Você está na página 1de 8

Nephrol Dial Transplant (2015) 30: 1104–1111

doi: 10.1093/ndt/gfu289
Advance Access publication 11 September 2014

Full Review

Acid–base disturbances in intensive care patients: etiology,


pathophysiology and treatment

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


Mohammed Al-Jaghbeer and John A. Kellum
Center for Critical Care Nephrology, CRISMA Center, Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, PA, USA

Correspondence and offprint requests to: John A. Kellum; E-mail: kellumja@ccm.upmc.edu

case, either approach may be used successfully by the skilled


A B S T R AC T practitioner.
Acid–base disturbances are very common in critically ill and
injured patients as well as contribute significantly to morbidity
M E TA B O L I C AC I D – B A S E D I S O R D E R S
and mortality. An understanding of the pathophysiology of
these disorders is vital to their proper management. This Metabolic acidosis can arise when the concentration of
review will discuss the etiology, pathophysiology and treat- organic anions (e.g. lactate or β-hydroxybutyrate) increases,
ment of acid–base disturbances in intensive care patients— when there is a loss of sodium bicarbonate (NaHCO3; e.g. due
with particular attention to evidence from recent studies to diarrhea or renal tubular acidosis) or there is a gain of ex-
examining the effects of fluid resuscitation on acid–base and ogenous anions (e.g. iatrogenic acidosis or poisonings). Meta-
its consequences. bolic alkaloses occur when there is loss of strong anions in
Keywords: acid–base physiology, acidosis, alkalosis, anion excess of strong cations (e.g. vomiting and diuretics), or,
gap, strong ion difference rarely, by administration of strong cations in excess of strong
anions (e.g. transfusion of large volumes of banked blood con-
taining sodium citrate).

INTRODUCTION
M E TA B O L I C AC I D O S I S
The modern intensive care unit is a place where complex acid–
base and electrolyte disorders are common, with one study, Traditionally, metabolic acidoses are categorized according to
showing that 64% of critically ill patients have acute metabolic the presence or absence of unmeasured anions, inferred by cal-
acidosis [1]. Although it is generally believed that most cases culating the anion gap (AG). The differential diagnosis for a
of acid–base derangement are mild and self-limiting, extremes ‘positive-AG’ acidosis is summarized in Table 1. Non-AG
of blood pH in either direction, especially when happening acidoses can be divided into three types: renal, gastrointestinal
quickly, can have significant multiorgan consequences. Ad- and iatrogenic.
vances in evaluating acid–base balance have helped in under- The potential effects of metabolic acidosis and alkalosis on
standing the impact of fluids in the critically ill [2]. In this vital organ function are presented in Table 2. Metabolic and re-
review, we will discuss common causes of acid–base abnormal- spiratory acidosis may have different implications with respect
ities in critically ill patients, and examine the evidence that to survival, an observation that suggests that the underlying
these conditions result in harm to patients and briefly consider disorder is perhaps more important than the absolute degree
various management strategies. We will discuss these disorders of acidemia [3]. If metabolic acidemia is to be treated, con-
using concepts from traditional and (more controversial) sideration should be given to the likely duration of the disorder.
physical–chemical perspectives. It is our belief that these ap- In situations where the underlying cause is readily reversible
proaches are complementary not contradictory and in any (e.g. diabetic ketoacidosis), facilitating respiratory compensation

© The Author 2014. Published by Oxford University Press 1104


on behalf of ERA-EDTA. All rights reserved.
Table 1. Causes of a metabolic acidosis added (or equal amounts of strong acid removed) to return the
Increased anion gap equilibrium to normal. If one uses the physical–chemical ap-
Renal failure proach [4] (our preference), this change can be viewed as a
Ketoacidosis change in strong ion difference, the difference between com-
Diabetic
Alcoholic
pletely dissociated anions and cations. So, for example, the
Starvation plasma [Na+] would have to increase by 10 mEq/L for NaHCO3
Metabolic errors administration to completely repair the acidosis.
Lactic acidosis NaHCO3 administration is associated with certain disad-
Toxins vantages. Large (hypertonic) doses given rapidly may lead to
Methanol
Ethylene glycol
hypotension [5] and have the potential to cause a sudden and
Salicylates marked increase in PaCO2 [6]. Accordingly, it is important to
Paraldehyde 5-oxoproline (pyroglutamic acid) assess the patient’s ventilatory status before NaHCO3 is admi-
Sepsis nistered, particularly in the absence of mechanical ventilation.
Hyperchloremic (non-anion gap) NaHCO3 infusion also affects circulating [K+] and [Ca2+],

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


Renal tubular acidosis
Gastrointestinal
which need to be monitored closely.
Diarrhea Tromethamine (Tris buffer or Tham) is an organic buffer
Small bowel/pancreatic drainage that readily penetrates cells [7]. It is a weak base ( pK 7.9);
Iatrogenic modern formulations do not alter the strong ion difference
Parenteral nutrition nor do they affect plasma [Na+]. Accordingly, tromethamine
Saline
Carbonic anhydrase inhibitors
is sometimes used when administration of NaHCO3 is contra-
Anion exchange resins indicated because of hypernatremia. This agent has been avail-
able since the 1960s, but limited data are available on its use in
humans with acid–base disorders.
Table 2. Clinical effects of metabolic acid–base disorders

