritional assessment for patients with CKD has recently been published (Table 85-5) (National Kidney Foundation, 2009). The World Health Organization Growth Standards of length-for-age, weight-forage, weight-for-length, body mass index–­for-age and head circumference–for-age should be used as the reference for children from birth to 2 years (World Health Organization, 2006). Whereas nutritional intervention is indicated in children with CKD for findings that include an impaired ability to ingest or tolerate oral feedings, a body mass index value less than the 5th percentile of height-for-age, an acute weight loss of 10% or more or a length/height ratio more than 2 standard deviations (SD) less than the mean, neonates with CKD should be considered to be at nutritional risk if they are preterm or have any of the following: ll ll ll

Low birth weight (<2500 g) A birth weight z score less than –2 SD for gestational age Polyuria and an inability to concentrate urine

TABLE 85-6  Equations to Estimate Energy Requirements for Children at Healthy Weights

Age (mo)

EER (kcal/d) =  Total energy expenditure + Energy deposition

0-3

EER = 89 × Weight (kg) – 100 + 175

4-6

EER = 89 × Weight (kg) – 100 + 56

7-12

EER = 89 × Weight (kg) – 100 + 22

13-35

EER = 89 × Weight (kg) – 100 + 20

Modified from National Kidney Foundation: KDOQI clinical practice guideline for nutrition in children with CKD: 2008 update, Am J Kidney Dis 53(Suppl 2):S36, 2009. EER, Estimated energy requirement.

In children with CKD, the spontaneous energy intake decreases with deteriorating kidney function. Energy requirements should, however, be considered to be 100% of the estimated energy requirement for chronologic age (Table 85-6) (Food and Nutrition Board, 2002; Ruely et al, 1989). Because energy intake is the principal determinate of growth during infancy, malnutrition has the most marked negative effect on the growth of children with congenital disorders leading to CKD (Betts et al, 1977). Supplemental nutritional support is indicated when the voluntary intake by the child fails to meet energy requirements and the child is not achieving expected rates of weight gain or growth for age. In infants requiring fluid restriction because of their impaired kidney function, oral intake of an energy-dense diet with a milk formula that has a caloric density greater than 20 kcal/oz, and the appropriate phosphorus content for CKD stage is preferred. If poor appetite or vomiting preclude an adequate oral intake, tube feedings (e.g., nasogastric, gastrostomy, gastrojejunostomy) should be considered and provided by either bolus or continuous infusion. The development of repeated emesis in children fed per nasogastric tube has prompted the use of gastrostomy as the preferred route of therapy (Warady et al, 1996). In infants younger than 1 year, an initial infusion rate of 10 to 20 mL/h or 1 to 2 mL/ kg/h is generally well tolerated, to be followed by a daily increase of 5 to 10 mL per 8 hours or 1 mL/kg/h toward achieving the treatment goal. It is imperative that tube-fed

1218

PART XVII  Renal and Genitourinary Systems

TABLE 85-7  Recommended Dietary Protein Intake in Children with CKD Stages 3 to 5 and 5D

Dietary Reference Intake (DRI)

Age

DRI (g/kg/d)

Recommended for CKD Stage 3 (g/kg/d) (100%-140% DRI)

Recommended for CKD Stages 4-5 (g/kg/d) (100%-120% DRI)

Recommended for PD (g/kg/d)†

Recommended for HD (g/kg/d)*

0-6 mo

1.5

1.5-2.1

1.5-1.8

1.6

1.8

7-12 mo

1.2

1.2-1.7

1.2-1.5

1.3

1.5

1-3 y

1.05

1.05-1.5

1.05-1.25

1.15

1.3

Modified from National Kidney Foundation: KDOQI clinical practice guideline for nutrition in children with CKD: 2008 update, Am J Kidney Dis 53(Suppl 2):S49, 2009. *DRI + 0.1 g/kg/d to compensate for dialytic losses. †DRI + 0.15 – 0.3 g/kg/d depending on patient age to compensate for peritoneal losses.

