Diagnosis of metabolic acid-base disturbances in critically ill patients.

We compare two commonly used diagnostic approaches, one relying on plasma bicarbonate concentration and "anion gap," the other on "base excess," with a third method based on physicochemical principles, for their value in detecting complex metabolic acid-base disturbances. We analyzed arterial blood samples from 152 patients and nine normal subjects for pH, PCO(2), and concentrations of plasma electrolytes and proteins. Ninety-six percent of the patients had serum albumin concentration < or = 3 SD below the mean of the control subjects. In about one-sixth of the patients, base excess and plasma bicarbonate were normal. In a great majority of these apparently normal samples, the third method detected simultaneous presence of acidifying and alkalinizing disturbances, many of them grave. The almost ubiquitous hypoalbuminemia confounded the interpretation of acid-base data when the customary approaches were applied. Base excess missed serious acid-base abnormalities in about one-sixth of the patients; this method fails when the plasma concentrations of the nonbicarbonate buffers (mainly albumin) are abnormal. Anion gap detected a hidden "gap acidosis" in only 31% of those samples with normal plasma bicarbonate in which such acidosis was diagnosed by the third method; when adjusted for hypoalbuminemia, it reliably detected the hidden abnormal anions. The proposed third method identifies and quantifies individual components of complex acid-base abnormalities and provides insights in their pathogenesis.

Anion gap and hypoalbuminemia.

OBJECTIVES: To show how hypoalbuminemia lowers the anion gap, which can mask a significant gap acidosis; and to derive a correction factor for it. DESIGN: Observational study. SETTING: Intensive care unit in a university-affiliated hospital. SUBJECTS: Nine normal subjects and 152 critically ill patients (265 measurements). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Arterial blood samples analyzed for pH, PCO2, and concentrations of plasma electrolytes and proteins. Marked hypoalbuminemia was common among the critically ill patients: 49% of them had serum albumin concentration of <20 g/L. Each g/L decrease in serum albumin caused the observed anion gap to underestimate the total concentration of gap anions by 0.25 mEq/L (r2 = .94). CONCLUSIONS: The observed anion gap can be adjusted for the effect of abnormal serum albumin concentrations as follows: adjusted anion gap = observed anion gap + 0.25 x ([normal albumin] [observed albumin]), where albumin concentrations are in g/L; if given in g/dL, the factor is 2.5. This adjustment returns the anion gap to the familiar scale of values that apply when albumin concentration is normal.

Stewart's quantitative acid-base chemistry: applications in biology and medicine.

We review P.A. Stewart's quantitative approach to acid-base chemistry, starting with its historical context. We outline its implications for cellular and membrane processes in acid-base physiology; discuss its contributions to the understanding and analysis of acid-base phenomena; show how it can be applied in clinical problems; and propose a classification of clinical acid-base disturbances based on this general approach.

Dissociation between dyspnea and respiratory effort.

Breathlessness induced by hypercapnia may be related to the sensation of respiratory effort or to the central or peripheral effects of CO2. To examine the relationship among breathlessness, respiratory effort, and hypercapnia, we studied eight normal naive subjects. By using a visual feedback system, subjects maintained a constant ventilation of 50-60 L/min. PETCO2 was held at 40 mm Hg during the first 2 min of each trial (control period), then for 4 min (test period) was either kept at 40 mm Hg or elevated to 50 mm Hg. At the end of each control and test period, subjects were asked to give separate ratings for dyspnea (an unpleasant urge to breathe) and for the sense of respiratory effort (analogous to lifting a weight) on a 50-cm visual analog scale. Hypercapnia was associated with a significant reduction in effort ratings (-7.3 +/- 6.4, mean +/- SD, p < 0.05) and a concomitant increase in dyspnea (+6.6 +/- 6.0, p < 0.05). We conclude that dyspnea associated with hypercapnia is dissociated from changes in respiratory effort, and that CO2 has a direct central effect that leads to breathlessness. Our data also suggest that the sense of effort at a given level of ventilation is less when the ventilation is the result of "reflex" stimuli to breathe rather than "voluntary" signals to the respiratory muscles.

Serum proteins and acid-base equilibria: a follow-up.