Metabolic acidosis Metabolic alkalosis


Cardiovascular Cardiovascular
Decreased inotropy (Ca2+ entry) AG AND STRONG ION GAP

FULL REVIEW
Decreased inotropy
Conduction defects Altered coronary blood flow
Arterial vasodilatation Digoxin toxicity The AG is estimated from the differences between the serum
Venous vasoconstriction Neuromuscular cations (Na+ and K+) and anions (Cl− and HCO 3 ). Normally,
Oxygen delivery Neuromuscular excitability this ‘gap’ is made up by albumin and, to a lesser extent, phos-
Decreased oxy-Hb binding Encephalopathy seizures
phate. Sulfate and lactate also contribute a small amount, nor-
Decreased 2,3-BPG (late) Metabolic effects
Neuromuscular Hypokalemia mally <2 mEq/L. Plasma proteins other than albumin in the
Respiratory depression Hypocalcemia aggregate tend to be neutral, except in rare cases of abnormal
Decreased sensorium Hypophosphatemia paraproteins, such as multiple myeloma [8].
Metabolism Impaired enzyme function In practice, the AG is calculated as follows:
Protein wasting Oxygen delivery
Bone demineralization Increased oxy-Hb affinity AG ¼ ð½Naþ  þ ½Kþ Þ  ð½Cl  þ ½HCO
3 Þ
Catecholamine, PTH and Increased 2,3-BPG (delayed)
aldosterone stimulation Owing to its low and narrow extracellular concentration range,
Insulin resistance K+ is often omitted from the calculation. The normal value for
Free radical formation
AG is 12 ± 4 (if [K+] is considered) or 8 ± 4 mEq/L (if [K+] is
Gastrointestinal
Emesis not considered).
Gut barrier dysfunction If the AG is increased, the explanation almost invariably
Electrolytes will be found among five disorders: ketosis, lactic acidosis, poi-
Hyperkalemia soning, renal failure or sepsis [9]. However, several conditions
Hypercalcemia
prevalent among patients with critical illness can alter the ac-
Hyperuricemia
curacy of AG estimation [10]. Hypoalbuminemia decreases
BPG, biphosphoglycerate; PTH, parathyroid hormone.
the AG and it has been recommended to ‘correct’ the AG by
decreasing it by 2.5–3 meq/L for every 1 g/dL decrease in
serum albumin concentration [11]. Respiratory and metabolic
should be considered as a first-line intervention. However, if the alkaloses are associated with an increase of up to 3–10 mEq/L
disorder is likely to be more chronic (e.g. renal failure), therapy in the apparent AG. The basis for this effect is enhanced lactate
aimed at restoring metabolic acid–base balance is indicated. In production (from stimulated phosphofructokinase enzymatic
all cases, the therapeutic target can be estimated initially from activity), reduction in the concentration of ionized weak acids
the standard base excess (SBE). The SBE corresponds to the (A−) and, possibly, the additional effect of dehydration.
amount of strong acid or base that must be added in order to Other factors that can increase the AG are low Mg2+ con-
restore the pH to 7.4, assuming a PCO2 of 40 mmHg. Thus, if centration, administration of the sodium salts of poorly reab-
SBE is −10 mEq/L, 10 mEq/L of strong base would need to be sorbable anions (such as, beta-lactam antibiotics) [12] and

Acid–base disturbances 1105


certain parenteral nutrition formulations, such as those con- variety of mechanisms from inflammation to increased metab-
taining acetate. Citrate-based anticoagulants rarely can have olism to decreased clearance [22–26].
the same effect after multiple blood transfusions [13]. In shock, lactic acid was classically viewed as being pro-
However, these rare causes do not increase the AG very much duced from the musculature and the gut as a consequence of
and are usually easy to identify. tissue hypoxia. Moreover, the proportion of lactic acid produc-
Additional doubt has been cast on the diagnostic value of tion was believed to correlate with the severity of shock as
the AG, however [10, 14]. The primary problem with the AG reflected by the amount of oxygen debt and magnitude of
is its reliance on the use of a ‘normal’ range that depends on hypoperfusion. However, this view has been challenged by
normal circulating levels of albumin and to a lesser extent showing that, during endotoxemia, muscle can actually
phosphate. Plasma concentrations of albumin or phosphate consume lactate [22]. In addition, the lungs are considered a
are often grossly abnormal in patients with critical illness, source of lactic acid during acute lung injury [23]. Another ex-
leading to changes in the ‘normal’ range for the AG. Moreover, planation of elevated lactate has emerged that attributes ele-
because these anions are not strong anions, their charge is vated lactate to be due to increased metabolic activity,

