The explanation was that respiratory acidosis with accumulation of CO₂ caused a low pH. During this period, it is true to say that acid-base analysis went from being a physiological laboratory practice to a clinical necessity. Measuring total CO₂ with the Van Slyke apparatus and pH with another instrument, however, gradually came to be considered cumbersome.

Professor Poul Astrup said that he volunteered to ventilate polio patients at the Blegdams Hospital in Copenhagen. The respirators were operated by hand, and while he sat and pumped he wondered: “Am I doing this too fast, too slow, or just right? Where can I get an easy answer to my question? I must know the patient’s situation now, not in a few hours’ time.”

Astrup studied Van Slyke’s work and used the knowledge that the titration line for CO₂ to pH was approximately a straight line in the physiological measurement area (pCO₂ from 1.5-15 kPa), presuming that pCO₂ was marked logarithmically.

Capillary blood is taken in three capillary tubes (approximately 50 µL in each tube). Two of these samples are transferred to the equilibration chambers and equilibrated with special gas mixtures with known high and relatively low pCO₂ for 3-4 minutes. For the third sample, there is no preprocessing of the equilibration stage. The pH should be measured for all three samples. The titration line can be drawn based on the pH measurements in the two equilibrated samples. After that, the pCO₂ is read from the titration line with the measured pH from the blood sample, which was not equilibrated.

On a nomogram, which was later worked out by Professor Ole Siggaard-Andersen, it was possible to read base excess (BE), buffer base (BB) and standard bicarbonate parameters, which express the metabolic (non-respiratory) part of acid-base disturbance.

INTERPRETATION OF THE HENDERSON-HASSELBALCH EQUATION

In 1964, Lorentz Eldjarn from Norway created a box with a three-dimensional coordinate system. The pH was marked on the X axis. The pCO₂ was logged on the Y axis and the relevant bicarbonate level was marked on the Z axis. The graphic presentation was a double curved plane where the titration line for blood and spinal fluid was plotted.

It was possible to project this complicated plane down or into one of the three main symmetrical planes. Projection down or into the plane with the X and Y coordinates was Siggaard-Andersen’s curve nomogram where isobicarbonate lines were straight lines, projection into the plane with X and Z as coordinates became Davenport’s nomogram with plotted isobars for pCO2. The last projection with Y and Z as coordinates was called the Cohen and Kassirer nomogram with plotted iso-pH lines. A fourth nomogram with log pCO2 on the X axis and BE on the Y axis was called the Grogono nomogram and was also intended to simplify the interpretation of acid-base disturbances. In recent years I have thought a lot about the extent to which these nomograms have contributed to the understanding of acid-base disturbances, but there is no doubt that they did help quite a bit.
 

DIRECT MEASUREMENT OF ELECTRODES

Throughout the 1950s and the 1960s there was a lot of international focus on creating electrodes that could directly measure pO₂ and pCO₂ in blood samples. It was successful, eventually, and the most practically usable electrodes were subsequently called Clarke for the pO₂ electrode and Stow/Severinghaus for the pCO₂ electrode.

The three electrodes, pH, pCO₂ and pO₂, were then put together in one instrument, and analyzers for direct measurement of blood gases were available. Radiometer called its first analyzer of this type the Acid Base Laboratory 1 (ABL 1). This was the first fully microprocessor-controlled, fully-automated blood gas analyzer to become available commercially.

THE BIG TRANSATLANTIC ACID-BASE DEBATE

At the start of the 1960s, it could be said that Astrup and Siggaard-Andersen’s major input regarding acid-base put Copenhagen on the world map, and after a while the term “The Copenhagen School” began to be used. When the acid-base status of a patient was required, an “Astrup” was simply requested. I have heard such an acid-base status being requested as recently as 2011 at a hospital in Norway!

With the introduction of the metabolic parameters such as base excess (BE) and standard bicarbonate from “The Copenhagen School”, articles soon started to appear in leading journals such as The Lancet and the New England Medical Journal, etc., criticizing these new parameters! Two internists in Boston, William Schwartz and Arnold Relman, were particularly critical of the two parameters mentioned and claimed that BE, in particular, was not independent of pCO₂, especially in the case of hypercapnia. They pointed out that actual bicarbonate could well be used as the metabolic parameter. We call actual bicarbonate a mixed indicator that is present in both metabolic and respiratory acid-base disturbances. The presence of actual bicarbonate is most pronounced in metabolic acid-base disturbances.

A few years later, BE was modified and called “base excess extracellular fluid” (BEecF). The basis for this was that in vivo O₂ and CO₂ do not just equilibrate with blood but with all the extracellular fluid space. The buffer capacity is much higher in blood with hemoglobin present compared to the interstitial space, which does not contain hemoglobin and subsequently has a much lower buffer capacity.

It was then thought that the blood was diluted with extracellular fluid (fluid in the interstitial space). The concentration of hemoglobin in this diluted fluid would then be approximately one third of that of blood. Siggaard-Andersen had now worked out an “alignment” nomogram, where BE could be easily read. This meant, for example, that with a hemoglobin concentration of 15 g/dL or 9 mmol/L, you would use a hemoglobin concentration of 5 g/dL or 3 mmol/L when making your calculation and thus obtain the value of the BEecF.

Another element that led the Boston School, in particular, to not recognize BE, was that BE was a mystical, artificial parameter and was calculated from complicated algorithms. The various manufacturers of blood gas analyzers worked out their own algorithms for calculating BE. These algorithms contained pH, pCO₂ and hemoglobin. Some also included pO₂.

