Pathogenesis and Management of Metabolic Acidosis and Alkalosis
Acid–base disorders occur commonly in clinical medicine. Although the degree of acidosis or alkalosis that results is rarely life threatening, the careful evaluation of the acid–base status of the patient often provides insight into the underlying medical problem. Moreover, the pathophysiology and differential diagnosis of these disorders can be approached quite logically with a minimum of laboratory and clinical data. An effective approach to clinical acid–base disorders is accomplished most easily with a stepwise pathophysiologic approach.
Human acid–base homeostasis normally involves the tight regulation of CO2 tension by respiratory excretion and plasma bicarbonate [HCO3−] concentration by renal HCO3− reabsorption and elimination of protons (H+) produced by metabolism. The pH of body fluids (which can be sampled easily) is determined by the CO2 tension (in arterial blood, Paco2) and the [HCO3−]. Primary derangements of CO2 tension are referred to as respiratory disturbances, whereas primary derangements of [HCO3−] are called metabolic disturbances (1).
In this chapter, we first review acid–base chemistry and physiology and then present a pathophysiologic approach to the diagnosis and management of metabolic acidosis and alkalosis.
Acid–Base Chemistry and Physiology
The chemistry of acids, bases, and buffers and the normal physiology of acid and bicarbonate excretion (2,3) are described in detail in several excellent reviews and are summarized only briefly in this section.
Clinical acid–base chemistry basically is the chemistry of buffers. For clinical purposes, we may define an acid as a chemical that donates an H+ and a base that accepts an H+. For any acid (HA), we can define its strength or tendency to donate H+ by its dissociation constant K according to the relationship:
If we rearrange this equation and apply a log transformation, we arrive at the familiar relationship:
Buffering refers to the ability of a solution containing a weak or poorly dissociated acid and its anion (a base) to resist change in pH when a strong acid (i.e., highly dissociated acid) or alkali is added. To illustrate this important point, suppose 1 mL of 0.1 mol/L HCl is added to 9 mL of distilled water, the [H+] would increase from 10−7 to 10−2 mol/L. In other words, the pH would fall from 7.0 to 2.0. However, if we added 1 mL of 0.1 mol/L HCl to 9 mL of a 1 mol/L phosphate buffer (pK = 6.9) at pH 7.0, most of the dissociated H+ from HCl would combine with dibasic phosphate (HPO42−) and only slightly change the ratio of dibasic to monobasic (H2PO4−) phosphate. In fact, the pH would fall by only about 0.1 pH units. The addition of acid has been buffered by the phosphate dissolved in water. Another way to think about this is that the pH was stabilized by substances that bound the free H+ released by the HCl, in this case the phosphate. Such substances are called buffers (4).
Biochemical Determinants of pH
The bicarbonate buffer system is the most important buffer in the extracellular space in humans. Proteins and inorganic phosphate are less important buffers. Inorganic phosphate is probably the most important buffer in the intracellular space followed by bicarbonate and intracellular proteins. Although intracellular pH (pHi) is probably more important in predicting physiologic and clinical consequences than extracellular pH (5), it is difficult to measure in vivo without using sophisticated investigational techniques, such as 31P nuclear magnetic resonance (NMR) spectroscopy (6), laser scanning cytometry (7), and fluorescence lifetime imaging (8), which are not available for routine clinical applications. Therefore, our clinical efforts are focused on classifying disease states based on what is measurable, that is, extracellular pH. In particular, our attention focuses on the bicarbonate buffer system (2). We can assume that equilibrium conditions apply because there is abundant carbonic anhydrase in blood. Therefore, we can view the bicarbonate buffer system as the equilibrium reaction:
H2CO3 is defined by the partial pressure of CO2 and the solubility of CO2 in physiologic fluids, which is a constant S to all intents and purposes. We can therefore rearrange this equation as
which is attributed to Henderson in 1909.
Taking the antilog of both sides gives the following:
which is called the Henderson–Hasselbalch equation, first described by Hasselbalch in 1916. In blood at 37°C, the pK of the bicarbonate buffer system is 6.1 and the solubility coefficient for CO2 is 0.03. Therefore, we can simplify our expression to
In the above equations, [HCO3−] is expressed in mM (or mEq/L) and PaCO2 is expressed in torr (or mm Hg). Convenient expression allows us to view acid–base disorders as being attributable to the numerator of the ratio (metabolic processes), the denominator (respiratory processes), or both (mixed or complex acid–base disorders) (Fig. 3-1) (1).
Total Body Acid–Base Metabolism
A myriad of enzymatic reactions involve the loss or gain of protons that occurs with ongoing catabolism and anabolism. However, one simply has to examine the initial substrates and final products to understand whether acid or base is produced. To do this, it is helpful to think of acids and bases as “Lewis” acids and bases, in other words, to consider acids as electron acceptors rather than proton donors. In concrete terms, acid is generated when a substrate is metabolized to something more anionic (e.g., glucose is metabolized to lactate through the Embden–Meyerhof glycolytic pathway). Conversely, if a substrate is metabolized to something more cationic (e.g., lactate is metabolized to CO2 and H2O via the tricarboxylic acid [TCA] cycle), then acid is consumed (9). Because of the importance of the bicarbonate buffer system in the overall acid–base homeostasis, we generally consider the addition of a proton as equivalent to the decrease in total body HCO3− and loss of a proton as a gain in HCO3− (9).
This approach to understanding acid–base metabolism has led to its practical application in treating the acidosis of chronic renal failure with peritoneal dialysis. Although the lactate-based dialysate (generally at about 35 mM) has a pH of only 6.0, virtually all of the lactate is ionized; bicarbonate is generated and the acidosis is corrected via metabolism largely through the TCA cycle to CO2 and H2O (10).