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


affected by pH. glycolysis and pyruvate production reflecting disease severity
An alternative to the traditional AG is the strong ion gap rather than an oxygen debt [21, 24] in addition to decreased
(SIG). By definition, the strong ion difference must be equal lactate clearance in the critically ill [25, 26]. Finally, numerous
and opposite to the negative charges contributed by [A−] and drugs are associated with lactic acidosis including metformin,
total CO2. The sum of the charges from [A−] and total CO2 zidovudine and isoniazid. Administration of epinephrine may
concentration has been termed ‘effective strong ion difference’ be a common cause of lactic acidosis in patients with critical
[15]. The ‘apparent strong ion difference’ is obtained by illness [27] presumably by stimulating cellular metabolism
measurement of each individual ion. Both the apparent and (e.g. increased glycolysis in skeletal muscle).
effective strong ion differences should be equal; otherwise, un- When lactic acidosis is suspected, it is best to measure it
measured ions must exist. If the apparent strong ion difference directly. Monitoring blood pH, base deficit or AG may fail to
is greater than effective strong ion difference, these ions are detect hyperlactatemia as they can be affected by ventilator
anions and if the apparent strong ion difference is less than status, renal failure and other complex acid–base disorders
the effective strong ion difference, the unmeasured ions are [28]. A lactate-to-pyruvate ratio greater than 25:1 is considered
cations. This difference has been termed the SIG to distinguish to be evidence of anaerobic metabolism [29]. This approach
FULL REVIEW

it from the AG [16]. Unlike the AG, the SIG is normally zero makes biochemical sense, because pyruvate is reduced to
and does not change with changes in pH or albumin concen- lactate during anaerobic metabolism, thereby increasing the
tration (its primary advantage over the AG). However, critical lactate-to-pyruvate ratio. Unfortunately, pyruvate is very un-
illness itself appears to result in changes in acid–base balance stable in solution and, therefore, is difficult to measure accur-
that alter the SIG and the AG through mechanisms not clearly ately in the clinical setting, greatly reducing the clinical utility
identified [17]. of lactate/pyruvate determinations.
Treatment of lactic acidosis is generally supportive and spe-
cific therapy depends on the underlying etiology. The use of
NaHCO3 is controversial and of unproven value [30]. At best,
P O S I T I V E - A G AC I D O S E S it can be viewed as a temporizing measure. The use of renal re-
placement therapy (RRT) in conjunction with NaHCO3 is
Lactic acidosis likely to be more effective, but even this approach will fail if
Lactic acidosis is a common etiology of metabolic acidosis in the underlying condition is not reversed. Lactate is easily
the ICU [18]. Blood concentration of lactate has been shown to removed by RRT; however, use of RRT is rarely the primary
correlate with outcome in patients with hemorrhage [19] and therapy.
sepsis [20]. Arterial blood lactate concentration reflects a
balance between production and metabolism. Lactate is pro- Ketoacidosis
duced in the cytoplasm according to the following reaction: Ketones are formed by beta-oxidation of fatty acids, a
lactatedehydrogenase
process that is inhibited by insulin. In insulin-deficient states,
Pyruvate þ NADH þ Hþ ! lactate ketone bodies (acetone, β-hydroxybutyrate and acetoacetate)
þ NAD þ increase substantially because elevated blood glucose concen-
trations promote an osmotic diuresis, leading to intravascular
This reaction favors lactate formation, yielding a 10-fold lactate/ volume contraction. This state is associated with elevated cir-
pyruvate ratio, and depends on the NADH/NAD+ ratio. Under culating cortisol and catecholamine levels, which further sti-
physiologic conditions, lactate is mainly produced by the mulates free fatty acid production [31]. In addition, increased
muscles (25%), skin (25%), brain (20%), intestine (10%) and glucagon levels, relative to insulin levels, decreases intracellular
red blood cells (20%) [21]. Traditionally, lactic acidosis has sub- concentrations of malonyl co-enzyme A and increases the
divided into a ‘type A’, in which the mechanism is thought to activity of carnitine palmityl acyl transferase, effects that
be tissue hypoxia, and ‘type B’, in which there is no hypoxia. promote ketogenesis.
However, this distinction is largely artificial. Some disorders, Ketoacidosis may result from diabetes or excessive alcohol
such as sepsis, may be associated with lactic acidosis due to a consumption. The diagnosis is established by measuring

1106 M. Al-Jaghbeer and J.A. Kellum


serum ketone levels. However, it is important to understand normal response by the kidney is to increase Cl− excretion. If
that the nitroprusside reaction only measures acetone and the kidney fails to increase Cl− excretion appropriately, then
acetoacetate, and not β-hydroxybutyrate. There is a risk of impaired renal function is at least part of the problem, causing
confusion during treatment as ketone levels, as measured by acidosis. Extrarenal causes of hyperchloremic acidosis are
nitroprusside reaction, appear to increase as β-hydroxybutyrate exogenous Cl− loads (iatrogenic acidosis) or loss of cations
is converted to acetoacetate as acidosis is clearing. Hence, it is from the lower gastrointestinal tract without proportional
better to monitor therapy by measuring blood pH and AG losses of Cl−.
than by serum ketones.
Treatment of diabetic ketoacidosis includes infusing insulin Renal tubular acidosis
and large amounts of fluid; 0.9% saline is usually recommended Examination of the urine and plasma electrolytes and pH
but the predictable consequence is that patients will develop and calculation of the urine apparent strong ion difference
hyperchloremic metabolic acidosis (see below). Potassium allow one to correctly diagnose most cases of renal tubular
replacement is often required as well. This problem is further acidosis [39]. However, caution must be exercised when the
exacerbated Cl− reabsorption which increases as ketones are plasma pH is >7.35, because urinary Cl− excretion is normally