TABLE 85-8  Dietary Reference Intake for Healthy Children for Sodium, Chloride, and Potassium

Sodium (mg/d)

Chloride (mg/d)

Potassium (mg/d)

Age

AI

Upper Limit

AI

Upper Limit

0-6 mo

120

ND

180

ND

400

AI

Upper Limit ND

7-12 mo

370

ND

570

ND

700

ND

1-3 y

1000

1500

1500

2300

3000

ND

Modified from National Kidney Foundation: KDOQI clinical practice guideline for nutrition in children with CKD: 2008 update, Am J Kidney Dis 53(Suppl 2):S49, 2009. AI, Adequate intake; ND, not determined.

infants be encouraged to continue some oral intake or have oral stimulation (e.g., pacifier) if persistent feeding dysfunction is to be prevented. Whereas the spontaneous dietary protein intake (DPI) is reduced in progressive CKD in a manner similar to that of energy intake, the DPI is typically far in excess of the average requirements. At the same time, there is no evidence that strict dietary protein restriction (120% recommended daily allowance) has any nephroprotective effect, nor does this level of intake compromise growth. Because moderate protein restriction reduces the accumulation of nitrogenous waste products and helps to lower dietary phosphorus intake, it is appropriate to gradually lower the DPI toward 100% of the dietary reference intake (DRI) as CKD progresses toward the need for dialysis (Ruely et al, 1989). More specifically, a DPI of 100% to 140% DRI for CKD stage 3, 100% to 120% DRI for CKD stage 4 to 5, and 100% DRI for CKD stage 5D (dialysis) has been proposed (Table 85-7). In patients receiving dialysis, the protein requirements are increased to account for dialysis-related protein losses. Whereas restriction of sodium and water is often indicated in children with CKD complicated by sodium and fluid retention and systemic hypertension, infants and children with CKD secondary to obstructive uropathy or renal dysplasia often have polyuria and may experience substantial urinary sodium losses despite advanced stages of CKD. This scenario can result in contraction of the extracellular volume, in addition to having an adverse effect on growth and nitrogen retention (Wassner and Kulin, 1990). Infants and children with salt-wasting forms of CKD who do not receive salt supplementation may in turn experience vomiting, constipation, and significant growth retardation (Parekh et al, 2001). The same holds true for infants receiving peritoneal dialysis even if they are anuric, because most patients lose significant quantities of sodium in the

dialysate. Sodium depletion in the PD population has resulted in cerebral edema and blindness (Lapeyraque et al, 2003). Individualized therapy can be achieved by first prescribing at least the age-related DRI of sodium and chloride (Table 85-8), with subsequent modification of therapy based on regular assessment of clinical and laboratory data. Potassium homeostasis in children with CKD is usually unaffected until the GFR falls to less than 10% of normal. However, infants and children with disorders such as renal dysplasia and reflux nephropathy often demonstrate resistance to aldosterone and may experience hyperkalemia, even when the GFR is preserved. The hyperkalemia can be exacerbated by volume contraction, as can be seen in patients with salt-wasting forms of CKD. In patients who remain hyperkalemic despite repletion of salt and water, restriction of dietary potassium intake is critical. For infants and young children, 40 to 120 mg (1 to 3 mmol/kg/d) of potassium may be a reasonable start. Breast milk has a lower potassium content (546 mg/L; 14 mmol/L) than commercial milk–based infant formula (700 to 740 mg/L; 18 to 19 mmol/L) (National Kidney Foundation, 2009). Pretreatment of infant formula with a potassium binder, treatment of constipation, and attention to medications that can exacerbate hyperkalemia (e.g., potassium-sparing diuretics, angiotensin-­converting enzyme inhibitors, angiotensin-receptor blockers) may also be necessary in many patients (Bunchman et al, 1991; Fassinger et al, 1998).

Ethics of Initiating or Withdrawing Renal Replacement Therapy Decisions to withdraw or withhold treatment have to be made for many patients in neonatology units, and for as many as 30% to 58% of patients in pediatric intensive care units. On rare occasions, such decisions pertain to