Renal Acid Excretion
In adult men studied at sea level, the kidneys regulate the [HCO3−] at approximately 24 mM and the lungs control the PaCO2 at about 38 torr, thus producing an arterial pH of approximately 7.42; for women, the corresponding values for [HCO3−], PaCO2, and pH are 24 mM, 37 torr, and 7.43, respectively (11). The kidneys regulate plasma [HCO3−] and acid–base balance by reclaiming filtered HCO3− and generating new HCO3− to replace that lost internally in titrating metabolic acid and externally (e.g., from the gastrointestinal [GI] tract). A normal “Western diet” generates approximately 1 mmol of acid per kilogram of body weight per day. This acid load must be excreted by the kidney to maintain acid–base homeostasis. The easiest way to understand the molecular processes involved in renal acid excretion is to separate renal acid–base handling itself into two functions: bicarbonate reabsorption and net acid excretion (NAE) (12).
RENAL CELLULAR MECHANISMS OF PROTON EXTRUSION
In recent years, our understanding of the renal cellular transport proteins that effect H+ extrusion has expanded significantly. We now know that in addition to the sodium–proton exchanger (Na+/H+ exchanger) that exchanges one H+ for one sodium molecule, the sodium phosphate symporter that transports one sodium with one monobasic phosphate molecule, and the vacuolar H+ ATPase that directly pumps H+ into the tubular lumen (13–19), other transport proteins may be of considerable importance. These other transport proteins include the family of chloride–bicarbonate symporters and exchangers, the “colonic” H+/K+ ATPase, and the Na+/K+ ATPase (20–23). These transport proteins are expressed to different degrees in the different nephron segments depending on the function of the cell type within that nephron section.
On a more global basis, there is a tight link between acid secretion and the reclamation of filtered bicarbonate as well as the production of new bicarbonate by the kidney. For example, the reclamation of HCO3− filtered from the blood occurs when HCO3− formed inside the renal tubular cells by either H+ secretion or ammonium (NH4+) synthesis is transported back into the blood via the basolateral Na+(HCO3−)3 symporter (23) or a Cl−/HCO3− antiporter (13,24). Alternatively, bicarbonate can be secreted by certain cells of the collecting duct in exchange for chloride (25).
RENAL ACID–BASE METABOLISM
When plasma is filtered at the glomerulus, HCO3− enters the tubule lumen. Mechanistically, each HCO3− that is reclaimed requires the epithelial secretion of one H+. This is accomplished largely by an Na+/H+ exchanger on the luminal membrane, although an electrogenic H+ ATPase also may be involved. On an organ physiology level, HCO3− reabsorption can be considered in terms of the plasma threshold (PT) for bicarbonate, that is, the plasma HCO3− concentration at which HCO3− begins to appear in the urine. In terms of the maximal net activity of tubular HCO3− reabsorption (also called Tmax), assuming glomerular filtration rate (GFR) of 100 mL/min and a plasma [HCO3−] of 24 mM, the renal tubules must secrete about 2.4 mmol of H+ per minute to reclaim all of the filtered HCO3−. Therefore, HCO3− reclamation by the tubules involves a tremendous amount of H+ secretion. Bicarbonate reclamation is coupled tightly to sodium reabsorption and also is sensitive to a number of other influences. As the Tmax for HCO3− increases, the PT for HCO3− increases. Conversely, decreases in Tmax result in decreases in the PT. In particular, states of extracellular fluid (ECF) expansion and decreases in PCO2 decrease the apparent Tmax for HCO3−, whereas ECF contraction and increases in PCO2 increase the apparent Tmax for HCO3−. Parathyroid hormone (PTH) inhibits proximal tubule HCO3− reabsorption and lowers the apparent Tmax and PT for HCO3−. Most (but not all) of this HCO3− reabsorption (about 85–90%) occurs in the proximal tubule (3).
Carbonic anhydrase that is present both intracellularly and on the tubular surface of the brush border of the proximal tubule allows the secreted H+ that combines with tubular fluid HCO3− to form H2CO3. This H2CO3 rapidly dissociates to form H2O and CO2 that can readily permeate proximal tubule cell membranes. Intracellularly, carbonic anhydrase catalyzes the formation of H2CO3 again, which subsequently dissociates into HCO3− and H+. Finally, HCO3− leaves the cell via several bicarbonate transport proteins, including the Na+(HCO3−)3 symporter and the Cl−/HCO3− exchanger (26). The secreted H+ also titrates citrate, another form of alkali (27). This process is shown schematically in Figure 3-2.
Net Acid Excretion
NAE is the net amount of H+ eliminated from the body. If we postulate that an excreted HCO3− molecule negates the value of an excreted H+, then we can consider NAE by the kidney to be the amount of H+ (both buffered and free) excreted in the urine minus the amount of HCO3− excreted in the urine. As discussed, H+ secretion into the tubule lumen mandates 1:1 stoichiometric HCO3− transport across the basolateral segment into the extracellular space; therefore, NAE represents the amount of new HCO3− generated by the kidneys and added to the body stores.
NAE is accomplished primarily through elimination of titratable acid (which is mostly phosphate) and nontitratable acid (in the form of NH4+) (12). These terms refer to clinical chemistry titration techniques by which known amounts of alkali were added to urine until a color change with a pH indicator (e.g., phenolphthalein) occurred. This color change occurred above the pK for the phosphate buffer system but below the pK for ammonia–ammonium (about 9). It must be stressed that the NAE is relatively insensitive to the urine pH. This concept is illustrated by the observation that addition of 1 mmol of HCl to 1 L of distilled water results in a pH of 3 (corresponding to an H+ concentration of 10−3 mol/L). Therefore, in the extreme case (i.e., no buffers in the urine at all), extremely acidic urine (i.e., a very low pH) could eliminate very few protons from the body. There are several clinical settings discussed later in which acidic urine is elaborated, but NAE is insufficient. NAE requires the adequate function of both the proximal tubule to synthesize NH4+ (which generates an HCO3− molecule) and the distal tubule and collecting tubules where H+ and NH4+ secretion occur (26).
Proton secretion by the distal nephron involves the production of an electrogenic gradient that favors H+ secretion produced by removal of sodium from the luminal fluid and direct pumping of H+ into the tubular lumen. The latter is accomplished by the activities of vacuolar H+ ATPase and the H+/K+ ATPase in type A intercalated cells. Notably, chloride exchange with bicarbonate on the basolateral side of these distal tubular cells allows for proton secretion to be translated into bicarbonate addition to the blood. Finally, the epithelial membrane must not allow back leak of H+ or loss of the electrogenic gradient. Under normal circumstances, humans can elaborate a urine pH as low as 4.4, representing a 1,000:1 gradient of H+ between tubular fluid and cells. However, the excretion of NH4+, which is discussed later, is of much greater importance in terms of NAE than is the level of urine pH achieved (28).