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


excreted in the urine. An increase in the tubular Na+ load pro- decreased when pH is this high. In such circumstances, it may
duces electrical–chemical forces favoring Cl− reabsorption [32]. be necessary to infuse sodium sulfate or furosemide. These
Nevertheless, administration of NaHCO3 is rarely necessary agents stimulate Cl− and K+ excretion and can be used to
and should be avoided except in extreme cases [33]. Again, unmask the defect and probe K+ secretory capacity.
RRT can be used, especially if renal function is impaired, as Type IV renal tubular acidosis is caused by aldosterone de-
ketones are easily removed by RRT; however the primary goal ficiency or resistance. These disorders are diagnosed by the
of therapy should remain focused on the underlying metabolic presence of high serum [K+] and low urine pH (<5.5). Note
derangement. that occasionally partial urinary tract obstruction may present
the same way. Treatment is usually most effective if the cause
Renal failure can be removed; most commonly drugs such as nonsteroidal
Renal failure, especially when chronic, leads to accumula- anti-inflammatory agents, heparin or potassium-sparing diu-
tion of sulfates and other acids, widening the AG, although retics, are responsible. Particularly in diabetics, the condition
this increase usually is not large [34]. Similarly, uncomplicated may require management with loop diuretics in conjunction

FULL REVIEW
renal failure rarely produces severe acidosis, except when ac- with dietary modification. Occasionally, mineralocorticoid re-
companied by a high rate of acid generation, such as occurs placement is required.
with a highly catabolic state [35]. In all cases, the strong ion
difference is decreased and remains so unless some therapy is Gastrointestinal acidosis
provided. Hemodialysis removes sulfate and other ions and Fluid secreted into the gut lumen contains higher amounts
allows normal Na+ and Cl− balance to be restored, thus return- of Na+ than Cl−. Large losses of these fluids, particularly if
ing the strong ion difference to normal (or near normal). volume is replaced with fluids containing equal amounts of
However, patients not yet requiring dialysis and those who are Na+ and Cl−, result in a decrease in the plasma Na+ concentra-
between treatments often require some other therapy to in- tion relative to the Cl− concentration and a decrease in strong
crease the strong ion difference. NaHCO3 can be used in ion difference. Such a scenario can be avoided if lactated
caution as long as the plasma Na+ concentration is not already Ringer’s solution is used instead of normal saline to replace
elevated. gastrointestinal losses.

Other and unknown causes Iatrogenic acidosis


Even when very careful methods are applied, using the SIG Two of the most common causes of a hyperchloremic
or similar strategies, unmeasured anions have been detected in metabolic acidosis are iatrogenic and both are due to adminis-
the blood of patients with sepsis [36] and liver disease [37]. tration of Cl−. Modern parenteral nutrition formulas contain
Furthermore, unknown cations also appear in the blood of weak anions, such as acetate, in addition to Cl−. The propor-
some critically ill patients [38]. The significance of these find- tions of each anion can be adjusted, depending on the acid–
ings remains to be determined. Various toxins that can cause base status of the patient. If an insufficient amount of weak
metabolic acidosis are listed in Table 1. A complete discussion anion is provided, the plasma Cl− concentration increases and
of these is beyond the scope of this review. acidosis results. A similar condition can arise when saline is
used for fluid resuscitation. While usually transient and of
minor importance in healthy subjects, in critically ill and
N O N - A G ( H Y P E R C H LO R E M I C ) AC I D O S E S injured patients saline-induced acidosis is an important
problem [40].
Hyperchloremic metabolic acidosis occurs as a result of either Administration of saline causes acidosis because this solu-
the increase in [Cl−] relative to strong cations, especially Na+, tion contains equal amounts of Na+ and Cl−, whereas the
or the loss of cations with retention of Cl−. As seen in Table 1, normal Na+ concentration in plasma is 35–45 mEq/L greater
these disorders can be separated by history and by measure- than the normal Cl− concentration. Administration of 0.9%
ment of urinary Cl− concentration. When acidosis occurs, the saline increases the Cl− concentration relatively more than the

Acid–base disturbances 1107


Na+ concentration. Many critically ill patients have a signifi- Table 3. Differential diagnosis of a metabolic alkalosis (an increased
cantly lower strong ion difference than do healthy individuals, strong ion difference)
even when there is no evidence of a metabolic acid–base de- Chloride loss <sodium
rangement [41]. One alternative to using normal saline to re- Chloride-responsive (urine Cl− concentration <10 mmol/L)
Gastrointestinal losses
suscitate patients is to use Ringer’s lactate solution. This fluid Vomiting
contains a more physiologic difference between [Na+] and Gastric drainage
[Cl−], and thus, its strong ion difference is closer to normal Chloride wasting diarrhea (villous adenoma)
(28 mEq/L compared with 0 mEq/L for normal saline). Post-diuretic use
Morgan et al. [42] recently showed that a solution with a Post-hypercapnia
Chloride-unresponsive (urine Cl− concentration >20 mmol/L)
strong ion difference of ∼24 mEq/L results in a neutral effect Mineralocorticoid excess
on the pH as blood is progressively diluted. Primary hyperaldosteronism (Conn’s syndrome)
The clinical impact of iatrogenic hyperchloremic acidosis Secondary hyperaldosteronism
has been in debate [43]. However, mounting evidence sup- Cushing’s syndrome
Liddle’s syndrome