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

the infant with ESRD confronted with the prospect of a lifetime of dialysis and transplantation. With the advent of advanced technology in the 1980s that made possible the provision of safe and effective peritoneal dialysis to even the smallest infant came increasing ethical dilemmas and significant variations in practice (Fauriel et al, 2004; Shooter and Watson, 2000; Watson and Shooter, 2004). In one of the most interesting studies on the topic, Geary (1998) conducted an international survey on the attitudes of pediatric nephrologists regarding the management of ESRD during infancy (Geary, 1998). More than 200 physicians from eight countries replied to a series of questions pertaining to the provision of RRT to infants younger than 1 month of age versus those 1 to 12 months old. The factors that most often influence the decision to initiate or withhold RRT were found to be the presence of coexistent serious medical disorders and the anticipation of significant morbidity for the child. Only 25% of respondents believed it was usually ethically acceptable for parents to refuse RRT in infants beyond the 1st month of life compared with 50% with the same opinion concerning younger infants. Overall, only 41% of respondents offered RRT to all infants younger than 1 month, and 53% offered this therapy to all infants 1 to 12 months old, despite recent evidence supporting the use of this therapy as a reasonable treatment option (Carey et al, 2007). Evidence for this variation in practice has been seen in other surveys as well (Fauriel et al, 2004). Most clinicians agree that there is more to their skill than the indiscriminant application of technology. Factors to consider when making the decision regarding initiation or withdrawal of RRT during infancy include quality of life concerns, allocation of resources, legal issues, and most importantly the opinions of the hospital team and the parents. The role of hospital ethics committees in the process remains extremely variable. In the end, clinicians and parents often struggle bravely to reach a compassionate decision with as much agreement as possible. Principles of practice that may provide valuable assistance in this process have been published previously (Watson and Shooter, 2004).

Renal Replacement Therapy End-stage renal disease is an uncommon disorder in children, with an incidence in the United States of approximately 14 patients per 1 million children of a similar age (U.S. Renal Data System, 2008). The incidence varies within the pediatric population with a rate of 29 per 1 million for children 15 to 19 years old, in contrast to a rate of 9 per 1 million for children 0 to 4 years old. In a recent study by the NAPRTCS, Carey et al (2007) estimated the incidence of dialysis-treated neonatal ESRD as 0.045 cases per 1 million population per year (Carey et al, 2007). Although kidney transplantation is the nearly universal goal for children who develop ESRD, approximately 75% of pediatric patients initially receive chronic peritoneal dialysis or hemodialysis before a kidney transplant.

Peritoneal Dialysis Peritoneal dialysis is the preferred chronic dialysis modality for infants with ESRD. Subsequent to the development of continuous ambulatory PD in 1976, the availability of

1219

small dialysate bags resulted in a number of reports of infants receiving continuous ambulatory PD in the middle 1980s. The later emergence of automated cycling machines that were able to deliver fill volumes as low as 50 mL, with incremental adjustments of 10 mL, allowed for the provision of safe and effective PD in neonates and infants. The preference for PD, and specifically automated PD as the chronic dialysis modality in infants and young children, is reflected by data from NAPRTCS in which 93% of patients 0 to 1 year old and receiving dialysis were receiving PD at dialysis initiation (Carey et al, 2007; NAPRTCS, 2008). Peritoneal dialysis makes use of the peritoneal membrane as a natural dialyzing membrane. Dialysis solution is instilled and dwells within the peritoneal cavity, during which time bloodstream-derived solutes move down a concentration gradient based on diffusion, and fluid is removed as a result of the osmotic gradient created by the dextrose component of the dialysis fluid. The inflow, dwell, and drainage of dialysate characterize a single dialysis cycle or exchange. The peritoneal catheter is the cornerstone of successful PD, and the PD prescription accounts for the dextrose concentration and amount of dialysis solution (initially 600 to 800 mL/m2 body surface area in infants) during each exchange and the length of the exchange (National Kidney Foundation, 2006). The infant with ESRD who receives chronic PD is at risk for a variety of treatment-related complications that are ideally either prevented or identified early and treated aggressively if morbidity and mortality are to be minimized. The single most serious complication is peritonitis. The incidence of this infection is greatest during infancy, with a rate of 1 infection every 14.2 patient months reported by the NAPRTCS (2008). Whereas gram-positive organisms account for the majority of infections, gram-negative episodes of peritonitis are common in infants and young children (Zurowska et al, 2008). In turn, when peritonitis is suggested, empiric antibiotic therapy should provide coverage for gram-positive and gram-negative organisms (Warady et al, 2000). In some cases, infants may experience hypogammaglobulinemia in this situation and may benefit from replacement therapy (Neu et al, 1998). Other treatment-related complications that occur most frequently during infancy include anterior ischemic optic neuropathy and sudden blindness secondary to hypovolemia, excessive loss of protein across the peritoneal membrane and hernia formation (Lapeyraque et al, 2003; Quan and Baum, 1996; Warady et al, 2009).