A mathematic model that described the acid-base behavior of blood plasma has been revised to incorporate pK values of individual histidine residues on human serum albumin determined by nuclear magnetic resonance spectroscopy. With the insights derived from the model a method for evaluation of the strong ion difference has been developed. Thus if pH, PCO2, and the concentrations of serum albumin and phosphate are measured, all independent variables, which physically determine "acid-base balance" in plasma, can be quantified. New ways to evaluate "unidentified anions" in metabolic acidosis can be explored with this approach.

Oral mucosal stimulation modulates intensity of breathlessness induced in normal subjects.

Patients with chronic obstructive pulmonary disease (COPD) often report an increase in breathlessness when they breathe through a mouthpiece. We hypothesized that stimulation of receptors in the oral mucosa modulates the sensation of breathlessness. We studied 10 normal naive volunteers in whom breathlessness was induced by having them breathe for 4 min with an inspiratory resistive load (18 cm H2O/L/s) while breathing was stimulated by CO2 inhalation (end-tidal PCO2 maintained at 55 mm Hg). Initially, subjects breathed with a tight-fitting face mask and inspiratory flow was displayed on a storage oscilloscope. In subsequent trials, the subjects were asked to match this trace, which controlled ventilation and the pattern of breathing. Subjects performed eight trials, four with the tight-fitting mask only (M) and four with a mouthpiece and the mask (MM). M and MM were alternated; the initial condition was chosen at random. Following each of the trials, subjects rated the intensity of their breathlessness by choosing a number from a modified Borg scale. On the average, subjects were more breathless while breathing with the mask and mouthpiece than with the mask alone (mean ratings of breathlessness 6.6 +/- 1.1 and 5.6 +/- 1.8 units, p less than 0.01). Six subjects repeated the protocol on 2 additional days: 1 day with inhalation of warm (34 degrees C), humidified air and 1 day after topical application of 4% lidocaine to the oral mucosa. Both these interventions abolished the differences in breathlessness between mask and mouthpiece and mask alone. We conclude that afferent information from oral mucosal stimulation influences the intensity of breathlessness.

Effect of chest wall vibration on breathlessness in normal subjects.

This study evaluated the effect of chest wall vibration (115 Hz) on breathlessness. Breathlessness was induced in normal subjects by a combination of hypercapnia and an inspiratory resistive load; both minute ventilation and end-tidal CO2 were kept constant. Cross-modality matching was used to rate breathlessness. Ratings during intercostal vibration were expressed as a percentage of ratings during the control condition (either deltoid vibration or no vibration). To evaluate their potential contribution to any changes in breathlessness, we assessed several aspects of ventilation, including chest wall configuration, functional residual capacity (FRC), and the ventilatory response to steady-state hypercapnia. Intercostal vibration reduced breathlessness ratings by 6.5 +/- 5.7% compared with deltoid vibration (P less than 0.05) and by 7.0 +/- 8.3% compared with no vibration (P less than 0.05). The reduction in breathlessness was accompanied by either no change or negligible change in minute ventilation, tidal volume, frequency, duty cycle, compartmental ventilation, FRC, and the steady-state hypercapnic response. We conclude that chest wall vibration reduces breathlessness and speculate that it may do so through stimulation of receptors in the chest wall.

The role of serum proteins in acid-base equilibria.

Serum proteins act as weak acids and participate in acid-base balance. Their effects are imprecisely quantified; in particular, the roles of albumin and globulins need reevaluation. We approached the problem in three steps. First, in artificial solutions resembling serum but with human serum albumin as the only protein moiety, we varied the strong ion difference (SID), partial pressure of carbon dioxide (Pco2) and the concentration of albumin [( Alb]) and fixed the concentration of inorganic phosphate [( Pi]). We measured pH and derived the charges on albumin. Second, extending the work of Stewart (Stewart PA. How to understand acid-base. A quantitative acid-base primer for biology and medicine. New York: Elsevier, 1981:1-286), we developed a mathematical model that solves for pH and for the charges on albumin as functions of SID, Pco2, [Pi], and [Alb]. The calculated values fit the observed values well; that is, the model describes well the behavior of these solutions over a wide range of simulated complex acid-base disturbances. Finally, in human serum samples containing both albumin and globulins, we varied SID, Pco2, and total protein concentration [( TP]); we fixed [Pi] and then measured pH and derived the charges on proteins as above. When we applied to these data the computer model developed for albumin alone, the calculated pH and derived charges on albumin values agreed well with the observed pH and derived charges on proteins. We conclude first that human serum globulins play a negligible role in acid-base equilibria, and second, that in normal human serum at pH 7.40 with [TP] = 7 and [Alb] = 4.3 gm/dl, the charges attributed to proteins are approximately 12 mEq/L; this is substantially less than the value of approximately 17 mEq/L given by many contemporary texts, based on work of van Slyke et al. (van Slyke DD, Hastings AB, Hiller A, Sendroy J Jr. Studies of gas and electrolyte equilibria in blood. XIV. Amounts of alkali bound by serum albumin and globulin. J Biol Chem 1928;79:769-80). These findings should be considered when evaluating acid-base balance in patients with abnormal serum albumin concentration, for example, when interpreting values of the anion gap.