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


ports the position that this adverse biochemical effect results
Bartter’s syndrome
in important adverse clinical events, especially for the kidney. Exogenous corticoids
Shaw et al. [44] examined the records of >30 000 patients Excessive licorice intake
treated with 0.9% saline for perioperative fluid management in Ongoing diuretic use
the setting of open abdominal surgery and compared them Exogenous sodium load (>chloride)
with just under 1000 patients treated with a balanced crystal- Sodium salt administration (acetate, citrate)
Massive blood transfusions
loid solution. Both overall and in a matched patient analysis, Parenteral nutrition
adverse effects including the use of dialysis were more frequent Plasma volume expanders
in the saline-treated patients. In a single-center, before–after Sodium lactate (Ringer’s solution)
study in Australia, Yunos et al. [45] showed that restricting the Other
use of ‘chloride-rich’ solutions including saline was associated Severe deficiency of intracellular cations
Magnesium, potassium
with significant reductions in the rates of acute kidney injury
and use of dialysis. However, it should be noted that this was
not a randomized trial and therefore susceptible to bias from
FULL REVIEW

secular trends. alkalosis more rapidly. Hypokalemia is also a known etiology


Given this evidence, we recommend avoiding saline for of metabolic alkalosis [47]. Proposed mechanisms are in-
fluid resuscitation, especially when large volumes required or creased H+ secretion in both proximal and distal nephron, in-
in acidemic patients, and favor use of more balanced solutions. creased bicarbonate reabsorption at proximal tubule and
We recognize that lactated Ringer’s solution is not ideal, given stimulating ammoniagenesis [48]. Since hypokalemia will
its slight hypotonicity, but the toxicity of saline appears to be a exacerbate metabolic acidosis, management should include
more significant concern. repletion of potassium.

M E TA B O L I C A L K A LO S I S R E S P I R AT O R Y AC I D – B A S E D I S O R D E R S

Metabolic alkalosis occurs as a result of an increase in strong Respiratory disorders are easier to diagnose and treat than
ion difference or a decrease in ATOT. These changes can occur metabolic disorders. Normally, alveolar ventilation is adjusted
secondary to the loss of anions (e.g. Cl− from the stomach and to maintain PaCO2 between 35 and 45 mmHg. When alveolar
albumin from the plasma) or the retention of cations (rare). ventilation is increased or decreased out of proportion to CO2
Sometimes, the loss of Cl− is temporary and can be treated ef- production, a respiratory acid–base disorder exists.
fectively by replacing the anion; metabolic alkalosis in this cat- Normal CO2 production by the body (∼220 mL/min) is
egory is said to be ‘chloride-responsive’. Indeed, the term equivalent to 15 000 mM/day of carbonic acid [49]. This
‘chloride depletion’ alkalosis is preferable to ‘contraction’ al- amount compares to <500 mM/day for all nonrespiratory
kalosis, because chloride repletion corrects the alkalosis in the acids that are handled by the kidney and gut. Pulmonary venti-
face of depleted volume but not vice versa [46]. In other cases, lation is adjusted by the respiratory center in response to
hormonal mechanisms produce ongoing losses of Cl−. Thus, changes in PaCO2 , blood pH and PaO2 as well as other factors
at best, the Cl− deficit can be offset only temporarily by Cl− (e.g. exercise, anxiety and wakefulness). PaCO2 changes in com-
administration; this form of metabolic alkalosis is said to be pensation for alterations in arterial pH produced by metabolic
‘chloride resistant’ (Table 3). Similar to hyperchloremic acid- acidosis or alkalosis in predictable ways (Table 4).
osis, these disorders can be distinguished by measurement of
the urine Cl− concentration. Diuretics and other forms of
volume contraction produce metabolic alkalosis predominant- R E S P I R AT O R Y AC I D O S I S
ly by stimulating aldosterone secretion, as discussed earlier.
However, diuretics also induce K+ and Cl− excretion directly, When CO2 elimination is inadequate relative to the rate of tissue
further complicating the problem and inducing metabolic production, PaCO2 increases to a new steady-state determined by

1108 M. Al-Jaghbeer and J.A. Kellum


Table 4. Observational acid–base patterns

Disorders HCO
3 (mEq/L) PCO2 (mmHg) SBE (mEq/L)
Metabolic acidosis <22 (1.5 × HCO
3 )+8 Less than −5
40 + SBE
Metabolic alkalosis >26 (0.7 × HCO 3 ) + 21 Greater than +5
40 + (0.6 × SBE)
Acute respiratory acidosis [(PCO2 − 40)/10] + 24 >45 0
Chronic respiratory acidosis [(PCO2 − 40)/3] + 24 >45 0.4 × (PCO2 − 40)
Acute respiratory alkalosis 24 − [(40 − PCO2)/5] <35 0
Chronic respiratory alkalosis 24 − [(40 − PCO2)/2] <35 0.4 × (PCO2 − 40)
From Kellum [50] with permission.

alveolar ventilation and CO2 production. Acutely, the increase in Chronic hypercapnia requires treatment when there is an
PaCO2 increases both the [H+] and the [HCO