Hemodialysis The use of HD during infancy is dictated often by the presence of a medical condition that contraindicates use of the peritoneal membrane (e.g., omphalocele, gastroschisis, diaphragmatic hernia, bladder exstrophy). The procedure is complicated, and limited clinical experience has revealed a high incidence of patient morbidity (Al-Hermi et al, 1999; Kovalski et al, 2007; Shroff et al, 2007). Whereas the infrequent use of HD in infants precludes the generation of an evidence base upon which to guide maintenance therapy, there are some key treatment-related principles to consider when conducting the procedure. Despite the

1220

PART XVII  Renal and Genitourinary Systems

availability of neonatal-sized tubing, the extracorporeal volume usually comprises more than 10% of the patient’s blood volume in infants weighing less than 8 kg; this necessitates circuit priming with 5% albumin or packed red blood cells diluted to a hematocrit of 35% (Warady et al, 2009). Blood access is typically in the form of a 7F or 8F dual-lumen catheter, which allows blood pump flow rates of 30 to 50 mL/min. Despite the use of HD machines with volumetric control capability, the accuracy is only to 50 to 100 mL, thus mandating the use of continuous monitoring of the patient’s weight with a digital scale when conducting HD on neonates. The complicated nature of the procedure mandates that it be performed only in highly qualified centers.

Transplantation The topic of kidney transplantation in patients who develop ESRD as neonates or young infants is complicated, and a lengthy discussion is beyond the scope of this chapter. In short, transplantation is a viable alternative for these young patients and is their best hope for long-term survival. In a review of NAPRTCS data, Carey et al (2007) found that 3 years after dialysis initiation, 80% of 63 neonates had undergone transplantation. What is often most important for the neonatologist is recognizing the need to develop a collaborative strategy with members of the pediatric nephrology and urology team for management of congenital structural abnormalities of the urinary tract that are present in the patient with severe CKD/ESRD, the majority of which will ultimately require transplantation (Sarwal and Salvatierra, 2004). Minimizing the use of interventions that greatly increase the risk of central venous thrombosis should also be encouraged. The most recent report of the U.S. Renal Data System reveals that 51 infants younger than 1 year and 89 children aged 1 to 4 years received a kidney transplant in 2006 (U.S. Renal Data System, 2008). The 3-year graft survival for patients who received a living donor transplant in 2003 was 83.9% and 94.6% for patients who were less than 1 and 1 to 4 years old, respectively, at the time of transplant. In the case of deceased donor transplants, the graft survival for those aged 1 to 4 years was 89.7%, while the number of deceased donor transplants performed during infancy was exceptionally small.

OUTCOMES Growth and Development Children with CKD often experience some degree of growth failure. This failure is especially concerning when CKD occurs in infancy, a time of rapid growth. By 2 years of age, approximately one third of postnatal growth has occurred (Haffner, 2008); therefore early intervention to address treatable causes of CKD-associated growth failure is essential to maximize the growth of these infants. The disordered growth in patients with CKD is a multifactorial process. Protein-calorie malnutrition, metabolic acidosis, electrolyte disarray, renal osteodystrophy, and changes in the gonadotropic hormone axis in the face of uremia, corticosteroid treatment, or both are factors that contribute to this challenging problem (Geary, 1998; Haffner, 2008).

