Loading...
 

May - August 2004: 
Volume 17, Issue 2

Click on the image to download the Issue in PDF format.

ARCHIVE

Latest developments in graphic diagnostic approach of arterial blood gases disturbances
Abstract
The objective of the present study was to investigate the diagnostic validity of a new diagrammatic approach to the diagnosis of acid-base balance disorders as proposed by A. Grogono. In order to examine the diagnostic validity of the proposed diagram, 3,122 arterial blood gas samples were drawn from 114 ICU patients and were interpreted using the Grogono diagram and the following approaches as comparators: 1. Τhe Siggaard-Andersen (S-A) chart, 2. The “Oxygen Status Algorithm” (OSA) software and 3. Τwo physicians with more than 10 years of experience in ICU, considered experts in acid-base balance disorders. Our results showed that: 1. Τhe Grogono diagram presents the higher diagnostic agreement with the other methods ranging from 59.7% to 72.5%; 2. The diagnostic agreement between the Grogono diagram and the OSA software (70.5%) is significantly higher (p<0.001) in comparison to that between the Grogono diagram and the S-A chart (59.7%); 3. Τhe ratios of diagnostic agreement between the Grogono diagram and each one of the two experts (72.5% and 62.1%, respectively) are significantly higher (p<0.001) than those calculated for the “S-A chart” (48.3% and 56.4%, respectively); 4. Τhe two expert physicians disagreed with each other in the diagnosis of 1/3 of the cases of arterial blood gases disturbances. Moreover, they presented significant diagnostic variation in the comparisons with the other methods. In conclusion, the Grogono diagram, although superior to the S-A chart, cannot be safely used for the diagnosis of acid-base balance disorders in everyday clinical practice, because it has been shown to provide inaccurate diagnoses in at least 25% of the cases. Pneumon 2004, 17(2):150-158.
Full text

INTRODUCTION

The first official description of acid-base balance disorders was pub lished in Lancet during the Asian cholera outbreak in England in 18311. Since then, tremendous scientific and technological advances in both laboratory investigation and diagnosis of acid-base disorders have occurred. The technological revolution that has taken place in the last 30 years led to the development of rapid, user friendly and small size arterial blood gas analyzers that have made blood gas analysis a routine test available to all clinicians. Furthermore, the diagnostic evaluation of acid-base balance disorders is not a subject dealt with only by specialized physicians any more; instead, physicians that are not experts in this field, as well as nurses manage such cases effectively in routine practice2,3.

Nevertheless, the enormous technological advance in laboratory methods and instruments that has occurred in the last decades has not brought a respective advance in the field of diagnosis. Diagrammatic methods of acid-base disorder diagnosis, with Siggaard-Andersen diagram and Davenport diagram being classic representatives of these methods, are the standard from the 1970's till now4-10, whereas the international literature indicates minimal progress in the diagnosis of acid-base balance disorders in recent years. A two-axial diagram proposed by A. Grogono is an exception, though11,12. The partial pressure of carbon dioxide (CO2) in arterial blood (respiratory component) is plotted on the horizontal axis and the standard base excess (SBE) (metabolic component of acid-base balance) on the vertical axis of the Grogono diagram. The objective of the present study was to assess the diagnostic validity of this diagram in comparison to classic methods of diagrammatic diagnosis and the opinion of expert physicians.

PATIENTS AND METHODS

The study included 114 patients treated in the Intensive Care Unit (ICU) of the "Korgialenio & Benakio" General Regional Hospital of the Greek Red Cross.

In total, 3,122 arterial blood samples were drawn from all patients. All blood samples were arterial, blood was drawn anaerobically with a 2.5 mL syringe applying all re­commended safety and infection control pre­­cau­­­tions13-16. The time between sample acquisition and measurement was less than 2 minutes.16 For the purposes of the study, each sample was separately assessed and had a code number. Two blood gas analyzers (Radiometer Copenhagen ABL 300; Radiometer Copenhagen ABL 510) were used.

