Key Points

• 
  

  There are well-described renal anatomic and physiologic adaptations associated with aging.

  
  
• 
  

  While a loss of glomerular filtration rate (GFR) is not an absolute finding with aging, in general, GFR has an inverse relationship with age; estimation equations for GFR exist and should guide therapeutic decision making in the elderly with stable renal function.

  
  
• 
  

  The typical renal functional changes of aging restrict the kidney’s ability to maintain homeostasis in the face of physiologic challenges, resulting in an increased incidence of dysnatremias and other electrolyte disorders such as hypercalcemia and hyperkalemia.

  
  
• 
  

  Acute kidney injury (AKI) is not exclusive to the elderly but is more common and is associated with an increased risk of nonrecovery of renal function in the aged compared with the young.

  
  
• 
  

  A host of cellular mechanisms associated with senescence may collectively increase the risk of AKI and retard the regenerative capacity of the kidney, limiting renal recovery from AKI in the aged.

  
  
• 
  

  The hazard of a renal biopsy is not increased by age itself, and specimen quality is not diminished; thus, in situations of diagnostic uncertainty of AKI, renal biopsy should not be withheld in the elderly because it may offer high diagnostic yield and effectively guide therapeutic treatment strategy.

  
  
• 
  

  The mainstay of renal replacement therapy in the elderly is prevention of complications, particularly those related intradialytic hemodynamic instability and malnutrition.

  
  
• 
  

  Current guidelines pertaining to the management of patients with AKI are age agnostic, yet elderly patients do have an increased risk of AKI and persistent renal dysfunction. This disconnect underscores the need for further research to establish evidenced-based critical care management strategies of the very elderly with AKI.

Introduction

Renal dysfunction is common in the critically ill elderly. In contrast, an understanding of the mechanisms contributing to this dysfunction is limited. Adopting a Bayesian approach, whereby available knowledge of normal renal senescence and acute kidney injury in the critically ill population are considered, the expected disruptions to renal homeostasis in the hospitalized elderly with life-threatening illness can be predicted and mitigated.

Renal Function and Aging: Epidemiology

A reduction in glomerular filtration rate (GFR) in the elderly compared with younger people was noted more than 60 years ago and has been confirmed by numerous subsequent studies [1]. In healthy adults, the average loss of GFR with increasing age is estimated to be approximately 0.75 ml/min per year [2]. Closer inspection, however, shows substantial heterogeneity in GFR loss across as well as within age strata. For example, a substantially higher rate of GFR decline is described in older healthy adults (40–80 years of age) compared with younger individuals (1.51  versus 0.26 ml/min, respectively) [2]. In addition, there is variation in the rate of decline within elderly cohorts, with over one-third of subjects manifesting no decrement in renal function with age [2,3], even when using GFR estimation equations in lieu of creatinine clearance [4] and among the “oldest old” [5]. Not surprisingly, the rate of decline of GFR in older populations that include subjects with comorbidities is substantially higher than the average reported for healthy cohorts, approximating 2.6 ml/min per year [6].

Thus, with aging, loss of GFR is not an absolute finding. Rather, there is variation in the decline in GFR with aging, with some healthy adults maintaining GFR, whereas in most adults the GFR declines at 1 ml/min per year and perhaps faster in the presence of additional comorbidities.

Estimating Renal Function in the Elderly

While serum creatinine (SCr) remains the most commonly used biomarker for estimating GFR, reduced dietary protein intake and changes in body composition with aging (e.g., sarcopenia) both reduce SCr. As a consequence, reliance on SCr alone leads to overestimation of GFR, and the presence of renal disease can be missed even when the SCr is within the laboratory reference range. This has led to a search for more reliable biomarkers of GFR as well as the development of equations with less variability in GFR estimation.

The Cockcroft-Gault (C-G) formula estimates creatinine clearance (eCrCl) using an equation derived from a small study of Caucasian men aged 18 to 92 years, with and without chronic kidney disease (CKD) [7]. As is true for measured urinary creatinine clearance, eCrCl typically overestimates GFR due to tubular secretion of creatinine.