Infants with CKD are especially susceptible to proteincalorie malnutrition because of the high energy demands in this age group. Several studies have shown that infants have impaired growth at initiation of chronic dialysis. There are conflicting reports regarding improvement in growth with renal replacement therapy. Most studies suggest that young children on dialysis fail to grow well, despite meeting 100% of the recommended daily allowance of caloric intake ­(Shroff et al, 2003). Growth outcomes may be better after transplantation (Ledermann et al, 1999), and Hijazi et al (2009) report that current growth outcomes of infants with ESRD have improved over time, possibly because of advances in medical and surgical therapies. To continue this trend of improvement, clinicians must address comorbidities including protein-calorie malnutrition in these infants with CKD. Recent reports have described improved longitudinal growth and sustained catch-up growth in infants with chronic renal failure in whom growth hormone treatment was initiated in the 1st year of life. Mencarelli et al (2009) reported 12 infants initiated on recombinant human growth hormone at a median age of 0.5 years had significantly better height SDS scores and change-in-height SDS scores versus a control group not treated with growth hormone, despite similar nutritional intake. In addition to growth impairments, neurodevelopmental impairments may also be present in children with CKD during infancy and early childhood. Like many children with chronic illnesses, children with CKD may have delayed achievement of many developmental milestones. In addition, delayed cognitive performance especially in regard to general intelligence, attention, executive function, language, visual-spatial abilities, and memory, has been repeatedly reported in children with CKD (Geary, 1998; Gipson, 2008). Whereas some children have renal involvement as part of a larger syndrome that may involve structural or functional abnormalities of the central nervous system, other comorbidities of CKD (e.g., anemia, hypertension, cerebral vascular accidents, adverse effects of therapy) have been implicated in the neurodevelopmental impairments in many patients (Geary, 1998). The brain undergoes rapid growth during infancy, reaching half of its adult weight by 6 months of age (Harris, 2006). Postnatal brain growth includes neuronal differentiation, dendritic branching, and axonal myelination (Gipson, 2008). Renal impairment in infancy, a crucial time of neural development, raises concerns regarding the neurodevelopmental outcomes in these children. In one study, 28 patients initiating chronic peritoneal dialysis by 3 months of age underwent formal neurodevelopmental testing. Only 6 of the 28 patients were functioning below the average level. Nineteen of these patients were retested after their fourth birthday. Fifteen of the 16 school-aged patients were fulltime students in age-appropriate classrooms, and all of the children younger than 5 years were in preschool (Warady, 1999). In another longitudinal study of 31 patients with ESRD diagnosed in infancy, 18 were attending regular school, 13 had significant neuropsychologic impairments, 9 required special education classes, and 4 were severely impaired (Hijazi et al, 2009). Eleven patients with ESRD diagnosed before 2 years of age were followed; six had appropriate milestones, whereas five had special education

CHAPTER 85  Acute Kidney Injury and Chronic Kidney Disease

needs (Shroff et al, 2003). Yet another study followed 16 patients with ESRD diagnosed in infancy and found that 14 achieved age-appropriate developmental milestones (Ledermann et al, 1999). Most recently, Madden et al (2003) reported on the cognitive and psychosocial outcome of 16 infants who began PD during the 1st year of life; a majority (75%) had a functioning transplant at the time of their reassessment at a mean age of 5.8 years. Ten (67%) children had an intelligence quotient in the normal range, whereas 13 of 15 (87%) were within 2 standard deviations of the mean (Madden et al, 2003). Thus, although existing reports are largely case series with relatively small sample sizes, more recent literature suggests improved neurodevelopmental outcomes for children with CKD and ESRD in infancy compared with reports before 2000. It appears that at least 25% of infants and toddlers who have severe renal insufficiency are reported to have developmental delay, whereas the effects of milder forms of CKD on the neurodevelopment of infants remains less clear. Large multicenter longitudinal studies may provide more insight into the developmental outcomes of infants with CKD.

Hospitalization There is emerging evidence to support the clinical expectation that a majority of neonates and infants with CKD or ESRD require frequent hospitalization throughout childhood. In one study of 18 children requiring chronic hemodialysis by 2 years of age, the median number of hospital admissions while receiving dialysis was 6 (range 3 to 16). Of those hospitalized, the median hospitalization rate per patient was 8.2 admissions per year with the duration ranging from 63 to 399 days (Shroff et al, 2003). Another study divided 698 children requiring chronic dialysis by 2 years old into those initiating dialysis by 1 month of age and those initiating dialysis between 1 month and 24 months of age. Approximately 80% of children in both groups required hospitalization at some point in the 13-year follow-up period. Among children ever hospitalized, those initiating dialysis as neonates were hospitalized more frequently than were children starting dialysis later (mean number of hospitalizations, 54 versus 39; p <0.001) and had longer hospital stays (Carey et al, 2007).

Renal Osteodystrophy There is a paucity of outcome data in the neonatal and infant populations regarding many of the other comorbidities associated with CKD. There is evidence suggesting that infants with secondary hyperparathyroidism from CKD have improvement with chronic renal replacement therapy. One study followed 17 patients initiating hemodialysis between birth and 2 years of age and found that the percentage of patients with intact parathyroid hormone concentrations less than twice the upper limit of normal increased after 3 months of hemodialysis (41% at initiation versus 69% after 3 months; Shroff et al, 2003). Another study of 20 infants on long-term PD revealed similar results (58% after 6 months of PD versus 100% after 1 year of PD) (Ledermann, 2000). Further study regarding the prevalence of renal osteodystrophy in this population is needed.