Distinguishable types of dyspnea in patients with shortness of breath.

Dyspnea frequently accompanies a variety of cardiopulmonary abnormalities. Although dyspnea is often considered a single sensation, alternatively it may encompass multiple sensations that are not well explained by a single physiologic mechanism. To investigate whether breathlessness experienced by patients represents more than one sensation, we studied 53 patients with one of the following seven conditions: pulmonary vascular disease, neuromuscular and chest wall disease, congestive heart failure, pregnancy, interstitial lung disease, asthma, and chronic obstructive pulmonary disease. Patients were asked to choose descriptions of their sensation(s) of breathlessness from a dyspnea questionnaire listing 19 descriptors. Cluster analysis was used to identify natural groupings among the chosen descriptors. We found that patients could distinguish different sensations of breathlessness. In addition, we found an association between certain groups of descriptors and specific conditions producing dyspnea. These findings concur with those in an earlier study in normal volunteers in whom dyspnea was induced by various stimuli. We conclude that different types of dyspnea exist in patients with a variety of cardiopulmonary abnormalities. Furthermore, different mechanisms may mediate these various sensations.

Distinguishable sensations of breathlessness induced in normal volunteers.

Various theories about the genesis of dyspnea have often assumed that the sensation is similar from patient to patient and is generated by a single underlying mechanism. To investigate whether breathlessness induced in normal volunteers by different stimuli represents one or more than one sensation, we studied 30 subjects in whom breathlessness was induced by each of 8 different stimuli: breath-holding, CO2 inhalation, inhalation of CO2, with ventilation voluntarily targeted below the level dictated by chemical drive, breathing with a resistive load, breathing with an elastic load, voluntary elevation of functional residual capacity, voluntary limitation of tidal volume, and exercise. For each stimulus, subjects were asked to choose description of their sensation(s) of breathlessness from a questionnaire listing 19 descriptors. The responses from this questionnaire were evaluated using cluster analysis to search for relationships among descriptors and to identify natural groupings. We found that distinct groups of descriptors emerged, i.e., subjects could distinguish different sensations of breathlessness. In addition, we found an association between certain descriptor groups and stimuli. We conclude that the term breathlessness may encompass multiple sensations, and, therefore, may not be explainable by a single physiologic mechanism.

Breathlessness induced by dissociation between ventilation and chemical drive.

Suppression of ventilation by tasks such as talking may produce breathlessness in normal individuals under conditions when a strong respiratory drive exists, e.g., during exercise, and in patients with severe lung disease. To investigate the nature of breathlessness produced by a dissociation between ventilation and chemical drive, we studied ten naive normal subjects who breathed at various levels of ventilation while end-tidal PCO2 (PETCO2) was held at 55 mm Hg. After a 10-min equilibration period of free breathing at PETCO2 = 55 mm Hg, subjects used a visual target to adjust ventilation to five different levels ranging from 50% below to 50% above the chemically driven ventilation (CDV). Ratings of breathlessness were made on a visual analogue scale relative to the intensity of breathlessness experience at CDV. As ventilation was targeted to levels below CDV, all subjects became increasingly breathless; the response was more variable when ventilation was targeted to levels above CDV. Overall, the relationship between ventilation and breathlessness was described by a hyperbolic function, for which the coefficient of determination (R2) was 0.92. Ventilation was suppressed below CDV without recruitment of antagonistic muscles during inspiration. The intensity of breathlessness was not correlated with measures of respiratory effort. We conclude that suppressed ventilation is a useful model for the study of breathlessness not fully explained by measures of respiratory effort and we speculate that the dissociation between chemical drive and afferent signals produced by motion of the lung and chest wall is important in modulating the sensation of breathlessness.