NAE is sensitive to a variety of factors, including the plasma K concentration (increases in plasma K decrease NH4+ excretion, whereas decreases enhance H+ secretion by the distal nephron) and the effects of aldosterone. By stimulating the renin–angiotensin–aldosterone system, ECF contraction enhances distal acid secretion (29).
The traditional view that NH4+ excretion was determined by simple passive trapping of NH4+ in the tubular lumen has been revised considerably. Recent studies have led to new insights into mechanisms of renal ammonia transport and metabolism (30). Although many of these proteins are primarily involved in the transport of H+ or K+, they also transport NH4+ (31). The role of aquaporins and the identification of mammalian Rh glycoproteins (discussed below) are two new exciting areas of research in understanding ammonia transport in normal acid–base homeostasis (30). Importantly, proximal tubule cells deaminate glutamine to form alpha ketoglutarate (αKG) and NH4+. Proximal tubule cells then secrete NH4+ into the lumen, probably via substitution for an H+ using the luminal Na+/H+ antiporter. A key feature is that the further metabolism of αKG generates a new HCO3− molecule. Therefore, complete metabolism of each glutamine produces two NH4+ and two HCO3− ions. Ammonium is later reabsorbed in the thick ascending limb of Henle (via substitution for K+ using the Na+/K+/2Cl− cotransporter in the apical membrane), then across the basolateral membrane via the sodium–hydrogen exchanger isoform 4 (NHE4) to the tubular interstitial space (32). This ultimately results in the increase in the medullary concentration of NH4, which is highest in the inner medulla. This NH4+ is then taken up by distal convoluted tubule and collecting duct cells, substituting for K+ using the basolateral Na+/K+ ATPase of the inner medullary collecting duct or as NH3 (after losing H+) utilizing ammonia-specific transporters Rhesus Glycoproteins (Rhbg and Rhcg) along with a component of diffusive absorption. Collecting duct ammonia secretion involves parallel secretion of NH3 utilizing Rhcg of the apical membrane or by its diffusive capacity along with cytosolic H+ generated by carbonic anhydrase II–mediated mechanism. H+ secretion across the apical membrane is primarily mediated by H+-ATPase and H+-K+ ATPase pumps. In the tubular lumen, H+ titrates luminal NH3 to form NH4+ and maintains low NH3 concentration in the urinary space, which is necessary for continued NH3 secretion (33). The net generation of any HCO3− from αKG metabolism is ultimately dependent on the excretion of NH4+. This is because if this NH4+ molecule is not excreted in the urine but rather is returned via the systemic circulation to the liver, it is used to form urea at the expense of generating an H+. Thus, the HCO3− molecule that was generated by the metabolism of the αKG will be neutralized, and no change in acid–base status will occur (30).
Difficulties in the routine clinical measurement of urinary NH4+ concentrations have delayed our appreciation of its importance in net acid–base balance during pathophysiologic conditions including chronic metabolic acidosis or acid loads (31). However, recent observations by Batlle et al. (34) suggest that the urinary [NH4+] may be inferred fairly accurately by calculations based on urinary electrolyte concentrations as is discussed subsequently.
Although the topic under discussion is “renal” acid–base metabolism, the liver may be significantly involved. Hepatic glutamine synthetase expression appears to be regulated by pH as well as protein ingestion (35–37). Perhaps more importantly, administration of amino acids via a parenteral rather than the enteral route often is accompanied by acid retention (38).
Clinical Approach to Acid–Base Disorders
DEFINITIONS AND CAUSES
Metabolic acidosis is a systemic disorder characterized by a primary decrease in [HCO3−]. This may occur in three ways: (a) the addition of strong acid that is buffered by (i.e., consumes) HCO3−; (b) the loss of HCO3− from body fluids, usually through the GI tract or kidneys; and (c) the rapid addition to the ECF of nonbicarbonate solutions (dilutional acidosis). No organic anion is generated when HCO3− is lost or diluted. Reciprocal increases in the serum chloride concentration occur to preserve electroneutrality. Thus, these forms of metabolic acidosis generally are referred to as hyperchloremic or nonanion gap metabolic acidosis. When an organic acid consumes HCO3−, its organic anion is generated and may be retained in the ECF and serum. The serum chloride concentration does not increase with organic acidosis. An increase in the anion gap marks the existence and concentration of the organic anion (39).
DEFENSE OF SYSTEMIC pH DURING METABOLIC ACIDOSIS
The hallmark of metabolic acidosis is a fall in plasma [HCO3−]. We stress that the fall in [HCO3−] always is mitigated by the participation of other buffers in both the ECF and the intracellular fluid (ICF). Roughly one-half of an administered acid load is buffered by nonbicarbonate buffers (40). Bone is an important buffer pool in states of chronic metabolic acidosis. In fact, the leaching of calcium from bone is one of the major deleterious effects of chronic metabolic acidosis (41–43).
A fall in PaCO2 is a normal compensatory response with simple metabolic acidosis. The failure of this normal adaptive response is indicative of the presence of respiratory acidosis in the setting of a complex or mixed acid–base disturbance. Conversely, an exaggerated fall in PaCO2 producing a normal pH indicates the presence of respiratory alkalosis in the setting of a complex or mixed acid–base disturbance (44). The mechanism by which metabolic acidosis induces hypocapnia appears to be mediated in part by peripheral pH receptors in the carotid body, but mostly by central nervous system (CNS) pH receptors. This point is supported by the time course, specifically the temporal delay observed for respiratory compensation seen in experimental metabolic acidosis (45). The degree of chronic compensation varies from person to person; however, based on a large volume of clinical data, we can state with some confidence that the appropriate fall in PaCO2 (in torr) should be 1 to 1.5 × the fall in [HCO3−] (in mM) (44). Oral acid loading in normal subjects produced a rapid fall in PCO2 that reached a steady state after 30 minutes that was 0.85 times the fall in HCO3−, thus providing direct evidence for the rapid respiratory response to metabolic acidosis in humans (46).