I was able to demonstrate, in 1980, that the calculation of BE varied quite a bit when analyzers from different manufacturers were compared, and this could have clinical consequences in some instances, particularly in metabolic alkalosis. Throughout the 1980s and the 1990s the calculation was standardized so that today Van Slyke’s algorithm or modifications of it are generally used. pO₂ is not included in this calculation.

The Copenhagen School have subsequently admitted that it would have been better to call the BE parameter “titratable acid” but with the sign reversed. This was certainly a familiar term from the titration of urine for measuring the excretion of acids or bases.

STEWART’S APPROACH TO THE INTERPRETATION OF ACID-BASE DISTURBANCES

In 1981, the Canadian Peter Stewart introduced a new calculation method that he felt would improve our understanding of acid-base disturbances, particularly metabolic disturbances. The traditional definition of acids and bases was abandoned, and two new parameters were introduced: SID = strong ion difference, which in simplified form is Na⁺ + K⁺ – Cl^-, and A_tot, which is the total amount of weak acids present.

Using these two parameters and pCO₂, he formulated six equations and created a “summation equation” that gave detailed information about the acid-base status. Furthermore, it may be mentioned that the SID was identical to Buffer Base that was introduced by Singer and Hastings in 1948.

Another feature, which should be mentioned with regard to Stewart’s approach, was that there was greater focus on the role of albumin and chloride in the understanding of acid-base disturbances. I include an example, which has led to discussion in international media: According to Stewart’s theory an acute isolated hypoalbuminemia would be regarded as a mild metabolic alkalosis.

The negative albumin charges will, due to a reduction of the albumin concentration, mean that these will have to be replaced to maintain electrical neutrality.This is primarily achieved by increasing the bicarbonate concentration, which is then interpreted as metabolic alkalosis. This is not an acid-base disturbance and should not be treated with acid therapies!

Stewart’s method became popular among researchers, but was not very user-friendly in the busy everyday clinical setting due to complicated equations and the need for programmable calculators.

NEW ELECTRODES FOR MEASURING ELECTROLYTES INTRODUCED IN BLOOD GAS ANALYZERS

From the end of the 1960s, ion-selective electrodes were developed for the macro electrolytes in blood plasma. It was particularly electrodes for sodium, potassium, chloride and calcium that were used and that were found to be reliable. These electrodes were built into the blood gas analyzer.

Simon’s discovery in 1969 that valinomycin was a particularly suitable carrier substance in a potassium electrode was of major importance. With regard to calcium, two answers had to be reported as ionized calcium varied depending on the pH.

When the pH was between 7.60 and 7.20, the “adjusted” calcium value was calculated to pH = 7.40. In this way, it was possible to distinguish between, for example, hypercalcemia, which was just secondary to acidosis, and actual hypercalcemia! After a while, several types of directly measuring electrodes were built into the blood gas analyzers, such as glucose and lactate electrodes.

SPECTROPHOTOMETRY AND OXIMETRY

Spectrophotometric determination of oxygen saturation was introduced at the start of the 1930s. During the Second World War the need to measure oxygen saturation increased due to the military needs of flying at high altitudes without stabilization of the hypobaric conditions.

During the 1970s and 1980s, there was a vast development in multiwavelength oximetry. Absorption spectra for hemoglobin derivatives, particularly within the wavelength area of 500-700 nm, were recorded. Measurements at over 100 different wavelengths were recorded to calculate the four common hemoglobin derivatives: oxyhemoglobin, deoxyhemoglobin, CO-hemoglobin and methemoglobin. It was also possible to measure sulfhemoglobin if this was present in the case of sulfa therapy.

Simultaneous measurement of oxygen saturation and partial pressure of oxygen soon became a clinical necessity for plotting the position of the oxyhemoglobin dissociation curve (ODC) and also the pO₂ 50 %, which could tell about the hemoglobin’s affinity for oxygen.

With the introduction of oximetry to the blood gas analyzers, the blood sample had to be split; one part went to the electrometry measurements and one part went to the oximetry measurements so that the modern blood gas analyzer now gives at least 11 directly measured parameters, four from the oximeter and seven from the electrometer.

In addition, glucose and lactate measurements became possible, so that on the printout there are 13 directly measured values as well as a number of calculated parameters.

QUALITY CONTROL

Many different test solutions have been prepared to control the quality of the measurements in the blood gas analyzers. What is common to them all is that the gases are very difficult to control, particularly low pO₂ (hypoxemia). The reason is that oxygen is not particularly water-soluble, and when the test ampoule is opened, oxygen in the air quickly diffuses into the solution. When using hemoglobin-containing solutions, it is difficult to produce these with deoxyhemoglobin, because this is immediately converted into oxyhemoglobin when it comes into contact with air.

In major international quality control tests, where many types of instruments are tested, the total coefficient of variation can be up to 15 % for low pO₂ values of 7-8 kPa. Ranges for these results can be 6.0-9.5 kPa! The large variation is caused by both preanalytical and analytical factors! Tonometry has also proved to be a good method for carrying out quality tests on pCO₂ and pO₂.

CONCLUSION

When you read the two-volume work by Peters and Van Slyke from 1934, it is surprising how much biochemical and physiological knowledge they had then! There has, however, particularly following the Second World War, been an incredible technological development resulting in today’s blood gas analyzers playing a very important role, particularly within intensive care medicine.

Otherwise it may be pedagogically correct in the future to rename BE and call it hydrogen ion excess with the opposite sign!