Naloxone does not alter response to hypercapnia or resistive loading in chronic obstructive pulmonary disease.

To assess the role of endogenous opioid peptides in ventilatory control in patients with chronic obstructive lung disease, we measured the ventilatory and mouth occlusion pressure responses to hypercapnia and the compensatory response to an inspiratory resistive load in 11 male patients with COPD before and after intravenous administration of naloxone or placebo on 2 separate days. There were no statistically significant differences between naloxone and placebo administration in any index of ventilatory response to CO2 or resistive loading. When an inspiratory resistive load was added during CO2 rebreathing, minute ventilation at PETCO2 = 50 mm Hg in all 11 patients decreased significantly (p less than 0.05) with placebo and naloxone. In response to the inspiratory resistive load, in eight of the 11 patients mouth occlusion pressure (P0.1) did not increase; these eight subjects were classified as noncompensators. Naloxone did not affect the P0.1 response to inspiratory resistive loading, either in the group as a whole or in the subgroup of eight patients classified as noncompensators. Our study was unable to demonstrate that increased activity of endogenous opioid peptides suppresses the ventilatory response to CO2 or resistive loading in patients with chronic obstructive lung disease.

Acid-base disorders in critical care medicine.

In critically ill patients, nonrespiratory (metabolic) alkalosis is the most common acid-base disturbance; it is caused by hypochloremia and/or by hypoproteinemia. Information on the concentration of plasma proteins should be included when evaluating acid-base status.

Hyperventilation with hypoproteinemia.

Hypoproteinemia by itself produces a metabolic alkalosis. It is not clear whether a respiratory compensation (hypercapnia) develops with this alkalosis; patients with liver cirrhosis, most of them with hypoproteinemia, are known to hyperventilate. We studied 23 clinically stable patients with hypoproteinemia, with very low albumin-to-globulin ratios (range 0.4 to 1.1), who had either liver cirrhosis (n = 12) or other medical conditions (n = 11). In both groups, there was marked hypocapnia, accompanied by alkalemia (PaCO2 values (mean +/- SD) 31 +/- 2 and 32 +/- 3 torr; pH (mean +/- SD) 7.45 +/- 0.03 and 7.47 +/- 0.03, for the patients with cirrhosis and those without, respectively). Hypoxemia was not the stimulus provoking hyperventilation. The lowering of PaCO2 was proportional to the reduction of serum albumin and total protein concentrations; no detectable difference was seen between the patients with cirrhosis and those without cirrhosis in this apparent dependence of PaCO2 on the concentration of serum proteins. Many of these clinically stable patients with hypoproteinemia, with or without liver cirrhosis, had appreciable concentrations of unidentified anions in plasma (inappropriately high anion gap). Whatever the nonrespiratory acid-base status of the patients with hypoproteinemia, their pulmonary ventilation (hypocapnia) appeared excessive when compared with subjects (presumably) without proteinemia who had similar nonrespiratory acid-base states. The mechanism responsible for the hyperventilation in hypoproteinemia and the nature of the unidentified anions in this condition are obscure.

Detection of hypercapnia by normal subjects.

1. To investigate whether changes in PaCO2 can be detected independently of the CO2-induced changes in pulmonary ventilation, we tested five normal subjects for the ability to distinguish different levels of end-tidal PCO2 (PETCO2) while holding minute ventilation constant. 2. Helped by a visual feedback system, the subjects maintained a constant ventilation targeted at a level that was higher than that dictated by the chemical drive at PETCO2 = 50 mmHg (6.7 kPa). End-tidal PCO2 was held at 40 mmHg (5.3 kPa) during the first 2 min of each test trial ('control period'); then, for 4 min ('test period'), PETCO2 was either elevated to 50 mmHg or kept at 40 mmHg. Twelve runs were performed by each subject. 3. In 24 out of the total 30 trials (80%) in which PETCO2 was raised during the test period to 50 mmHg, the subjects detected the changes. There was one false positive result (3%), when PETCO2 kept at 40 mmHg during the test period was reported as different from control. In four out of the five subjects the ability to detect the change in PETCO2 from 40 to 50 mmHg was statistically significant. 4. We conclude that increases in PETCO2 can be detected independently of changes in the absolute level of ventilation.