The kidney provides the mechanism for the third line of defense to pH changes. However, this mechanism is rather slow compared to buffering (which begins immediately) and respiratory compensation (which begins within 15–30 minutes), since it takes up to 5 days to become maximal. NAE is increased in response to either metabolic acidosis (unless the kidney is the cause) or respiratory acidosis. This increase in NAE is largely NH4+ excretion because titratable acid excretion is limited by the amount of excreted phosphate, which changes very little. Metabolic acidosis increases the processing of glutamine into NH4+, which, in turn, leads to enhanced generation of HCO3−, by both transcriptional and translational regulations of key enzymes in this pathway (47). Chronic metabolic acidosis increases renal endothelin-1 activity that activates the NHE3 sodium–hydrogen ion antiporter on the proximal tubule brush border (48). Thus, both the generation of new HCO3− via the glutamine system and the enhancement of HCO3− reabsorption and titratable acid formation are stimulated. Of interest, the hypocapnia that occurs because of respiratory compensation actually limits renal correction in metabolic acidosis (49). Note that renal correction never corrects the pH to normal until the disorder causing HCO3− loss or acid generation is halted.
BIOCHEMICAL AND PHYSIOLOGIC EFFECTS OF METABOLIC ACIDOSIS
Mild degrees of acidemia generally are well tolerated, at least acutely, and even may afford some physiologic advantages, such as favorable oxygen delivery from hemoglobin. However, with marked acidemia, pH < 7.10, myocardial contractility is depressed and peripheral resistance falls (50,51). These manifestations may be a result of the effect of acidosis to depress both vascular and myocardial responsiveness to catecholamines as well as innate myocardial contractility. Both myocardial β-adrenergic receptor density and physiologic responses to β agonists appear to be decreased by metabolic acidosis (52,53).
Clearly, metabolic acidosis induces an intracellular acidosis in myocytes (48,49). This intracellular acidosis in turn impairs contractile responses to normal or even elevated cytosolic calcium concentrations (54,55). Alterations in troponin I–troponin C interactions mediated by low pH appear to shift the sensitivity of troponin C to calcium (56). Additionally, impairment of actin–myosin cross-bridge cycling caused by increases in the concentration of inorganic phosphate in the monovalent form may be involved in the decreased calcium sensitivity and contractile dysfunction seen with acidosis (57). The increase in monovalent inorganic phosphate results both from the acidic environment, which increases the ratio of monovalent to divalent inorganic phosphate, and from an impairment of myocardial energy production, which increases the total intracellular concentration of inorganic phosphate (58,59). Metabolic acidosis and hypoxia appear to additively or synergistically impair myocardial function, a phenomenon consistent with the monovalent inorganic phosphate hypothesis (60). The vasodepressor effect of acidosis likely results from similar molecular mechanisms (61).
Clinically, one can observe an increase in ventilatory effort with even mild degrees of acidosis. With severe metabolic acidosis (e.g., pH < 7.20), respirations become extremely deep and rapid (Kussmaul). Mild degrees of acidosis do not appear to markedly impair hemodynamic stability, at least in subjects with otherwise normal cardiovascular function. However, severe metabolic acidosis may lead to hypotension, pulmonary edema, and, ultimately, ventricular standstill (50,62,63). Chronic metabolic acidosis, even if fairly mild, causes hypercalciuria and bone disease as bone buffering of acid leads to marked calcium losses from the bone because of increases in prostaglandin E2 (PGE2) production (64). This aspect is extremely important in determining treatment of renal tubular acidosis (RTA) or the acidosis of chronic renal failure.
Simple metabolic acidosis is characterized by a decrease in blood pH, [HCO3−], and PCO2 (through compensation). Note that a failure to lower the PCO2 by 1 to 1.5 × the fall in [HCO3−] indicates the coexistence of respiratory acidosis (44). The clinical implications of this are quite profound because this failure of compensation may signify impending severe respiratory failure. Serum electrolytes reveal a fall in [TCO2]. Acidosis tends to shift potassium out of cells in a rather complex manner (65), and renal potassium excretion tends to increase in many states of metabolic acidosis. Normal or increased serum potassium in the face of decreased total body potassium stores occurs commonly in cases of metabolic acidosis (Chapter 5) (65). Some states of metabolic acidosis are characterized by the retention of an organic anion generated in concert with HCO3− consumption (organic acidosis), whereas others are not (hyperchloremic). The screening of plasma for such organic anions is not practical on a routine, immediate basis; thus, a calculation performed on the serum electrolytes called the serum anion gap (SAG) is employed (66).
Serum Anion Gap
The SAG is a concept used in acid–base pathophysiology to infer whether an organic or mineral acidosis is present. We use the routine venous blood serum electrolytes to calculate the SAG (58):
Here we use TCO2 as an index of HCO3−. We define unmeasured cations (UCs) as cations that are not Na+ (e.g., K+, Mg2+, Ca2+) and unmeasured anions (UAs) as anions that are not Cl− or HCO3− (e.g., SO42−, H2PO4−, HPO42−, albumin, organic anions). Thus, electroneutrality demands that
Normally, the SAG is about 9 (6–12 mEq/L). Although the SAG is used routinely in the differential diagnosis of metabolic acidosis, we stress that it is a relative rather than absolute indicator of the underlying pathophysiology. Note that the maintenance of stoichiometry (i.e., 1 mEq increase in anion gap for every 1 mmol decrease in HCO3−) depends on the clearance mechanisms for the anion as well as the myriad factors that influence HCO3− concentration. Therefore, some organic acidoses may manifest trivial or even no increase in the anion gap, whereas some hyperchloremic acidoses may have coincidental increases in the anion gap. This must be kept in mind when evaluating the differential diagnosis of metabolic acidosis. However, a major increase in the anion gap (e.g., SAG >26 mEq/L) always implies the existence of an organic acidosis (67).