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


3 ] in blood accord- acute deterioration. In this setting, it is important to recognize
ing to the carbonic acid equilibrium equation. Thus, the change that the goal of therapy is not a normal value for PaCO2 (35–
in [HCO− 3 ] is mediated simply by the dissociation of H2CO3 45 mmHg), but rather restoration of the patient’s baseline
into H+ and HCO− 3 , not by an active physiologic adaptation re- PaCO2 (if known). If the baseline PaCO2 is not known, a target
sponse. Similarly, the increase in [HCO− 3 ] does not ‘buffer’ the PaCO2 of 60 mmHg is reasonable. Overventilation has two un-
increase in [H+]. There is no change in the strong ion difference desirable consequences. First, life-threatening alkalemia can
and hence no change in SBE. Cellular acidosis always occurs in occur if the PaCO2 is rapidly normalized in a patient with
respiratory acidosis, since CO2 builds up in the tissues. If the chronic respiratory acidosis and an appropriately large strong
PaCO2 remains increased, active compensatory mechanisms are ion difference. Secondly, even if the PaCO2 is corrected slowly,
activated and the strong ion difference increases to restore [H+] the patient will reduce the plasma strong ion difference over
toward normal. time, making it impossible to wean the patient from mechan-
Primarily, compensation is accomplished by removal of ical ventilation.
Cl− from the plasma space. Since movement of Cl− into the Noninvasive ventilation is another treatment option that is
tissues or red blood cells results in intracellular acidosis, Cl− useful in selected patients, particularly those with normal sen-

FULL REVIEW
must be removed from the body to achieve a lasting effect on sorium [51]. Rapid infusion of NaHCO3 in patients with re-
the strong ion difference. The kidney is the primary organ for spiratory acidosis can induce acute respiratory failure, if
Cl− removal. Accordingly, patients with renal disease have a alveolar ventilation is not increased to adjust for the increased
difficult time adapting to chronic respiratory acidosis. When CO2 load. Thus, if NaHCO3 is used, it must be administered
renal function is intact, Cl− is eliminated in the urine and, slowly and alveolar ventilation adjusted appropriately. If in-
after a few days, the strong ion difference increases to the level creasing the plasma [Na+] is not possible or not desirable,
necessary to return blood pH to ∼7.35. According to the Hen- NaHCO3 should be avoided.
derson–Hasselbalch equation, the increased pH will result in It may be useful to reduce CO2 production by reducing the
an increased [HCO 3 ] for a given PCO2. Thus, the ‘adaptive’ in- carbohydrate load in the nutritional support regimen, lower-
crease in [HCO 3 ] ‘results from’ the increase in pH and is not ing the temperature in febrile patients and providing adequate
the ‘cause for’ the increase in pH. sedation for anxious or combative patients. Treatment of shi-
Although the change in HCO 3 concentration is a conveni- vering in the postoperative period can reduce CO2 production.
ent and reliable marker for the metabolic compensation, it is However, it is unusual to control hypercarbia with these tech-
not the mechanism. This point is more than semantic because niques alone.
only changes in the independent variables of acid–base
balance (PCO2, ATOT and strong ion difference) can affect the
plasma [H+], and [HCO 3 ] is not an independent variable. R E S P I R AT O R Y A L K A LO S I S
The primary threat to life in cases of respiratory acidosis
comes not from acidosis but from hypoxemia. If the patient is Respiratory alkalosis may be the most frequently encountered
breathing room air, PaCO2 cannot exceed 80 mmHg before acid–base disorder. It occurs in a number of pathologic condi-
life-threatening hypoxemia results. Accordingly, supplemental tions, including salicylate intoxication, early sepsis, hepatic
oxygen is always required. When the underlying cause can be failure and hypoxic respiratory disorders. Respiratory alkalosis
addressed quickly (e.g. reversal of narcotics with naloxone), it also occurs with pregnancy and with pain or anxiety. Hypo-
may be possible to avoid endotracheal intubation. Mechanical capnia appears to be a particularly bad prognostic indicator in
support is indicated when the patient is unstable or at risk for patients with critical illness [52]. As in acute respiratory acid-
instability or when central nervous system function deterio- osis, acute respiratory alkalosis results in a small change in
rates. Furthermore, in patients who are exhibiting signs of re- [HCO 3 ] as dictated by the Henderson–Hasselbalch equation.
spiratory muscle fatigue, mechanical ventilation should be If hypocapnia persists, the strong ion difference will begin to
instituted before overt respiratory failure occurs. Thus, it is not decrease as a result of renal Cl− reabsorption. After 2–3 days,
the absolute PaCO2 value that is important but rather the clin- the strong ion difference assumes a new, lower, steady state
ical condition of the patient. [53]. Severe alkalemia is unusual in patients with respiratory