A personal database for each individual patient was then constructed; results of all laboratory measurements, and a variety of additional parameters, such as type of respiration (spontaneous breathing or mechanical ventilation), parameters of mechanical ventilation, use of an artificial kidney (dialysis), time and kind of surgical treatment, as well as major changes in the patient's clinical condition were entered in this database on a daily basis. Blood gas analysis results were entered automatically to the database, through an on-line connection with the blood gas analyzer. To this purpose, a local network between the computer and the two blood gas analyzers was created via a multiple serial inlet and outlet PCI module. An appropriate data processing software program was developed using the data acquisition language "Labview for Windows 98/ME/NT, version 5.0".

This software program consists mainly of two basic parts. The first part relates to the entry of patient data (full name, age, gender), date of hospital admission, and diagnosis at admission, and includes a medical history section. The second part relates to the entry of the results of primary acid-base measurements. Of all acid-base balance parameters calculated by blood gas analyzers, only those that are actually measured (not derived), i.e. only pH, PaCO2, PaO2, hemoglobin concentration (cHb) and saturation (sHb), are transferred through the PCI port to the computer. The remaining parameters (secondary measurements), that are calculated by the blood gas analyzer using the primary measurements, were not processed by the software program, because the various types of blood gas analyzers use different equations for these calculations. Hence, to avoid calculation errors and to ensure a uniform way of calculating secondary parameters of acid-base balance, the software program used the latest formulae published in the international literature17-19 to calculate the following parameters:

1. plasma bicarbonate concentration [HCO3-]p,

2. standard bicarbonate concentration [HCO3-]st,

3. change in bicarbonate concentration Δ[HCO3-]p,

4. base excess (BEb),

5. standard base excess (BEst),

6. standard pH (pHst),

7. anion gap (AG) and its relative change (ΔΑG),

8. anion gap change to bicarbonate concentration change ratio (ΔΑG/Δ[HCO3-]p),

9. total blood carbon dioxide content (total CO2 cont),

10. pH and PaCO2 values corrected for patient temperature [pH(T) and PaCO2(T)]20.

All arterial blood gas samples were then evaluated using the following methods:

1. The Grogono diagram. The initial concept for the construction of the Grogono diagram (Figure 1) was based on the previously described Siggaard-Andersen4 and J. Severinghaus diagrams21. It's a new two-axial diagram with the partial pressure of CO2 (PaCO2), which represents the respiratory component of the acid-base balance, plotted on the horizontal axis and the standard base excess (SBE), which represents the metabolic component, plotted on the vertical axis. The diagram consists of 23 clearly defined areas, each one representing a particular acid-base balance disorder. Each area is defined on the basis of a formula that describes as accurately as possible a particular acid-base balance disorder. This formula was derived by the analysis of the acid-base parameters from all patients with a particular acid-base balance disorder published in the international literature. The final version of the Grogono diagram was published in 199811; its electronic form as a software program was developed using the JAVA programming language and is available through the Internet on the website of A. Grogono ((http://www.acid-base.com/homepage.html)12.

2. The Siggaard-Andersen chart (S-A chart). It was developed by Siggaard-Andersen in 1971 and it's in fact a revised edition of the two diagrams previously described by the same investigator. It is considered the standard diagram for the diagnosis of acid-base balance disorders not only because it is widely used by clinical doctors in their routine practice, but also because it has been the basis for the development of many diagrams that followed. It is a two-axial diagram that plots pH (horizontal axis) as a linear function of logPaCO2 (vertical axis), as derived by the Henderson-Hasselbalch equation.