If we look at the changes in the SAG and compare it to changes in HCO3− concentrations, additional insights can be made. This is often referred to as the “Delta gap” = Delta SAG – Delta Serum [HCO3−] or the “Delta–Delta.” It takes advantage of the assumption that with a pure organic metabolic acidosis, the fall in HCO3− and the increase in SAG should be equal (with the disclaimers stated above). Ergo, if the Delta gap is very positive (i.e., >6), it suggests that there must be another process generating alkali. Conversely, if the Delta gap is very negative (i.e., <−6), it suggests that there is another source of acid that does not result in an increase in SAG, ergo a concomitant nonanion gap metabolic alkalosis.
Urine Anion Gap
To address a different problem, urinary electrolytes have been used to estimate the quantity of NH4+ in the urine, a measurement that has been difficult to develop into a routine clinical test. The concept is quite similar to that described for the SAG. In the urine, because of electroneutrality
When urine pH is <6, UA does not include appreciable amounts of HCO3−, but consists primarily of phosphate (H2PO4− more than HPO42−) and, to a lesser degree, sulfate (SO42−) and organic anions. UC is made up mostly of NH4+. Therefore, if we define the urinary anion gap (UAG) as
then we see that this is determined largely by the amount of NH4+ in the urine that holds true in clinical studies of metabolic acidosis (34). In contrast to the SAG, which is useful in many settings of clinical acid–base diagnosis and therapy, the UAG has a very narrow clinical application in the differentiation of renal from nonrenal causes of nonanion gap metabolic acidosis (68).
DIFFERENTIAL DIAGNOSIS OF METABOLIC ACIDOSIS
The differential diagnosis of metabolic acidosis generally is approached clinically by using the SAG. Those acidosis states associated with retention of an organic anion are classified as increased anion gap or simply anion gap metabolic acidosis. Those acidosis states not associated with retention of an organic anion are classified as nonanion gap or hyperchloremic metabolic acidosis (66). These disorders are listed in Table 3-1.
CAUSES OF HYPERCHLOREMIC METABOLIC ACIDOSIS
Gastrointestinal Loss of HCO3−
Diarrhea is the most common cause of hyperchloremic metabolic acidosis and always should be considered early in the differential diagnosis. The concentration of HCO3− in diarrheal fluid is generally greater than that of plasma. In the extreme case of cholera, patients may lose up to 20 L/day of fluid containing 30 to 50 mEq/L of HCO3− (69). However, hypovolemic shock likely will cause lactic acidosis and increase the anion gap in that situation. Ileostomy fluid is also rich in HCO3−, especially early after construction (70).
Normal Anion Gap (Hyperchloremic)
Increased Anion Gap (Organic)
Gastrointestinal Loss of HCO3−
Intestinal fistula or drainage
Anion exchange resins
Ingestion of CaCl2 or MgCl2
Increased Acid Production
Renal Loss of HCO3−
Renal tubular acidosis
Carbonic anhydrase inhibitors
Inborn errors of metabolism
Toxic alcohol ingestion
Failure of acid excretion
Recovery from ketoacidosis
Addition of HCl
Acute renal failure
Chronic renal failure
The diagnosis of diarrheal loss of HCO3−, however, may be difficult in the very young or very old (71). In the former case, the distinction between diarrhea and an underlying RTA is extremely important. The UAG may be very helpful in this setting. Patients with diarrhea as a cause of metabolic acidosis typically have a very negative UAG (i.e., urinary chloride exceeds the sum of urinary NA+ + K+ by >10 mEq/L) reflecting the presence of ample urinary NH4+ concentrations, whereas patients with all forms of distal RTA have positive UAGs reflecting the inadequate urinary NH4+ concentrations present (34).
Gastrointestinal Drainage and Fistulas
The succus entericus and pancreatic and biliary secretions are rich in HCO3− and poor in Cl−. The succus entericus has a daily volume of 600 to 700 mL but may be increased in disease states. Biliary secretions amount to >1 L/day of fluid, with [HCO3−] approaching 60 mmol/L. Pancreatic secretion may exceed 2 L/day, with [HCO3−] approaching 120 mmol/L. Therefore, it is not surprising that GI drainage or fistulas could cause significant metabolic acidosis (72).
The technique of simultaneous kidney/pancreas transplant (SPK) has been used for the treatment of type I diabetic patients with end-stage renal disease (73). When urinary drainage of the HCO3−-rich exocrine secretions of the pancreatic allograft is employed rather than enteric drainage, this addition of pancreatic exocrine fluid causes urinary loss because the bladder cannot absorb the pancreatic secretion. Thus, SPK with urinary drainage is often associated with significant normal anion gap metabolic acidosis (74,75). Many of these patients require sodium bicarbonate supplementation (as high as 100–150 mEq/day) on a chronic basis. The incidence of metabolic acidosis can be significantly lowered with enteric drainage compared to urinary drainage because the intestine can absorb the HCO3− (76).
Urinary Diversion to Bowel
Patients may require urinary diversion from normal egress through the bladder for a variety of reasons. Approaches to this include the creation of an ileal loop conduit or, less commonly, drainage of the ureters into the sigmoid colon (ureterosigmoidostomy) (77). Although metabolic acidosis can develop with both procedures, it is more severe with ureterosigmoidostomy. The pathophysiology for both situations is the bowel mucosal secretion of HCO3− in exchange for Cl− during water reabsorption, which may lead to significant HCO3− losses in the GI tract effluent (72). Newer reconstructive procedures including ileal neobladders, which minimize the time of contact between urine and bowel mucosa, have been successful in limiting HCO3− loss (78).
Chloride-containing Anion Exchange Resins
Cholestyramine is a nonabsorbable anion exchange resin used to bind bile acids in the gut for a variety of purposes, including the treatment of obstructive liver disease as well as hypercholesterolemia and the management of acute diarrhea in children (79). However, this resin has some affinity for HCO3− and may exchange Cl− for HCO3− across the bowel mucosa. In conditions of renal insufficiency where new HCO3− generation is impaired, in volume depletion or patients taking spironolactone, hyperchloremic metabolic acidosis has been reported (80).
Calcium or Magnesium Ingestion
The divalent cations, calcium or magnesium, are absorbed incompletely in the GI tract. If large amounts of these cations are ingested in soluble form (e.g., as the Cl2 salts), then the unabsorbed Ca2+ or Mg2+ reacts with HCO3−, which has been exchanged across the mucosa for Cl−, to form an insoluble salt. Thus, plasma HCO3− falls to a moderate degree (81).
