Comparison of Hemoglobin Values Obtained by Arterial Blood Gas Analysis versus Laboratory Method during Major Head-and-Neck Surgeries.

Background: Accuracy of hemoglobin (Hb) measured by arterial blood gas (ABG) analyzer is considered inferior to laboratory (lab) measurements as it could overestimate Hb levels. Aim of the Study: The study aims to compare Hb measured using ABG versus conventional lab method at the time of major blood loss and in the preoperative and immediate postoperative periods. Settings and Design: It was a prospective, nonrandomized observational study conducted in a tertiary care center. Materials and
Methods: The study was conducted in 24 patients undergoing major head-and-neck surgeries. Simultaneous blood samples were sent for Hb measurement by ABG analysis and lab method at induction of anesthesia, when intraoperative blood loss exceeded maximum allowable blood loss, and in the immediate postoperative period. Statistical Analysis Used: Chi-square test, independent sample's t-test, and paired t-test were used for statistical analysis.
Results: Mean Hb values obtained by both techniques were significantly different at all time points. Hb obtained by ABG analysis was significantly higher than lab value preoperatively (12.78 ± 2.51 vs. 12.05 ± 2.2, P = 0.038), at maximum blood loss (11.00 ± 2.57 vs. 9.87 ± 2.06, P = 0.006), and in the immediate postoperative period (11.96 ± 2.00 vs. 10.96 ± 2.24 P < 0.001). ABG Hb values were found to be approximately 1 g.dL-1 greater than lab values.
Conclusion: Hb measured by ABG analysis was significantly higher than that measured by lab method at the time of major blood loss, preoperatively, and at the immediate postoperative period in patients undergoing major head-and-neck surgeries, with a good correlation of values obtained by both the techniques. Copyright:

INTRODUCTION

Peroperative transfusions are guided by blood loss assessment and hemoglobin (Hb) values. Hb measured by arterial blood gas (ABG) analyzer is convenient, easily available, and results being obtained faster than the conventional laboratory (lab) methods. However, accuracy of Hb measured from ABG analysis is inferior to lab measurements,[1] and it could overestimate Hb levels.[2] A reliable correlation of values obtained by gas analyzer and lab method in the surgical population is lacking.

The primary objective of the present study was to compare Hb concentration measured using ABG analyzer versus conventional lab method at the time of major blood loss in patients undergoing major head-and-neck surgeries. The secondary objectives were to determine the comparison of Hb values obtained by these two techniques preoperatively and in the immediate postoperative period. The hemodynamic parameters, blood loss, volume of crystalloids, colloids, and blood transfused, and use of vasopressors were also assessed. We also investigated to find out the correlation of the Hb values by both methods so as to find out whether it was possible to predict lab Hb value from the ABG values.

MATERIALS AND METHODS

This was a prospective, nonrandomized observational study conducted after obtaining Institutional Ethical Committee Clearance (IEC-AIMS-2019-ANES-272 dated December 28, 2019) and was registered prospectively in the Clinical Trial Registry India (CTRI/2020/01/022944) and all the procedures followed the guidelines laid down in the Declaration of Helsinki.

Twenty-four patients aged 18–60 years, undergoing major head-and-neck surgeries in which significant blood loss was anticipated, with a preoperative Hb value of >10 g.dL−1, belonging to American Society of Anesthesiologists Physical Status Classes of I-III were included in the study. Those on anticoagulants and diuretics and with bleeding disorders, uncontrolled hypertension, and ischemic heart disease were excluded. Written informed consent was obtained from all patients who were recruited for participation in the study, as well as for the use of data for research and educational purposes.

All patients were kept fasting for 6 h preoperatively. After shifting the patient to the operation theater, standard monitors such as pulse oximeter, noninvasive blood pressure monitor, and electrocardiograph were attached. Arterial cannulation under local anesthesia was performed with 20G cannula on one radial artery, and two large bore intravenous (i.v.) cannulae were also placed. Arterial blood was withdrawn as two samples before induction of anesthesia. One sample was sent for ABG analysis and other for conventional lab method for the estimation of Hb. ABG analyzer used was Radiometer ABL 800 Basic, Denmark, and lab Hb estimation was done with Sysmex XN-2000™ Hematology System, Japan, using sulfolyser method.

All patients received anesthesia following a standard protocol. After induction, patients were hydrated with Ringer's lactate at the rate of mL.kg−1.h−1. Maximum allowable blood loss (MABL) was calculated using the formula.[3]

MABL = (body weight × 70 [preoperative Hb − target Hb])/average Hb

Average Hb = (preoperative Hb + target Hb)/2.

Intraoperative blood loss was assessed by weighing blood soaked surgical swabs, measuring volume of blood loss in suction apparatus, and loss at surgical site was also taken into account. Up to MABL, intravascular volume was replaced with either Ringer's lactate or Kabilyte (based on lactate levels) 1.5–2 times the volume of loss or equal volume of colloid (hydroxyl ethyl starch [HES] up to 20 mL.kg−1) or albumin (if there was existing preoperative hypoalbuminemia) as appropriate.