Despite this limitation, for drug dosing, the eCrCl is the historical equation of choice for all ages, including the elderly. This is because eCrCl underestimates renal function [8] relative to the other commonly used GFR estimation equations (e.g., the Modification of Diet in Renal Disease [MDRD] formula, discussed below) [9]. GFR underestimation is especially preferred in the elderly because additional factors associated with aging may increase a drug’s pharmacologic activity and toxicity (e.g., altered volume of distribution, reduced serum albumin concentration, changes in tubular handling). Because these latter factors are not considered when dosing a drug based on GFR alone, the more conservative GFR estimation approach to dosing is recommended.

The second SCr-based GFR estimation equation, widely adopted for laboratory coreporting of GFR along with SCr, is the Modification of Diet in Renal Disease (MDRD) formula. Generated from the Modification of Diet in Renal Disease study, it is derived from a population of predominantly Caucasian nondiabetic US adults from 18 to 70 years of age with an eGFR of less than 60 ml/min/1.73 m² [10]. While reportedly “suitable for use across populations with chronic kidney disease (CKD)” [11] and more accurate for estimating GFR in the elderly than eCrCl, in the absence of a study specifically validating its performance in the elderly, this claim remains debatable [8].

A third eGFR method, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, is derived from the Chronic Kidney Disease Epidemiology Collaborative – a large (>10,000 subjects) racially and ethnically diverse population of men and women with and without CKD and diabetes [12]. In lieu of SCr, cystatin can be substituted as the serum biomarker. Regardless of the biomarker used, its relative precision as compared with the MDRD equation for estimating renal function in elderly patients with a GFR greater than 60 ml/min/1.73 m² remains to be determined.

Regardless of the formula used, all eGFR equations assume stability of SCr at the time of measurement, and thus laboratory-reported eGFR should be cautiously interpreted in patients with rapidly changing kidney function. Care is also warranted in the interpretation of GFR in the critically ill, where variable creatinine generation rates violate a second assumption of SCr-based equations. Further, SCr-based estimation of GFR may be inaccurate in patients with reduced muscle mass and those with irregular dietary intake, blood loss, or significant volume infusion, all of which are often seen in the elderly and/or critically ill [13].

Kidney Changes in Aging

The renal changes of aging are manifest throughout the nephron and are both structural and functional. Functionally, the aging renal vasculature demonstrates reduced endothelial cell capacity for nitric oxide production, as well as increased sensitivity to endothelin-1 and angiotensin II [14]. The elasticity of blood vessels also diminishes with age. The net result is a predictable rise in systolic arterial blood pressure (BP) with a widening of pulse pressure with progressive age [15]. Structurally, renal arteriolar hyalinosis accompanies the functional vascular changes of aging, with the relative contribution of hypertension versus aging as drivers of its development still under debate [15].

The restricted ability of the aging renal vasculature to effectively modulate blood flow in response to dynamic changes in arterial pressure results in increased glomerular capillary hydraulic pressure. Progressive glomerular sclerosis ensues, especially in the renal cortex, resulting in a loss of nearly 7,000 glomeruli per year after age 18 [16]. The involution of glomeruli results in atrophy of both afferent and efferent arterioles. As a consequence, the peritubular capillary density is eroded, and tubular atrophy and interstitial fibrosis become manifest. In juxtamedullary glomeruli, tuft sclerosis results in the development of direct arteriovenous shunts (aglomerular arterioles), accounting for the relative preservation of medullary blood flow in the senescent kidney [16]. Combined, the anatomic changes of aging result in reduced renal mass after the fourth decade, with cortical volume loss exceeding that of the medulla, and an average mass of less than 300 g by 80 years of age [16].

The tubular loss and interstitial fibrosis due to aging contribute to a variety of salt and water syndromes that are commonly manifest in the hospitalized as well community-dwelling elderly – notably the dysnatremias, hyperkalemia, and hypercalcemia.