Renal Loss of HCO3−
Renal Tubular Acidosis
RTA refers to a group of functional disorders characterized by impairment of renal HCO3− reabsorption and H+ excretion that is out of proportion to any reduction in GFR. In many cases, the RTAs exist in the presence of a completely normal GFR. Unfortunately, a nomenclature has evolved that confuses many experienced clinicians as well as trainees and students. We provide a pathophysiologic classification of these disorders while referring to this nomenclature. RTAs can be divided into those characterized by disturbed distal nephron function (i.e., impaired NAE) and those caused by impaired proximal HCO3− reabsorption (82). Distal RTAs are divided into those associated with hypokalemia (83) and those associated with hyperkalemia, which may be further subdivided into RTA caused by hypoaldosteronism and RTA characterized by a general distal tubular defect (84).
Proximal Renal Tubular Acidosis
Proximal RTA, also called type II RTA, is an uncommon but very interesting disorder (85,86). Basically, the acid–base disturbance is caused by impairment in proximal tubular reabsorption of HCO3−, the nephron site where 85% of HCO3− usually is reabsorbed. The delivery of HCO3−-rich fluid to distal nephron sites leads to substantial bicarbonaturia, when plasma levels of HCO3− are normal, as well as urinary losses of potassium and sodium. Thus, patients present with hypokalemia and hyperchloremic metabolic acidosis. When the plasma concentration of HCO3− is maintained at normal by administration of HCO3−, fractional HCO3− excretion (i.e., the fraction of filtered HCO3− that is excreted in the urine) exceeds 15%.
In physiologic terms, the apparent Tmax and PT for HCO3− are significantly reduced in patients with proximal RTA. However, once a level of plasma HCO3− is achieved that is below the patient’s PT for HCO3−, renal acid handling is normal. In other words, NAE equals dietary and endogenous acid production rates and the subject comes into a steady state of acid–base balance, albeit at a moderately reduced plasma HCO3− concentration (and a mild reduction in systemic pH). Because a steady state in acid handling is achieved, patients with proximal RTA have less severe acidosis as well as less nephrocalcinosis (which results from bone calcium mobilization from acidosis) than patients with distal RTAs (see the following).
The problem with proximal HCO3− reabsorption may occur independently but more commonly coexists with other defects in proximal nephron function, such as decreased reabsorption of glucose, amino acids, phosphate, and uric acid. The term Fanconi syndrome is employed when general proximal nephron function is disturbed (87). Patients with full-blown Fanconi syndrome may have severe osteomalacia and malnutrition in addition to the mild metabolic acidosis associated with proximal RTA (88). Proximal RTA may occur as a primary disorder and present in infancy or may be acquired in the course of other diseases or as a result of exposure to substances toxic to this nephron segment. A list of causes of proximal RTA is shown in Table 3-2. Treatment of this condition is approached by addressing the underlying cause, but if this is ineffective, administration of large amounts of HCO3− (10–15 mmol/kg/day) and potassium to compensate for ongoing potassium losses in the urine caused by the bicarbonaturia is necessary. This is necessary to avoid growth retardation in children and osteopenia, which may be produced by even mild degrees of acidemia (89).
Sickle cell disease/trait
Renal transplant rejection
Renal transplant rejection
Other interstitial disease
Medullary sponge kidney
Medullary cystic disease
Injury from kidney preservation
The distal RTAs are characterized primarily by impaired NAE, which is due, at least in part, to impaired NH4+ excretion. The central role of impaired NH4+ excretion in this disorder is highlighted by a recent clinical study in which all patients with either hypokalemic distal RTA (also called type I or classic distal RTA) or hyperkalemic distal RTA (previously referred to as type IV RTA), caused by either hypoaldosteronism or a generalized tubular defect, had a positive UAG reflecting decreased NH4+ excretion. How NH4+ excretion is impaired in this diverse set of clinical disorders is still incompletely understood (84,90).
Hypokalemic distal RTA has long been considered a disorder of the collecting duct in which the quantity of H+ secretion is inadequate to effect the necessary NAE for the subject to maintain acid–base balance. Clinically, patients with hypokalemic distal RTA present with hyperchloremic metabolic acidosis but are unable to acidify their urine (pH < 5.5 is commonly used) in response to an acid challenge. It must be stressed that the failure to acidify the urine does not fully explain the defect in NAE, which is primarily caused by an associated defect in NH4+ excretion (91). However, the failure to acidify the urine under conditions of systemic acidosis historically has been considered the clinical hallmark of hypokalemic distal RTA. The physiologic mechanisms for this impaired acidification have been a topic of interest for some time and are summarized with clinical examples in Table 3-3. Basically, four mechanisms have been suggested for impaired acidification by the distal nephron: (a) back leak through a leaky epithelium; (b) pump failure, where the H+ ATPase cannot pump sufficient amounts of H+; (c) voltage defect, where a favorable transepithelial voltage cannot be generated (e.g., decreased sodium delivery to the distal nephron or decreased sodium reabsorption in the distal nephron); or (d) rate defect/NH4+ defect, where urinary pH is reduced but NH4+ excretion and NAE cannot be increased to normal amounts. Hypokalemic distal RTA appears to be caused by either back leak or pump failure. Patients may have an isolated defect in the H+/K+ ATPase or the vacuolar ATPase (92). Hyperkalemic distal RTAs are probably caused by either voltage defect or rate defect/NH4+ defect (84).