Acid–base disturbances 1109


alkalosis, and management is therefore directed to the under- 13. Kang Y, Aggarwal S, Virji M et al. Clinical evaluation of autotransfusion
lying cause. Typically, these mild acid–base changes are clinic- during liver transplantation. Anesth Analg 1991; 72: 94–100
14. Kraut JA, Nagami GT. The anion gap in the evaluation of acid-base disor-
ally more important for what they can alert the clinician to, in ders: what are its limitations and can its effectiveness be improved. Clin J
terms of underlying disease, than for any threat they pose to Am Soc Nephrol 2013; 8: 2018–2024
the patient. In rare cases, respiratory depression with narcotics 15. Corey HE. Stewart and beyond: new models of acid-base balance. Kidney
is necessary. Int 2003; 64: 777–787
16. Kellum JA, Kramer DJ, Pinsky MR. Strong ion gap: a methodology for
exploring unexplained anions. J Crit Care 1995; 10: 51–55
17. Gunnerson KJ, Srisawat N, Kellum JA. Is there a difference between strong
P S E U D O R E S P I R AT O R Y A L K A LO S I S ion gap in healthy volunteers and intensive care unit patients? J Crit Care
2010; 25: 520–524
The presence of arterial hypocapnia in patients with profound 18. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992;
20: 80–93
circulatory shock has been termed ‘pseudorespiratory alkal- 19. Weil MH, Afifi AA. Experimental and clinical studies on lactate and pyru-
osis’ [54]. This condition can be seen when alveolar ventilation vate as indicators of the severity of acute circulatory failure (shock). Circu-

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


is supported but the circulation is grossly inadequate. In such lation 1970; 41: 989–1001
conditions, the mixed venous PCO2 is significantly elevated, 20. Mikkelsen ME, Miltiades AN, Gaieski DF et al. Serum lactate is associated
but the arterial PCO2 is normal or even decreased secondary to with mortality in severe sepsis independent of organ failure and shock. Crit
Care Med 2009; 37: 1670–1677
decreased CO2 delivery to the lungs and increased pulmonary 21. Levy B. Lactate and shock state: the metabolic view. Curr Opin Crit Care
transit time. Overall, CO2 clearance is markedly decreased and 2006; 12: 315–321
there is marked tissue acidosis, usually involving both meta- 22. Bellomo R, Kellum JA, Pinsky MR. Visceral lactate fluxes during early en-
bolic and respiratory components. The metabolic component dotoxemia in the dog. Chest 1996; 110: 198–204
comes from tissue hypoperfusion and hyperlactatemia. Arter- 23. De Backer D, Creteur J, Zhang H et al. Lactate production by the lungs in
acute lung injury. Am J Resp Crit Care Med 1997; 156: 1099–1104
ial oxygen saturation also may appear to be adequate despite 24. Marik PE, Bellomo R, Demla V. Lactate clearance as a target of therapy in
tissue hypoxemia. This condition is rapidly fatal unless cardiac sepsis: a flawed paradigm. OA Crit Care 2013; 1:3
output is rapidly corrected. 25. Levraut J, Ciebiera JP, Chave S et al. Mild hyperlactatemia in stable septic
patients is due to impaired lactate clearance rather than overproduction.
Am J Resp Crit Care Med 1998; 157: 1021–1026
26. Severin PN, Uhing MR, Beno DWA et al. endotoxin-induced hyperlactate-
C O N F L I C T O F I N T E R E S T S TAT E M E N T
FULL REVIEW