Packed red blood cells were transfused when the blood loss had exceeded MABL or when blood loss was less than MABL but associated with hypotension or at a Hb level of <8 g.dL. The second set of arterial blood samples was sent for analysis at this point of time. Fresh frozen plasma and platelets were transfused when massive blood transfusion was required.

Third sample was collected at the immediate postoperative period after shifting patient to the intensive care unit (ICU). Intraoperative blood loss, hemodynamic parameters such as heart rate (HR), mean arterial pressure (MAP), need of vasopressors or inotropes, and total volume of crystalloids, colloids, packed cells, and albumin, if used, were documented and analyzed.

Based on the results of mean and standard deviation of Hb of arterial blood samples measured by ABG analyzer (12.47 ± 1.41) and conventional lab method (10.72 ± 1.17) observed in an earlier publication[4] with 90% power and 95% confidence interval, the minimum sample size required to obtain statistically significant results was calculated to be eight patients. However, we included 24 patients in our study after taking informed consent before conduct of the study.

Statistical analysis

Paired t-test was used to compare the mean Hb of ABG and lab values. Pearson correlation was used to find the correlation of Hb between ABG and lab. Simple linear regression analysis was used to estimate the coefficient and prediction model for lab Hb by ABG. Statistical analyses were conducted using SPSS Version 20.0 for Windows (IBM Corporation, Armonk, NY, USA).

RESULTS

Twenty-six patients were assessed for eligibility, but only 24 patients comprising 75% males and 25% females were enrolled and their data were analyzed. Mean age of the patients was 58.65 ± 10.43 years, height 160 ± 7.65 cm, and weight 62 ± 12.64 kg. The mean Hb values obtained by both techniques were significantly different at all time points. Hb value obtained by ABG analysis was significantly higher than lab value preoperatively (12.78 ± 2.51 vs. 12.05 ± 2.2, P = 0.038), at maximum blood loss (11.00 ± 2.57 vs. 9.87 ± 2.06, P = 0.006), and in the immediate postoperative period (11.96 ± 2.00 vs. 10.96 ± 2.24, P < 0.001) [Figure 1 and Table 1].

Figure 1

Comparison of ABG and lab Hb at various time points. ABG = Arterial blood gas, Hb = Hemoglobin, Lab = Laboratory

Table 1

Comparison of mean hemoglobin values

Time	Mean±SD		Mean difference (95% CI)	P
	Hb lab	ABG Hb
Baseline	12.05±2.2	12.78±2.51	0.73 (0.04–1.42)	0.038
At maximum blood loss	9.87±2.06	11.00±2.57	1.13 (0.35–1.92)	0.006
End of surgery	10.96±2.24	11.96±2.00	1.00 (0.56–1.43)	<0.001

ABG=Arterial blood gas, SD=Standard deviation, Hb=Hemoglobin, CI=Confidence interval

Results of correlation and regression are shown in Table 2. A simple linear regression was calculated to predict the lab Hb based on ABG at three time points. At baseline, a significant regression equation was found (F(1, 21) = 32.890, P < 0.001), with a coefficient of determination (R2) of 0.610. Predicted lab Hb was equal to 3.221 ± 0.691 (ABG) g.dL−1 (t = 5.735, slop = 0.691). When ABG was measured in g.dL−1, the average lab Hb decreased 0.691 g.dL−1 for each g.dL−1 of value obtained from ABG. Linear regression of Hb preoperatively is shown in Figure 2.

Table 2

Correlation and regression of hemoglobin values and corresponding hemodynamic variables

Correlation and regression of hemoglobin values
Time	Correlation coefficient (r)	P
Baseline	0.781	<0.001
At maximum blood loss	0.717	<0.001
End of surgery	0.893	<0.001
Mean hemodynamic parameters at sample collection time points
Time points	HR (per min)	MAP (mmHg)
Postinduction	93.87±15.493	98.43±14.718
During major blood loss	83.09±18.133	64.70±9.093
End of surgery	85.04±18.758	89.74±15.259

HR=Heart rate, MAP=Mean arterial pressure

Figure 2

Linear regression of preoperative Hb values. ABG = Arterial blood gas, Hb = Hemoglobin, Lab = Laboratory

At the time of major blood loss, the regression equation was calculated (F(1, 21) = 22.172, P < 0.001), with a coefficient of determination (R2) of 0.514. Predicted lab Hb was equal to 3.544 ± 0.575 (ABG) g.dL−1 (t = 4.709, slop = 0.575). The average lab Hb decreased 0.575 g.dL−1 for each g.dL−1 of ABG. Figure 3 depicts linear regression of Hb values at major blood loss.

Figure 3

Linear regression of Hb at major blood loss. ABG = Arterial blood gas, Hb = Hemoglobin, Lab = Laboratory

At the end of surgery, the regression equation was calculated (F(1, 21] = 82.270, P < 0.001), with a coefficient of determination (R2) of 0.787. Predicted lab Hb was equal to 0.998 (ABG) ‒0.969 g.dL−1 (t = 9.070, slop = 0.998). The average lab Hb decreased 0.998 g.dL−1 for each g.dL−1 of ABG value. Figure 4 shows the linear regression at the immediate postoperative period.

