Utility of Sequential Blood Gas Testing in Medical Intensive Care Unit
Seema Agarwal, M.D., Stephen R Karbowitz, M.D., F.C.C.P
Department of Medicine, Pulmonary Critical Care Division
Introduction: Blood gas analysis is an important tool in assessing ventilatory and metabolic status of a patient in ICU. Arterial Blood Gas (ABG) has the potential to significantly influence patient management. However, repeated ABG may not have a significant bearing on the patient management. In this study we determined the utility of sequential ABGs in patient treatment. ABG were ABG as either is “Routine ABG” performed as morning blood work or physician-ordered ABG “Physician ABG”. In this study we evaluated usefulness of Physician ABGs versus Routine ABGs. We studied the issue both through literary review as well as quantitative analysis. This gave us insight into effectiveness of sequential blood gas analysis in ICU.
Material and Methods: All patients admitted to MICU at NYHQ during 9/22/07-10/19/07 who had 2 or more blood gas analyses in any two day period of stay were enrolled in the study. Patient demographic profile, reason for ABG, the data from ABG and management steps taken in response to information from ABG was noted. It was also noted whether ABG was part of daily blood work (Routine ABG) or Physician-ordered ABG. Blood gases were given a utility score of 1 to 4 in increasing order of usefulness of the management decision taken on basis of the ABG result. Data was analyzed to find determinants of utility score of blood gases.
Observations and results: We collected data for forty-five patients- ranging in age from 24 years to 88 years. Four hundred ABG were analyzed and their utility scores determined. Out of these four hundred ABG, 238 were Physician ABG and the rest were Routine ABG. We were able to quantitatively establish five trends. Older patients get more ABG. Physicians order ABG when it is expected to have a significant influence on patient management. Over a quarter of Routine ABG do not provide significant treatment benefit to the patient. Physicians often use ABG as a tool for managing acidosis. Finally, the utility of each subsequent ABG goes down as the number of sequential ABG increase.
ABG and Age: We found that the number of ABG increase with patient age. We divided the data into four age groups. The youngest age group (24 to 43 years) had an average of 5.6 ABG per patient, while the oldest age group (84+ years) had 11.5 ABG per patient. Average number of ABG was 6.0 and 10.3 for age groups 44 to 63 yrs and 64 to 83 years, respectively. The correlation coefficient between age and number of ABG ordered (30%) was significant at p<0.05. We further ranked the patients by age and the number of ABG. The rank correlation (41%) was significant at p<0.01.
Physician ABG vs. Routine ABG: 26% of the Routine ABG had a utility score of one, while the corresponding number for Physician ABG was 8%. Further, 20% of all Physician ABG had the utility score of 4 compared to 2% for Routine ABG. Physicians tend to order ABG when it is expected to have a significant influence on patient management. 26% Routine ABG do not provide significant treatment benefit to the patient.
Acidosis Management: Doctors use ABG as a tool for managing acidosis. Out of a total of 400 ABG, nearly half (202) were ordered for patients with normal pH. However, of the remaining 198 ABG, 144(73%) ABG were ordered for patients with acidosis (pH 7.3 and below). Remarkably, out of these 144 ABG, 81% were Physician ABG. 67% of this Physician ABG had a utility score of 3 or 4. This suggests physicians effectively utilize ABG in management of acidosis.
Marginal Utility of ABGs: The marginal utility of ABG goes down as the number per patient increase; 89% of all first ABG had a utility of 3 or 4. The comparable number for 2nd, 3rd, 4th and 5th ABG were 69%, 63%, 60% and 36%, respectively. The negative correlation between the “sequential serial number” of an ABG and its resultant utility is negative and significant at p<<. 01. We calculated an alternate measure of correlation; specifically the correlation between serial number of all ABG and the average utility score for that particular serial number; the correlation was even higher and significant at p<<0.01.
Conclusion: Performing ABG is an important tool in diagnosing and treating various ventilatory and metabolic derangements, especially in critical care settings. Our study brings five important observations. Older patients get more ABG and the ABG lose their marginal utility as increasing numbers of tests are done. A large number of ABG are done to manage acidosis and ICU physicians order most of these ABG; at least 2/3 rd of them was crucial for patient management (meaning had a high utility scores). At least one quarter of routine ABG did not add to management. This suggests a need to limit routine ABG in older patients, especially if they have a prolonged ICU stay.