3. The Oxygen-Status-Algorithm (OSA) software. This software was developed by the Radiometer-Copenhagen Company on the basis of the diagram and work

Image 1

Figure 1. Diagnostic areas of the Grogono diagram: 1 & 22: non-compensated metabolic alkalosis; 2: partially compensated me¬tabolic alkalosis; 3: maximally compensated metabolic alkalosis; 4: respiratory acidosis and metabolic alkalosis; 5: maximally compensated respiratory acidosis; 6: partially compensated respiratory acidosis; 7 & 8: non-compensated respiratory acidosis; 9 & 10: mixed respiratory and metabolic acidosis; 11 & 12: non-compensated metabolic acidosis; 13: partially compensated metabolic acidosis, 14: maximally compensated metabolic acidosis; 15: respiratory alkalosis and metabolic acidosis; 16: maximally compensated respiratory alkalosis; 17: partially compensated respiratory alkalosis; 18 & 19: non-compensated respiratory alkalosis; 20 & 21: Mixed respiratory and metabolic alkalosis; 23: Normal range.

done by Siggaard-Andersen22-24. This software was designed for personal computers using DOS and is installed in all blood gas analyzers manufactured by Radiometer-Copenhagen.

4. Two pulmonary medicine specialists with great experience in the interpretation of arterial blood gases and the treatment of ICU patients. These two specialists were members of the ICU medical staff in the "Korgialenio & Benakio" General Regional Hospital. It should be noted that there was no cooperation between the two specialists. Furthermore, there were no restrictions regarding the methods of diagnosis the specialists were allowed to use.

Statistical analysis: The diagnostic agreement between the Grogono diagram and the Siggaard-Andersen diagram or the OSA software was assessed using paired comparisons of the diagnoses provided by each method for all arterial blood gas samples. Then, we assessed the di agnostic agreement between each one of the three diagrammatic methods of diagnosis and each one of the specialist doctors separately, but also with both specialists combined. Agreement ratios higher than 80% were considered high; ratios between 60% and 80% moderate and ratios lower than 60% low. Variation between agreement ratios was assessed using the chi-square test for comparisons between ratios with the Yates correction.

RESULTS

Table 1 lists the results of all possible comparisons between the methods used for the diagnostic evaluation of all arterial blood gas samples in this study. An initial and principal remark regarding the data presented in this table is that ratios of diagnostic agreement between the Grogono diagram and all other methods, although higher than all the rest, range from 59.7% to 72.5%, i.e. they are moderate. Interestingly, the lower agreement ratio of the Grogono diagram is that with the Siggaard-Andersen diagram (59.7%) and the higher that with specialist 1 (72.5%).

Ratios of diagnostic agreement between Siggaard-Andersen diagram and the other methods are generally low (48.3% to 59.7%). It should be stressed that Siggaard-Andersen diagram shows a higher ratio of agreement with Grogono diagram. Moreover, it is noted that Siggaard-Andersen diagram presents lower ratios of agreement with the two specialists (48.3% to 56.4%) compared to the respective ratios of Grogono diagram and OSA software.

The combination of a computer software program with the Siggaard-Andersen diagram (OSA software) seems to improve the diagnostic value of the diagram. Hence, the OSA software presents higher ratios of agreement (p<0.001) with the other assessed methods compared to the Siggaard-Andersen diagram. The OSA software shows a higher ratio of agreement with the Grogono diagram (70.5%). Furthermore, although ratios of agreement between the OSA software program and specialist 1 or specialist 2 are moderate (61.1% and 67.4%, respectively), these ratios are significantly higher than those of the Siggaard-Andersen diagram.

Ratios of agreement between the two specialists and the other study methods show the greater variation ranging from 48.3% to 72.5%. Specialist 1 has a significantly higher ratio of diagnostic agreement (p<0.001) with Grogono diagram (72.5%), whereas specialist 2 presents a significantly higher diagnostic agreement with OSA software (67.4%, p<0.001). In addition, there was a significant difference (p<0.001) in the ratios of agreement between specialist 1 and the Siggaard-Andersen diagram (48.3%) and between specialist 2 and the Siggaard-Andersen diagram (56.4%). Finally, it is noted that there is a moderate agreement between the two specialists (67.6%).