A number of physiologic maneuvers have been used to examine these mechanisms clinically. The first and simplest test is that of a metered pH (i.e., using a pH meter rather than a dipstick) performed on urine collected under oil. If the subject is already acidemic (e.g., arterial pH < 7.35), then there is no need for ammonium chloride loading. In some cases, patients are able to maintain a normal plasma HCO3− concentration and systemic pH under most circumstances, but do not respond normally to increases in acid generation by increasing NAE. Recent studies have identified mutations in the vacuolar ATPase B1 subunit most likely responsible for this defect (93). This is called an incomplete distal RTA. If an incomplete distal RTA is suspected, ammonium chloride is administered to induce a mild case of metabolic acidosis. This test basically screens for back leak, pump failure, or voltage defects. An alternate method, the furosemide–fludrocortisone test, is a better-tolerated test for urine acidification (94). Furosemide increases distal Na+ delivery, and the mineralocorticoid fludrocortisone increases distal H+ secretion. Normal subjects will acidify urine to pH < 5.3 by 3 or 4 hours, whereas patients with distal RTA fail to do so. Infusion of sodium sulfate or sodium phosphate increases distal sodium delivery. The failure to lower urine pH after these maneuvers suggests pump failure or impaired voltage owing to inadequate distal sodium reabsorption. Another maneuver is to determine the urine to blood PCO2 gradient when the patient has bicarbonaturia (urine [HCO3−] >100 mM) induced by HCO3− administration. Under conditions where bicarbonaturia is induced, H+ secreted into the collecting duct lumen will combine with HCO3− and form H2CO3. Because carbonic anhydrase is absent in the lumen of this segment (as well as the bladder), conversion to CO2 and water is slow and occurs largely in the urinary collecting system (i.e., renal pelvis, ureters, and urinary bladder), where the surface area for CO2 absorption is small. This CO2 essentially is trapped, and, when normalized for the blood PCO2 (i.e., the difference between urine and blood), is a marker for the rate of distal H+ secretion. Patients with back leak or pump failure generally have a small difference between urine and blood PCO2 (<20 torr).
Rate defect/NH4+ defect
Hypokalemic distal RTA may be primary or associated with other diseases, most commonly, Sjögren syndrome (95,96) and toxin exposures. These toxins may vary from wasp sting (97) to toluene inhalation, the latter capable of causing several forms of acidosis (98). A list of causes of hypokalemic distal RTA appears in Table 3-2. Some of the causes also may result in a hyperkalemic distal RTA because of a generalized tubular defect (84). Urinary obstruction and some of the autoimmune disorders are such examples (99). Perhaps the best understood cause of hypokalemic distal RTA is that due to amphotericin toxicity, which results (at least experimentally) in acidification failure owing to back leak of normally secreted H+. Hypokalemic distal RTA usually occurs in young children in its primary form. The children never achieve a steady state of acid–base balance; therefore, they typically present with extremely severe metabolic acidosis, growth retardation, nephrocalcinosis, and nephrolithiasis (100). Hypokalemia, which usually is present, is actually caused by the associated sodium depletion and stimulation of the renin–angiotensin–aldosterone axis. Therefore, renal potassium losses actually decrease considerably when appropriate therapy with sodium bicarbonate is instituted. This is quite different from patients with proximal RTA, where urinary potassium losses increase considerably during therapy because of the bicarbonaturia-associated urinary K losses. Another contrasting point between proximal RTA and hypokalemic distal RTA is the amount of alkali therapy needed. Once the acute acidosis is corrected, patients with hypokalemic distal RTA only need enough alkali to account for the amount of acid generated from diet and metabolism; therefore, 1 to 3 mmol/kg/day generally is sufficient.
Hyperkalemic distal RTA from hypoaldosteronism occurs in several settings summarized in Table 3-2. Best understood is the case of either selective aldosterone deficiency or complete adrenal insufficiency. Probably the most common form of RTA is the hyporeninemic hypoaldosteronism often seen in patients with diabetic nephropathy. In patients with this form of RTA, urinary acidification as assessed by urine pH appears normal, but the patients are unable to raise NAE to appropriate levels. The defect, at least in some of these individuals, can be traced to impaired NH4+ synthesis in the proximal nephron, resulting directly from the hyperkalemia. Treatment of the hyperkalemia in some individuals with this disorder is sufficient to correct the disturbance in NAE. In patients with pure primary aldosterone deficiency, replacement of physiologic amounts of mineralocorticoid results in correction of the disturbance in acid–base metabolism and is both logical and appropriate therapy. However, in patients with the hyporeninemic hypoaldosteronism form, the renal defect requires pharmacologic amounts of mineralocorticoid (i.e., 5–10 times the usual physiologic dose) for efficacy. Moreover, the use of mineralocorticoid in this setting may be contraindicated because these patients often have mild renal insufficiency and tend to be total body sodium expanded rather than depleted (as is the case in the pure hypoaldosteronism form). Treatment of the hyperkalemia by increasing renal K excretion (e.g., with loop diuretics) or K excretion through the GI tract with potassium-binding resins (Kayexalate) may be the preferred approach in patients with the hyporeninemic hypoaldosteronism form.
Hyperkalemic distal RTA from a generalized tubular defect is considerably more common than either classic distal or proximal RTA. A list of causes appears in Table 3-2. Urinary obstruction may be the most common and important cause of this form of distal RTA. Other important causes in selected populations include cyclosporine nephrotoxicity and allograft rejection in the renal transplant patient, sickle cell nephropathy in patients homozygous and occasionally heterozygous for the sickle cell gene, and many autoimmune disorders such as lupus nephritis and Sjögren syndrome. Urinary acidification is impaired similarly to the hypokalemic distal RTA patients. Also in contrast to the hypoaldosteronism form, hyperkalemia plays a less significant role in the genesis of impaired NH4+ excretion, which is tied directly to the impaired distal nephron function.
Carbonic Anhydrase Inhibitors
Carbonic anhydrase inhibitors such as acetazolamide inhibit both proximal tubular luminal brush border and cellular carbonic anhydrase. The net effect is a pattern of impaired HCO3− reabsorption similar to that of proximal RTA. These drugs are commonly used topically to treat glaucoma, but their use may be complicated by systemic effects such as hyperchloremic metabolic acidosis (101). Topiramate is an antiseizure medication widely used in children that causes a mild to moderate proximal RTA by inhibiting carbonic anhydrase (102). The use of this drug has been associated with nephrolithiasis in both adults and children because of the resultant hypocitraturia, hypercalciuria, and elevated urine pH (103).