Figure 4

Linear regression of Hb at immediate postoperative period. ABG = Arterial blood gas, Hb = Hemoglobin, Lab = Laboratory

Based on the linear regression, it was seen that the lab Hb could be predicted at different time points from the Hb value obtained in ABG using the following formulae.

At maximum blood loss, the predicted lab Hb = 3.5 ± 0.6(ABG Hb) g.dL−1

Preoperative predicted lab Hb = 3.2 ± 0.7 (ABG Hb) g.dL−1

Postoperative predicted lab Hb= (ABG Hb) − 1 g.dL−1.

Thirteen (54%) patients received packed cell transfusion, the mean volume being 258.69 ± 257.35 mL. Three patients (12.5%) required noradrenaline infusion to maintain blood pressure. Mean blood loss was 978.26 ± 451.04 mL. Mean volume of crystalloids used when the second sample collected was 2717.39 ± 1106.04 mL and at end of surgery was 5360 ± 1675.964 mL. Nine patients (37.5%) received colloids and the mean volume of colloids used was 195.65 ± 249.50 mL HES. Three patients (12.5%) received albumin. Duration of surgery was 518.04 ± 178.81 min. Mean HR and MAP at three time points of sample collection are shown in Table 2.

DISCUSSION

The results of blood investigations processed by a central lab are invariably delayed by a few hours[5] as compared to blood gas analyzers, which provide the measurements at the point-of-care and usually in a time frame of 2 min.[6] Although literature search shows numerous studies that compared both techniques of Hb estimation, the results are not conclusive and contradicting most of the time. However, it is generally assumed that ABG analyzer provides a reasonable estimate of the Hb concentration over a broad array of values.

It was shown that ABG analysis typically overestimates the Hb value by approximately 4.3 g.dL−1 than lab method in critically ill adults.[1] A significant difference in the mean Hb concentration between the two modalities of 0.91 g.dL−1 (range: 0‒4.3 g.dL−1) was observed when samples from patients undergoing ABG analysis and simultaneous venous sampling without therapeutic intervention between sampling were compared.[7]

However, in another study, where paired tests from five ICUs were compared, they did not show any statistically significant difference between the results of the blood gas and lab autoanalyzers for Hb (mean difference −0.35 g.dL−1).[8] When paired blood gas analyzer and lab samples taken not >10 min apart were compared, it was seen that >95% of results of Hb concentrations fell within the defined clinically acceptable limits, though Hb serum levels were slightly lower than blood gas levels.[9] Another study based in the emergency department also did not reveal any statistical differences and biases between ABG- and lab-measured Hb.[10] Most of these studies were done in critically ill patients in ICU setting.

All patients in our study were fasted 6–8 h before surgery, and hence, the first sample pair demonstrated correlation of Hb values estimated by ABG versus lab method when there is a certain degree of dehydration. The second sample pair was taken when there was major blood loss, which was replaced with i.v. fluids and hence showed correlation of Hb values by both techniques in the presence of hemodilution. The third comparison of values at the end of surgery showed the reliability of values by both techniques following therapeutic interventions such as fluid resuscitation and blood transfusions.

Very little is known on the comparison of Hb measurements by ABG versus conventional lab method in surgical patients. The effects of dehydration, acute blood loss, hemodilution, and therapeutic intervention on different methods of Hb estimation are also not well analyzed. As these conditions mimic hemoconcentration, hemodilution, and near-normal conditions, respectively, we tried to find out different formulae for predicting lab Hb values from ABG at these three time points. Although it is recommended that during hemorrhagic surgery, a point-of-care hemoglobinometer measuring arterial blood provides values closest to central lab measurements even when Hb concentrations are rapidly changing.[11] However, during major surgeries with massive bleeding and rapid fluid shifts Hb values from ABG samples only will be immediately accessible.

Many factors other than blood loss contribute to intraoperative hypotension during long surgical procedures. Hypothermia being one of them is quite common during prolonged major surgeries which might have contributed to hypotension in our study. Use of different i.v. anesthetics and volatile agents pertaining to individual surgical demands also influences intraoperative hemodynamic parameters.

The major drawback of our study was that the blood loss estimation was conventional and hence not very accurate. Some cases even with Hb >8 g.dL−1 received blood due to hypotension or associated comorbidities. Three ABG machines from the same manufacturer were used in the study. There is a possibility of variation in readings from ABG machines of other manufacturers. However, the whole purpose of our study was to predict an approximate value of lab Hb from ABG intraoperatively. The predictions from the formulae provided will only be approximate, but useful to alert the anesthesiologists about need for an impending transfusion before a lab Hb value could be obtained intraoperatively. Further studies with larger sample size and using ABG machines of different manufacturers may warrant more accurate and uniform results.

CONCLUSION

In patients undergoing major head-and-neck surgeries, the Hb measured by ABG analysis was significantly higher than that measured by conventional lab method at the time of major blood loss, preoperatively, and in the immediate postoperative period. There was good correlation of Hb values measured by both the techniques at all these time points, allowing a scope for accurate prediction of Hb from the ABG analysis reports.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.