Table 1 also contains ratios of agreement between each one of the three study methods and both specialists combined for all of arterial blood gas samples. Relevant data point out that ratios of agreement between either the Grogono diagram or the OSA software and both specialists combined are low (54.2% and 51.9%, respectively), and their difference is not significant. On the other hand, the Siggaard-Andersen diagram has an extremely low ratio of agreement with both specialists combined that hardly exceeds 40% and is significantly different from the respective ratios of Grogono diagram and OSA software.

Table 1

Table 1. Ratios of agreement between the assessed diagnostic methods and the two specialist physicians for the total of arterial blood gas samples (n=3122).

DISCUSSION

In the present study, we assess the diagnostic validity of a new diagram proposed by A. Grogono. It is a two-axial SBE-PaCO2 plot, all points of which represent a pair of values (for SBE and PaCO2). The diagram includes 23 well-defined areas, each one representing a particular acid-base balance disorder, either pure or mixed. Assessment of the diagnostic value of the new diagram in the clinical setting was based on multiple comparisons with the classic Siggaard-Andersen diagram, the Oxygen Status Algorithm (OSA) software and evaluations performed by two experienced physicians. To ensure the validity of all blood gas measurements, irrespective of the analyzer model that performed these measurements, we developed a special software program17-19, which has the ability to calculate all secondary parameters of acid-base balance using the results of the primary measurements.

Our results show that the overall ratios of agreement between the new diagram proposed by A. Grogono and the other diagnostic methods that were also assessed in this study were moderate. It is pointed out that the ratio of agreement between the Grogono diagram and the OSA software is significantly higher compared to that between the Grogono and the Siggaard-Andersen diagrams. In addition, there is a significant diagnostic variation between the Grogono diagram and the two specialists. Ratios of agreement between the Grogono diagram and the two specialists, although moderate, are significantly higher than those between the Siggaard-Andersen diagram and the two specialists. This finding, in combination with the low ratios of agreement between the Siggaard-Andersen diagram and the rest of diagnostic methods might be explained by the fact that the Siggaard-Andersen diagram includes only 8 diagnostic areas that are not representative of all acid-base balance disorders (either pure or mixed). The inherent inability to provide a definite diagnosis for every point plotted on this diagram is indicated by the fact that in the present study 273 arterial blood gas samples, which comprise 8.7% of all samples, using this method remained undiagnosed.

The combination of the Siggaard-Andersen diagram with a computer software program (Oxygen Status Algorithm software program) aims at eliminating the above-described limitations, thus improving the diagnostic value of the diagram. Hence, in all comparisons with other diagnostic methods, the OSA software shows significantly higher ratios of agreement compared to the classic Siggaard-Andersen diagram.

The overall ratios of agreement between the two physicians and the other diagnostic methods range from low to moderate. Consistent to previous reports,7-10 an impressive finding is the moderate overall ratio of agreement between the two physicians (67.6%). This ratio is considered rather low taking into account that both physicians were members of the same group of attending doctors that had the exclusive responsibility of the management and daily follow-up of all patients included in the present study. It might have been expected that the members of a treating group, especially physicians who have a long experience in the management of ICU patients, would present a much higher ratio of diagnostic agreement. The fact that they provide discordant diagnoses in 1/3 of the samples, as well as the significant variation between their opinion and the diagnoses provided by the assessed diagrammatic/computerized methods (specialist 1 "agrees" with the Grogono diagram, while specialist 2 with the OSA software) indicate that even today, despite rapid technological advance in laboratory investigation of acid-base balance disorders, discordance between experienced clinical doctors regarding diagnosis of blood gas disorders is quite common.

In order to investigate the great discordance between the two specialists, we made paired comparisons between the diagnosis derived from each one of the three methods for acid-base status assessment and the diagnosis of both specialists combined. We found that the Grogono diagram showed the greatest diagnostic agreement, which is not significantly different from that calculated for the OSA software. In contrast, the ratio of agreement between the Siggaard-Andersen diagram and the two specialists combined is much lower, confirming its lesser diagnostic value.