Endogenous opioids and ventilatory adaptation to prolonged hypoxia in goats.

To investigate whether endogenous opioid peptides mediate time-dependent changes in ventilatory control during prolonged hypoxia, we studied four adult goats at rest during 14 days at simulated high altitude in a hypobaric chamber (PB approximately 450 Torr). Arterial PCO2 fell during the first several hours of hypoxia, remained stable over the next 7 days, and then rose slightly (but without statistical significance) by day 14. Ventilatory responsiveness to CO2 increased during the first week of hypoxia. By day 14, while still greater than control, the ventilatory response to CO2 was less than that observed on day 7. Immunoactive beta-endorphin levels in plasma and CSF did not change during the 14-day period. Administration of naloxone on day 14 did not restore the ventilatory response to CO2 to the level observed during the first week of acclimatization. We conclude that in adult goats, time-dependent changes in ventilatory response to CO2 during acclimatization to prolonged hypoxia are not primarily attributable to alterations in endogenous opioid peptide activity.

Acid-base effects of altering plasma protein concentration in human blood in vitro.

We altered the concentration of plasma proteins in human blood in vitro by adding solutions with [Na+], [K+], and [Cl-] resembling those in normal blood plasma, either protein-free or with a high concentration of human albumin. After equilibrating the samples with a gas containing 5% CO2-12% O2-83% N2 at 37 degrees C, we measured pH, PCO2, and PO2; in separated plasma, we determined the concentrations of total plasma proteins and albumin and of the completely dissociated electrolytes (strong cations Na+, K+, Mg2+ and anions Cl-, citrate3-). With PCO2 nearly constant (mean = 35.5 Torr; coefficient of variation = 0.02), lowering plasma protein concentration produced a metabolic alkalosis, whereas increasing plasma albumin concentration gave rise to a metabolic acidosis. These acid-base disturbances occurred independently of a minor variation in the balance between the sums of strong cations and anions. We quantified the dependence of several acid-base variables in plasma on albumin (or total protein) concentration. Normal plasma proteins are weak nonvolatile acids. Although their concentration is not regulated as part of acid-base homeostasis, hypoproteinemia and hyperalbuminemia per se produce alkalosis and acidosis, respectively.

Hypoproteinemic alkalosis.

Hypoproteinemia by itself causes a nonrespiratory ("metabolic") alkalosis. On the average, a decrease in plasma albumin concentration of 1 g/dl produces an increase in "standard" bicarbonate of 3.4 mM/liter, and an apparent base excess of +3.7 meq/liter; it also reduces the value of the normal anion gap by about 3 meq/liter. Concentration of plasma protein should be measured as part of the analysis of acid-base status. Interpretation of acid-base data requires special consideration in "primary hypoproteinemic alkalosis."

Endogenous opioids and ventilatory responses to hypercapnia in normal humans.

Though administration of opioid peptides depresses ventilation and ventilatory responsiveness, the role of endogenous opioid peptides in modulating ventilatory responsiveness is not clear. We studied the interaction of endogenous opioids and ventilatory responses in 12 adult male volunteers by relating hypercapnic responsiveness to plasma levels of immunoactive beta-endorphin and by administering the opiate antagonist naloxone. Ventilatory responsiveness to hypercapnia was not altered by pretreatment with naloxone, and this by itself suggests that endogenous opioids have no role in modulating this response. However, there was an inverse relationship between basal levels of immunoactive beta-endorphin in plasma and ventilatory responsiveness to CO2. Furthermore, plasma beta-endorphin levels rose after short-term hypercapnia but only when subjects had been pretreated with naloxone. We conclude that measurement of plasma endorphin levels suggests relationships between endogenous opioid peptides and ventilatory responses to CO2 that are not apparent in studies limited to assessing the effect of naloxone.