Hypoaldosteronism is associated with a hyperkalemic distal RTA. This may be produced by pharmacologic antagonism of aldosterone action or impaired aldosterone secretion. Impaired aldosterone secretion may be caused by hyporeninemia (e.g., the hyporeninemic hypoaldosteronism associated with diabetes mellitus) or may be part of adrenal insufficiency (e.g., Addison disease). With hyporeninemic hypoaldosteronism, some have suggested that the disorder is, at least in part, an adrenal disorder because plasma potassium concentrations that typically induce aldosterone secretion do not in this disorder. However, permissive amounts of angiotensin II are necessary to allow potassium to be an effective aldosterone secretagogue (104,105).
Potassium-sparing diuretics, which either block aldosterone action (e.g., spironolactone, eplerenone) or impair distal nephron sodium reabsorption (e.g., amiloride, triamterene), also may produce a hyperchloremic acidosis in concert with hyperkalemia (106,107). The observation that aldosterone antagonists may ameliorate the progression of congestive heart failure (108) has led to more widespread use, but careful monitoring of plasma potassium is necessary (109).
Miscellaneous Causes of Hyperchloremic Acidosis
Recovery from Ketoacidosis
Although diabetic ketoacidosis (DKA) is one of the best described forms of increased anion gap metabolic acidosis, many patients during recovery from DKA may eliminate the organic anions through renal clearance faster than their acidosis corrects, leaving them with a nonanion gap or hyperchloremic metabolic acidosis (110). This also may occur in patients who drink enough to avoid volume depletion and consequent fall in GFR (111).
The rapid expansion of ECF volume with fluids that do not contain HCO3− leads to a dilution of HCO3− and mild metabolic acidosis. The fall in HCO3− produced in this manner is typically quite small (e.g., 10%) and usually is corrected fairly rapidly by renal generation of HCO3− (i.e., by renal correction) (112). This is also expected in the setting of therapeutic plasma exchange where large amounts of albumin are rapidly administered (113).
Addition of HCl
Administration of HCl or congeners (e.g., ammonium chloride or lysine chloride) rapidly consumes an HCO3− molecule without generating an organic anion, thus causing hyperchloremic metabolic acidosis (114).
Amino acid infusions without concomitant administration of alkali (or alkali-generating precursors) may produce hyperchloremic metabolic acidosis in a manner similar to addition of HCl. This problem can be avoided by replacing the chloride salt of these amino acids with an acetate salt, which is metabolized to HCO3− (115).
Ingested elemental sulfur or sulfur released during metabolism of sulfur-containing amino acids (e.g., methionine or cysteine) is oxidized to sulfate with accompanying H+ production. Sulfate is excreted rapidly by the kidneys, usually accompanied by sodium, whereas the excretion of H+ produced by sulfur metabolism lags, resulting in a hyperchloremic metabolic acidosis. A dietary intake rich in sodium compared to potassium and excessive consumption of sulfur-containing amino acids is a common feature of Western diets. Ingestion of 40 to 50 g/day of flowers of sulfur for several days, a folk remedy for constipation, has also produced profound hyperchloremic metabolic acidosis (116).
CAUSES OF INCREASED ANION GAP (ORGANIC METABOLIC ACIDOSIS)
Organic Acidosis Resulting from Increased Acid Production
Lactic acidosis is an extensively studied organic acidosis. Causes of lactic acidosis are summarized in Table 3-4. Lactic acid is the final product of mammalian anaerobic metabolism. In general, aerobic tissues metabolize carbohydrates to pyruvate, which then undergoes oxidative metabolism within mitochondria. This oxidative metabolism regenerates nicotinamide adenine dinucleotide (NAD)+ consumed at a more proximal site in the glycolytic pathway. When tissues must perform anaerobic glycolysis to regenerate this NAD+, the net effect is to generate lactic acid from carbohydrates and thus generate H+. Under normal conditions in humans, relatively small amounts of lactate, specifically the L-isomer, are formed during normal metabolism and are metabolized by the liver, maintaining relatively low plasma and urine levels of this metabolite. Lactic acidosis may develop under pathologic conditions associated with either local or systemic decreases in oxygen delivery (type A), impairments in oxidative metabolism (type B), or impaired hepatic clearance (117,118).
The diagnosis of lactic acidosis must be considered in all forms of metabolic acidosis associated with an increased anion gap, particularly those cases occurring in these clinical circumstances. Determination of the serum or plasma lactate level may confirm this diagnosis, although many clinical laboratories may not provide this information on an emergency basis (112). In cases of D-lactic acidosis (e.g., seen with blind intestinal loops colonized with D-lactate–producing organisms), the usual measurement of lactate performed in clinical laboratories using an enzymatic reaction does not detect this D-isomer. Nonroutine measurement techniques such as 1H NMR spectroscopy (which does not distinguish between D and L forms) or specific measurement of the D form with the appropriate enzymatic analysis may be necessary to document elevations of D-lactate in this unusual clinical circumstance (119,120).
Treatment of lactic acidosis must be directed at the underlying pathophysiology. Although the degree of acidemia in this setting may become deleterious in its own right, therapy with NaHCO3− to directly address the metabolic acidosis has not been found to be effective clinically (121) and is actually deleterious in several experimental models (50,122–124). This issue remains quite controversial at this time (125).
DKA results from insufficient insulin to metabolize glucose and excess glucagon, which generates short-chain fatty ketoacids, specifically β-hydroxybutyric and acetoacetic acids. These ketoacids are both relatively strong acids that dissociate almost completely at physiologic pH into H+ and the keto-anions and cause an anion gap metabolic acidosis. Interestingly, the amount of insulin needed for catabolism of short-chain fatty acids is significantly less than that necessary for glucose homeostasis. Thus, DKA is a common presentation in patients with insulin-dependent diabetes mellitus (111). However, DKA also occurs in patients with non–insulin-dependent diabetes mellitus (126). In addition, patients with non–insulin-dependent diabetes mellitus may present with marked increases in serum glucose concentrations without ketosis (127) (e.g., nonketotic hyperglycemic hyperosmolar coma). A newer class of hypoglycemic agents, the sodium–glucose cotransporter inhibitor, may predispose to ketone bodies generation and acidosis, especially at time of stress with relatively low glucose concentration (128).
Primary Decrease in Tissue Oxygenation
Excessive Energy Expenditures
Deranged Oxidative Metabolism
Intoxication (e.g., ethanol, iron, isoniazid, carbon monoxide, strychnine)
Impaired Lactate Clearance