The results of the present study indicate that the use of computers has led to the development of diagrams and special software programs, such as the Grogono diagram and OSA software, which are of greater diagnostic value compared to the currently, and for a long time, widely used Siggaard-Andersen diagram. In particular, the proposed Grogono diagram shows the higher ratios of diagnostic agreement in several of the comparisons made between this diagram and the other diagnostic methods or the two experts. However, despite major advances in diagrammatic diagnosis of blood gas disturbances, as regards diagnostic value, none of the available methods is clearly superior to the others, so as to gain the universal acceptance from all clinicians and become the "standard diagnostic method" for the diagnosis of acid-base balance disorders.

These considerations lead to the conclusion that if the diagnostic efficacy of proposed methods is to be improved, the use of appropriate algebraic equations that describe as precisely as possible the respective acid-base disorder is not enough; a thorough revision of the currently used diagnostic methodology is necessary. In particular, all diagnostic methods proposed till the present day, be it either simple diagrams or computer programs4-6,9,11,12,25,26, present two main drawbacks:

1. They suggest only one possible diagnosis for a blood gas sample and do not allow for more than one possible or alternative diagnoses. Suggestion of alternative diagnoses is essential in certain complex cases. Besides, one should bear in mind that such diagnostic methods are not substitutes for the clinical doctor or valid enough to provide indisputable diagnoses. On the contrary, they serve as leads employed by clinical doctors; these leads combined with theoretical knowledge, clinical experience, additional laboratory measurements, physical examination and patient follow up will indicate the most appropriate among suggested diagnoses.

2. They are two-dimensional, i.e. they are based on two acid-base balance parameters not taking into account the third and most important parameter: time.11,27 Each blood gas sample is evaluated separately, as if it were independent of previous samples and past relevant changes in acid-base balance in a particular patient.

Conclusively, after many years of minimal research activity in the international literature, a new two-axial diagram is proposed; this diagram has been developed as a result of the universal acceptance of the limitations of a static diagrammatic method for the diagnosis of acid-base balance disorders which is no longer helpful in the accurate diagnostic investigation of blood gas disturbances. The results of the present study demonstrate that the diagram proposed by A.W. Grogono, despite its higher diagnostic agreement compared to the classic Siggaard-Andersen diagram, is not superior to the Oxygen Status Algorithm software program.

Therefore, a thorough revision of the currently applied diagnostic methodology has become imperative. Individual attempts to design new software programs for the diagnosis of acid-base balance disorders based on the dynamic recording of changes in primary parameters over time, as well as the suggestion of one or more alternative diagnoses, depending on the complexity of a given case, indicate the required changes in clinical practice that have already been initiated and will be extensively discussed in the near future.

 

REFERENCES

1. Astrup P, Severinghaus J: The History of Blood Gases, Acids and Bases. 1st Edition, Munksgaard, Denmark, 1986; 264-295.

2. Williamson JC. Acid-base disorders: classification and management strategies. Am Fam Physician 1995; 52(2):584.

3. Fall PJ. A stepwise approach to acid-base disorders. Practical patient evaluation for metabolic acidosis and other conditions. Postgrad Med 2000; 107:249-258.

4. Siggaard-Andersen O: An acid-base chart for arterial blood with normal and pathophysiological reference areas. Scand J Clin Lab Invest 1971; 27:239-245.

5. Davenport HW. The ABC of acid - base chemistry. 6th edition, The University of Chicago Press, Chicago, 1974.

6. Winters RW, Dell RB. Acid - Base Physiology in Medicine. 3d Edition, Little Brown and Company, Boston, Massachusetts, 1978.

7. Theakos N, Loukos A, Vassiliou M, Moukas M, Zaka M, Behrakis P. Α simplified pH-logPaCO2 diagram for the diagnostic approach of simple acid-base disorders. Current Contents 1998; 27-31.