Hypernatremia

Hypernatremia (SNa⁺ > 145 mEq/liter) is relatively common in hospitalized patients with a reported incidence of roughly 1 percent in hospitalized patients as well as in long-term care facilities [17–19]. Hypernatremia is associated with substantial in-hospital mortality – up to 40 to 55 percent [17,18,20], although it should be noted that the mortality is typically related to the underlying disease process rather than the hypernatremia itself [19].

Hypernatremia is especially prevalent in the elderly; the mean age of patients admitted with this diagnosis is typically in the mid-70s [17,20]. Physiologic factors that predispose aging individuals to hypernatremia include the impaired renal concentration mechanisms that occur with age. Whereas young adults can achieve a urinary osmolality as high as 1,200 mOsm/kg, older individuals can achieve a maximum concentration of only 700 to 800 mOsm/kg [21]. Arginine vasopressin (AVP) levels are elevated in older individuals subjected to water deprivation, suggesting that the concentration deficit is due to renal tubular changes and not central deficiencies in osmoregulation. Studies in experimental animals suggest that this defect relates to decreased AVP-induced cyclic AMP generation and a consequent decrease in aquaporin expression [21–23]. In addition to these renal tubular epithelial (RTE) changes, healthy and cognitively intact elderly individuals do not have an adequate thirst response to water deprivation compared with younger subjects [22], and the risk is even greater in demented individuals [24].

Other clinical risk factors associated with hypernatremia in the elderly include female gender, age greater than 85 years, having more than four chronic conditions, limited mobility, infections, and altered mental status [19]. Additional contributing factors common in the elderly include febrile illness, diabetes, diarrhea, diuretics, gastrointestinal (GI) bleed, and intravenous (IV) solute [18], and a voluntary decrease in fluid intake due to concerns about urinary incontinence [19]. Increased water loss may be particularly problematic during the summer months [17]. Finally, hypernatremia may occur due to increased osmotic intake, whether via increased sodium/solute intake in patients with limited fluid intake or develop iatrogenically via the provision of hypertonic IV infusions, such as the use of IV bicarbonate in patients receiving cardiopulmonary resuscitation [19].

While the elderly are at particular risk of hypernatremia, the management of hypernatremia in the elderly does not differ substantially from that in any age group. The reader is thus referred to a general review of its management [25].

Hyponatremia

As with hypernatremia, renal functional changes leave elderly patients susceptible to hyponatremia. Hyponatremia is common in the elderly population, especially in institutionalized patients, in whom the prevalence is more than double that of age-matched ambulatory adults (18 versus 8 percent). Remarkably, 53 percent of nursing home denizens, particularly those with central nervous system or spinal cord lesions, experienced at least one episode of clinical hyponatremia in the preceding year [26], increasing the risk of hospitalization by 10 percent with a fourfold increased risk of an adverse outcome in hospitalized patients [27].

Clinically, the presentation of patients with hyponatremia depends not only on the severity of the hyponatremia but also on the rate at which it develops. Symptoms are often observed in cases of acute hyponatremia (i.e., the development of hyponatremia in < 48 hours). Patients with mild acute hyponatremia (SNa⁺ = 130–134 mmol/liter) may be asymptomatic or may report anorexia, cramping, headache, or irritability. Moderate acute hyponatremia (SNa⁺ = 125–129 mmol/liter) may lead to disorientation, confusion, weakness, or lethargy. Acute severe hyponatremia (SNa⁺ < 125 mmol/liter) can lead to nausea, vomiting, seizures, coma, respiratory arrest, or permanent brain damage [21,28].

Previously believed to be inconsequential, hyponatremia is, to the contrary, associated with a significant increase in the risk of fracture in the elderly [29–32], independent of the presence of osteoporosis [32]. Likely contributors to this heightened fracture risk include increased gait instability, attention deficits, and falls [30]. The identification of such neurologic deficits raises the question of whether hyponatremia is truly benign in the elderly and whether more aggressive management of the hyponatremia should be considered [33].

Elevated vasopressin levels are overwhelmingly present in hospitalized patients with hyponatremia [34]. A number of pathophysiologic mechanisms drive the nonosmotic release or increased effect of vasopressin, many of which are more likely to be present in the elderly [21,35] (Table 8.1).