mia results from decreased lactate clearance in hemodynamic ally stable


rats. Crit Care Med 2002; 30: 2509–2514
None declared. 27. Levy B, Bollaert P-E, Charpentier C et al. Comparison of norepinephrine
and dobutamine to epinephrine for hemodynamics, lactate metabolism,
and gastric tonometric variables in septic shock: a prospective randomized
study. Intensive Care Med 1997; 23: 282–287
REFERENCES 28. De backer D. Lactic acidosis. Intensive Care Med 2003; 29: 699–702
29. Gutierrez G, Wolf ME. Lactic acidosis in sepsis: a commentary. Intensive
1. Gunnerson KJ, Saul M, He S et al. Lactate versus non-lactate metabolic Care Med 1996; 22: 6–16
acidosis: a retrospective outcome evaluation of critically ill patients. Crit 30. Forsythe SM, Schmidt GA. Sodium bicarbonate for the treatment of lactic
Care Med 2006; 10: R22–R32 acidosis. Chest 2000; 117: 260–267
2. Story DA. Intravenous fluid administration and controversies in acid-base. 31. Alberti KGMM. Diabetic emergencies. Br Med J 1989; 45: 242–263
Crit Care Resuscitation 1999; 1: 151–156 32. Good DG. Regulation of bicarbonate and ammonium absorption in the
3. Hickling KG, Walsh J, Henderson S et al. Low mortality rate in adult re- thick ascending limb of the rat. Kidney Int 1991; 40: S36–S42
spiratory distress syndrome using low-volume, pressure-limited ventila- 33. Androque HJ, Tannen RL. Ketoacidosis, hyperosmolar states, and lactic
tion with permissive hypercapnia: a prospective study. Crit Care Med acidosis. In: Kokko JP, Tannen RL (eds). Fluids and Electrolytes. Philadel-
1994; 22: 1568–1578 phia, WB: Saunders, 1996; pp. 643–674
4. Stewart P. Modern quantitative acid-base chemistry. Can J Physiol Phar- 34. Widmer B, Gerhardt RE, Harrington JT et al. Serum electrolyte and acid
macol 1983; 61: 1444–1461 base composition: the influence of graded degrees of chronic renal failure.
5. Kette F, Weil MH, Gazmuri RJ. Buffer solutions may compromise cardiac Arch Intern Med 1979; 139: 1099
resuscitation by reducing coronary perfusion pressure. JAMA 1991; 266: 35. Harrington JT, Cohen JJ. Metabolic acidosis. In: Cohen JJ, Kassirer JP
2121–2126 (eds). Acid-Base. Boston: Little, Brown and Co., 1982; pp. 121–225
6. Hindman BJ. Sodium bicarbonate in the treatment of subtypes of lactic 36. Mecher C, Rackow EC, Astiz ME et al. Unaccounted for anion in metabolic
acidosis: physiologic considerations. Anesthesiology 1990; 72: 1064–1076 acidosis during severe sepsis in humans. Crit Care Med 1991; 19: 705–711
7. Bleich HL, Swartz WB. Tris buffer (THAM): an appraisal of its physio- 37. Kirschbaum B. Increased anion gap after liver transplantation. Am J Med
logic effects and clinical usefulness. N Engl J Med 1966; 274: 782–787 Sci 1997; 313: 107–110
8. Figge J, Mydosh T, Fencl V. Serum proteins and acid-base equilibria: a 38. Gilfix BM, Bique M, Magder S. A physical chemical approach to the analysis
follow-up. J Lab Clin Med 1992; 120: 713–719 of acid-base balance in the clinical setting. J Crit Care 1993; 8: 187–197
9. Levinsky NG. Acidosis and alkalosis. In: Braunwald E, Isselbacher KJ, 39. Battle DC, Hizon M, Cohen E et al. The use of the urine anion gap in the
Petersdorf RG et al. (eds): Harrison’s Principles of Internal Medicine. diagnosis of hyperchloremic metabolic acidosis. N Engl J Med 1988; 318: 594
11th edn. New York: McGraw-Hill, 1987 40. Wilkes NJ, Woolf R, Mutch M et al. The effects of balanced versus saline-
10. Salem MM, Mujais SK. Gaps in the anion gap. Arch Intern Med 1992; based hetastarch and crystalloid solutions on acid-base and electrolyte
152: 1625–1629 status and gastric mucosal perfusion in elderly surgical patients. Anesth
11. Gabow PA. Disorders associated with an altered anion gap. Kidney Int Analg 2001; 93: 811–816
1985; 27: 472–483 41. Kellum JA. Recent advances in acid-base physiology applied to critical
12. Whelton A, Carter GG, Garth M et al. Carbenicillin-induced acidosis and care. In: Vincent JL (ed). Yearbook of Intensive Care and Emergency
seizures. JAMA 1971; 218: 1942 Medicine. Heidelberg: Springer, 1998; pp. 577–587

1110 M. Al-Jaghbeer and J.A. Kellum


42. Morgan TJ, Venkatesh B, Hall J. Crystalloid strong ion difference deter- 49. Gattinoni L, Lissoni A. Respiratory acid-base disturbances in patients
mines metabolic acid base change during in vitro hemodilution. Crit Care with critical illness. In: Ronco C, Bellomo R (eds). Critical Care Nephrol-
Med 2002; 30: 157–160 ogy Dordrecht. The Netherlands: Kluwer Academic Publishers, 1998;
43. Gennari FJ. Intravenous fluid therapy: saline versus mixed electrolyte and pp. 297–312
anion solutions. Am J Kidney Dis 2013; 62: 20–22 50. Kellum JA. Determinants of blood pH in health and disease. Crit Care
44. Shaw AD, Bagshaw SM, Goldstein SL et al. Major complications, mortal- 2000; 4: 6–14
ity, and resource utilization after open abdominal surgery. Ann Surg 2012; 51. Brochard L, Mancebo J, Wysocki M et al. Noninvasive ventilation for
255: 821–829 acute exacerbations of chronic obstructive pulmonary disease. N Engl J
45. Yunos NM, Bellomo R, Hegarty C et al. Association between a chloride- Med 1995; 333: 817–822
liberal vs chloride-restrictive intravenous fluid administration strategy and 52. Gennari FJ, Kassirer JP. Respiratory alkalosis. In: Cohen JJ, Kassirer JP
kidney injury in critically ill adults. JAMA 2012; 308: 1566–1572 (eds). Acid-Base. Boston: Little, Brown and Co., 1982; pp. 349–376
46. Luke RG, Galla JH. It is chloride depletion alkalosis, not contraction alkal- 53. Cohen JJ, Madias NE, Wolf CJ et al. Regulation of acid-base equilibrium in
osis. J Am Soc Nephrol 2012; 23: 204–207 chronic hypocapnia: evidence that the response of the kidney is not geared
47. Hossain SA, Chaudhry FA, Zahedi K et al. Cellular and molecular basis of to the defense of extracellular [H+]. J Clin Invest 1976; 57: 1483–1489
increased ammoniagenesis in potassium deprivation. Am J Renal Physiol 54. Adrogue HJ, Madias NE. Management of life-threatening acid-base disor-
2011; 301: F969–F978 ders: part II. N Engl J Med 1998; 338: 107–111

Downloaded from https://academic.oup.com/ndt/article-abstract/30/7/1104/2324819 by guest on 15 October 2019


48. Hamm LL, Hering-smith KS, Nakhoul NL. Acid-base and potassium
homeostasis. Semin Nephrol 2013; 33: 257–264 Received for publication: 17.9.2013; Accepted in revised form: 5.8.2014

FULL REVIEW

Acid–base disturbances 1111