8. Loukos A, Theakos N, Vassiliou M, Fragoulidou E, Zaka M, Behrakis P. A simplified pH-logPaCO2 diagram for the diagnostic approach of mixed alkalemic acid-base disorders. Current Contents 1998; 41-45.

9. Loukos A, Theakos N, Vassiliou M, Fragoulidou E, Moukas M, Behrakis P. A simplified logPaCO2 diagram for the diagnostic approach of mixed acidemic acid-base disorders. Current Contents 1998; 47-51.

10. Θεάκος N, Λούκος A, Βασιλείου M, Μπεχράκης Π. Αξιοπιστία και πρακτικότητα ενός νέου απλοποιημένου διαγράμματος pH-PaCO2 για τη διάγνωση των διαταραχών της οξεοβασικής ισορροπίας. ΠΝΕΥΜΩΝ 1998; 11:133-143.

11. Schlichtig R, Grogono A, Severinghaus J. Human PaCO2 and standard base excess compensation for acid-base imbalance. Crit Care Med 1998; 26:1173-1179.

12. Grogono A. http://www.acid-base.com/homepage.html, 2002

13. Harsten A, Berg B, Inerot S, Muth L. Importance of correct handling of samples for the results of blood gas analysis. Acta Anaesthesiol Scand 1988; 32:365-368.

14. Ng R, Dennis R, Yeston N, McNamara-Ward J, Statland B. Factitious cause of unexpected arterial blood-gas results. N Engl J Med 1984; 310:1189-1190.

15. AARC clinical practice guideline. Sampling for arterial blood gas analysis. American Association for Respiratory Care. Respir Care 1992; 37:913-917.

16. AARC clinical practice guideline. In-vitro pH and blood gas analysis and hemoximetry. American Association for Respiratory Care. Respir Care 1993; 38:505-510.

17. Wooten E. Analytic calculation of physiological acid-base parameters in plasma. J Appl Physiol 1999; 86:326-334.

18. Maassen N, Boning D. Blood osmolality in vitro: dependence on base addition, buffer value and temperature. J Appl Physiol 1987; 62:2174-2179.

19. Siggaard-Andersen O, Wimberley P, Fogh-Andersen N, Gothgen I. Measured and derived quantities with modern pH and blood gas equipment: calculation algorithms with 54 equations. Scand J Clin Lab Invest 1988; 48:7-15.

20. Siggaard-Andersen O, Wimberley P, Gothgen I, Fogh-Andersen N, Rasmussen J. Variability of the temperature coefficients for pH, PCO2, and PO2 in blood. Scand J Clin Lab Invest 1988; 48:85-88.

21. Severinghaus J. Acid-base balance nomogram -a Boston-Copenhagen detente. Anesthesiology 1976; 45:539-541.

22. Siggaard-Andersen O, Siggaard-Andersen Μ. The oxygen status algorithm: a computer program for calculating and displaying pH and blood gas data. Scand J Clin Lab Invest Suppl 1990; 203:29-45.

23. Siggaard-Andersen M, Siggaard-Andersen O. Oxygen status algorithm, version 3, with some applications. Acta Anaesthesiol Scand Suppl 1995; 107:13-20.

24. Larsen V, Siggaard-Andersen O. The Oxygen Status Algorithm on-line with the pH-blood gas analyzer. Scand J Clin Lab Invest Suppl 1996; 224:9-19.

25. Hekking M, Lindemans J, Gelsema E. A computer program for the multivariate and graphical monitoring of acid-base data in an intensive care unit. Proc Ann Symp Comput Appl Med Care 1995; 52-56.

26. Williams A. ABC of oxygen: assessing and interpreting arterial blood gases and acid-base balance. BMJ 1998; 317:1213-1216.

27. Sasse S, Chen P, Mahutte C. Variability of arterial blood gas values over time in stable medical ICU patients. Chest 1994; 106:187-193..

References