1221

Survival Long-term survival of neonates with CKD or ESRD appears to be similar to that of older infants and young children (Carey et al, 2007, Hijazi et al, 2009, Wedekin et al, 2008). Factors reportedly associated with mortality include African American race, presence of comorbidities such as chronic lung disease, multiorgan dysfunction, diagnosis of a syndrome, and oliguria or anuria (Hijazi et al, 2009). Approximately 25% of neonates died over 18 months of follow-up in one study, which was similar to the mortality rate of older children (Carey et al, 2007). Another study found 22% mortality (Shroff et al, 2003). Yet another study following infants with CKD for 25 years reported 54% mortality, but found a 5-year survival rate of 87% among patients surviving the 1st year (Hijazi et al, 2009). In addition, Wedekin et al (2008) followed 119 infants with CKD and estimated the 1-, 2-, and 5-year survival rates to be 78%, 68%, and 63%, respectively. However, if the patients were receiving renal replacement therapy, survival improved to 91%, 83%, and 83% for 1, 2, and 5 years, respectively. Among the patients who received renal transplants, the 1-, 2-, and 5-year survival rates were greater than 95% (Wedekin et al, 2008). A majority of children with ESRD diagnosed as neonates undergo renal transplantation in the first few years of life. In a study by Carey et al (2007) 80% of neonates with ESRD diagnosed in the neonatal period received a kidney transplant within 5 years of initiating dialysis (Carey). It appears that younger children are receiving transplants more readily in recent years (Carey), which is encouraging in the face of more favorable survival.

SUGGESTED READINGS Andreoli SP: Acute renal failure in the newborn, Semin Perinatol 28:112-123, 2004. Askenazi DJ, Ambalavanan N, Goldstein SL: Acute kidney injury in critically ill newborns: what do we know? What do we need to learn? Pediatr Nephrol 24:265-274, 2009. Basile DP: The endothelial cell in ischemic acute kidney injury: implications for acute and chronic function, Kidney Int 72:151-156, 2007. Brenner BM, Lawler EV, Mackenzie HS: The hyperfiltration theory: a paradigm shift in nephrology, Kidney Int 49:1774-1777, 1996. Carey WA, Talley LI, Sehring SA, et al: Outcomes of dialysis initiated during the neonatal period for treatment of end-stage renal disease: a North American Pediatric Renal Trials and Collaborative Studies Special Analysis, Pediatrics 119:e468-e473, 2007. Geary DF: Attitudes of pediatric nephrologists to management of end-stage renal disease in infants, J Pediatr 133:154-156, 1998. Hijazi R, Abitbol C, Chandar J, et al: Twenty-five years of infant dialysis: a single center experience, J Pediatr 155:111-117, 2009. Mehta RL, Kellum JA, Shah SV, et al: Acute Kidney Injury Network (AKIN): report of an initiative to improve outcomes in acute kidney injury, Crit Care 11:R31, 2007. National Kidney Foundation: KDOQI clinical practice guideline for nutrition in children with CKD: 2008 update, Am J Kidney Dis 53(Suppl 2):S1-S124, 2009. Rees L: Management of the neonate with chronic renal failure, Semin Fetal Neonat Med 13:181-188, 2008. Shooter M, Watson A: The ethics of withholding and withdrawing dialysis therapy in infants, Pediatr Nephrol 14:347-351, 2000. Symons JM, Chua AN, Somers MJ, et al: Demographic characteristics of pediatric continuous renal replacement therapy: a report of the prospective pediatric continuous renal replacement therapy registry, Clin J Am Soc Nephrol 2:732738, 2007. Watson AR, Shooter M: The ethics of withholding and withdrawing dialysis in children. In Warady BA, Schaefer FS, Fine RN, Alexander SR, editors: Pediatric dialysis, Dordrecht, the Netherlands, 2004, Kluwer Academic Publishers, p 501. Wedekin M, Ehrich J, Offner G, Pape L: Aetiology and outcome of acute and chronic renal failure in infants, Nephrol Dial Transplant 23:1575-1580, 2008. White SL, Perkovic V, Cass A, et al: Is low birth weight an antecedent of CKD in later life? A systematic review of observational studies, Am J Kidney Dis 54:248-261, 2009. Complete references used in this text can be found online at www.expertconsult.com