Management of Hyponatremia

The management of hyponatremia in the elderly patient is similar to that in other adults, and thus the reader is referred to recent guidelines for detailed recommendations [36]. Noteworthy is that the most dreaded complication of hyponatremia, brain herniation, typically occurs with the development of hyponatremia during rapid cerebral uptake of water. Because this more often occurs in settings of excessive water intake (e.g., marathon running, ecstasy use), it is less common in the elderly.

In contrast, the elderly more often present with chronic hyponatremia and are at risk of iatrogenic injury from overly rapid correction of hyponatremia. Patients with liver disease, alcoholism, hypokalemia, malnutrition, or a SNa⁺ of less than 105 mEq/liter seem to be at particular risk of overly rapid correction. Current recommendations are that asymptomatic patients with chronic hyponatremia have their SNa⁺ corrected by no more than 6 to 8 mEq/liter per day because rare cases of osmotic demyelination have been reported with correction rates as low as 10 mEq/liter per day [37]. A minimum correction of 4 to 8 mEq/liter per day has been advocated as well.

Patients who develop severe hyponatremia (SNa⁺ < 120 mEq/liter) due to chronic thiazide use or hypovolemia are at particular risk of overly rapid correction. This is a consequence of a spontaneous aquaresis following volume-deficit correction. Also at high risk of overcorrection are patients whose hyponatremia develops as a result of cortisol deficiency or chronic desmopressin use. In such patients, therapeutic interventions to prevent rapid overcorrection are indicated. Serum Na⁺ and urine volume should be monitored every 4 to 6 hours. If the daily target has been achieved, further correction of the serum Na⁺ level via ongoing urinary water loss should be avoided by replacing it 1:1 with oral water or 5% dextrose or by parenteral desmopressin administration [36]. If the daily correction target has been exceeded, a relowering of the serum Na⁺ level via water provision and use of desmopressin is recommended [36].

Hyperkalemia

Hyperkalemia is another frequent electrolyte disorder of the elderly due to renal functional changes with aging. The decline in GFR with age predisposes elderly patients to hyperkalemia but is typically not sufficient to cause hyperkalemia until the GFR is less than 30 ml/min, suggesting that disorders in tubular function contribute to the propensity to hyperkalemia as well. A number of mechanisms inhibit the distal nephron’s ability to secrete potassium. Urinary tract obstruction, common among elderly males, triggers hyperkalemic renal tubular acidosis via a disruption of distal nephron potassium secretion [38]. Aldosterone is an important secretagogue of potassium in the distal nephron, but hyporeninemia, occurring as a by-product of the normal aging process or as a consequence of diabetes mellitus [39], may lead to hypoaldosteronemia. Certain medications, including nonsteroidal anti-inflammatory drugs, cyclooxygenase II inhibitors, and β-adrenergic blocking drugs, suppress renin secretion as well [40–43]. Selective hypoaldosteronism has also been described in association with amyloidosis, Sjögren syndrome, and cases of metastatic carcinoma to the adrenal gland [44–47] and is a well-known side effect of ketoconazole [39] and long-term heparin use [48]. Hyperkalemia is a common side effect of angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, and direct renin inhibitors. Amiloride and spironolactone directly antagonize the ability of the principal cell to secrete potassium, an activity shared by the antibacterial agent trimethoprim [43].

Potassium excretion in the distal nephron requires adequate sodium delivery and fluid flow [49]. Effective circulating volume depletion, either by hypovolemia (in the setting of diarrhea, sepsis, or liver disease or diuretic use) or hypervolemia (as seen with congestive heart failure) may lead to inadequate distal nephron sodium delivery, thus impairing the kidney’s ability to secrete potassium. Similarly, flow-sensitive channels in the distal nephron secrete potassium in response to high urine flow – failure to maintain high flow due to hypovolemia or inadequate fluid intake will similarly compromise renal potassium excretion [43]. Elderly patients often avoid excess fluid intake either due to impaired thirst mechanisms or due to concerns about frequent urination or incontinence and thus may be at particular risk for hypovolemia, thus impairing the kidney’s ability to secrete potassium.