Transforming Hemoglobin Measurement in Trauma Patients: Noninvasive Spot Check Hemoglobin Bellal Joseph, MD, FACS, Viraj Pandit, MD, Hassan Aziz, MD, Narong Kulvatunyou, MD, FACS, Bardiya Zangbar, MD, Andrew Tang, MD, FACS, Terence O’ Keeffe, MBChB, FACS, Qasim Jehangir, Kara Snyder, RN, Peter Rhee, MD, FACS
MD,
Technological advances now allow for noninvasive Hbg measurements. Previous studies have reported on the efficacy of continuous noninvasive Hgb devices. Recently, a new device, Pronto-7, a spot check pulse CO-oximeter has become available. The aim of our study was to assess noninvasive Hgb measurement in trauma patients. METHODS: We performed a prospective cohort analysis of all trauma patients presenting to our Level I trauma center. Invasive Hgb and spot check Hgb measurements were obtained simultaneously at presentation. Spot check was measured 2 times with each invasive Hgb value. Normal Hgb was defined as >8 mg/dL. Spearman correlation and Bland-Altman analysis was performed. RESULTS: A total of 525 patients had attempted spot check Hgb measurements with a success rate of 86% (n ¼ 450). A total of 450 invasive and 1,350 spot check Hgb measurements were obtained. Mean SD age was 41 21 years, 74% were male, and mean Injury Severity Score was 21 13. Thirty-eight percent (n ¼ 173) of patients had Hgb 8 mg/dL at presentation. Mean invasive Hgb was 11.5 4.36 g/dL, mean spot check Hgb 11.1 3.60 g/dL, and mean difference was 0.3 1.3 g/dL. Spot check Hgb values had strong correlation with invasive Hgb measurements (R2 ¼ 0.77; R ¼ 0.86; p ¼ 0.04) with 76% accuracy and 95.4% sensitivity. CONCLUSIONS: Spot check Hgb monitoring had excellent correlation with invasive Hgb measurements. Application of spot check has more clinical use as compared with previous continuous Hgb monitoring. This novel technology allows immediate and accurate Hgb measurements in trauma patients. (J Am Coll Surg 2015;220:93e98. 2015 by the American College of Surgeons)
BACKGROUND:
Hemorrhage remains a major cause of potentially preventable deaths.1-3 Early recognition of hypovolemia due to hemorrhage can be challenging, especially in trauma patients, as the clinical signs and symptoms are nonspecific and can be confounded by multiple factors, such as traumatic brain injury, anxiety, or drug intoxication.3
Additionally, hypotension is a late sign, and delaying therapy until this phenomenon occurs is associated with an increase in mortality.1-4 Hemoglobin measurements are one of the most frequently ordered laboratory tests in the trauma setting, as these can help guide therapeutic plan of management.5 Hemoglobin measurements have traditionally required an invasive blood draw and although the invasive method is accurate, it has its own limitations, such as delay in therapeutic plan.6,7 The ability to measure Hgb noninvasively can allow for a more rapid assessment of a patient’s condition and need for transfusion, enabling prompt and appropriate clinical management (requirement of blood transfusion, requirement of operative intervention).8 Our institution has recently shown a weak correlation between the Hbg measurements obtained from continuous noninvasive pulse oximeter and invasive blood draws.3 Recently, a new device, Pronto-7, a spot check pulse CO-oximeter, was developed. Pronto-7 spot check
CME questions for this article available at http://jacscme.facs.org Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. Abstract presented at the 27th Annual Scientific Assembly of the Eastern Association for the Surgery of Trauma, Naples, FL, January 2014. Received July 23, 2014; Revised September 9, 2014; Accepted September 25, 2014. From the Division of Trauma, Emergency Surgery, Critical Care, and Burns, Department of Surgery, University of Arizona, Tucson, AZ. Correspondence address: Bellal Joseph, MD, FACS, Division of Trauma, Critical Care and Emergency Surgery, Department of Surgery, University of Arizona, 1501 N Campbell Ave, Rm 5411, PO Box 245063, Tucson, AZ 85724. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
93
http://dx.doi.org/10.1016/j.jamcollsurg.2014.09.022 ISSN 1072-7515/14
94
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
pulse CO-oximeter has a spectrophotometric sensor (Rainbow DCI) that senses multiple wavelengths of light.9 The pulse CO-oximetry method discerns the distinctive light-absorption characteristics of different Hgb species and applies proprietary algorithms to determine Hgb levels.9 The efficacy and accuracy of a spot check Hgb measuring device in trauma patients has not been established. The aim of this study was to evaluate the accuracy of the spot check noninvasive Hgb device compared with invasive laboratory Hgb measurements in trauma patients.
METHODS After approval from the IRB at the University of Arizona, we performed a prospective cohort analysis of all trauma patients presenting to our Level I trauma center between April 2012 and October 2012. For this study, we obtained a waiver of consent for enrollment in the study, however, consent was obtained from the patients for assessing patients’ personal health information. Study population We included all trauma patients with trauma team activation presenting to our trauma center. Patients transferred from other centers and patients dead on arrival were excluded from the study. Data points collected We prospectively collected the following data points: patient demographics (age, sex, race, and ethnicity); mechanism of injury (blunt or penetrating); vital parameters on presentation, including heart rate, systolic blood pressure, temperature, and Glasgow Coma Scale score; invasive and noninvasive Hgb measurement on admission; fluids and blood products transfused in the emergency department; need for operative intervention; hospital and intensive care unit length of stay; and in-hospital mortality. The Abbreviated Injury Scale score and Injury Severity Score were obtained from the trauma registry. Study protocol and hemoglobin measurements Trauma patients were approached by a single investigator on admission. Invasive and noninvasive Hgb was recorded simultaneously in each patient on presentation to the trauma center. The noninvasive Hgb measurement probe was placed on the patient’s finger and a nurse collected a venous blood sample in an ethylenediamine tetraacetic acid tube, which was sent immediately to the hematology laboratory for Hgb measurement by the Advia 2120 (Siemens Medical Solutions Diagnostics) method, which is considered to be the be the gold standard measure for recording invasive Hgb.3 Noninvasive Hgb was measured 3 times in each patient using the Masimo Pronto-7 (version
J Am Coll Surg
2.1.9).9 The first read was obtained on presentation to the trauma unit; second and third reads were obtained 5 minutes after the first and second measurements, respectively. Three reading were recorded to assess the reliability of noninvasive Hgb device to measure Hgb in trauma patients. Intraclass correlation test was performed to confirm the correlation between the 3 readings. Mean of the 3 readings was calculated and used for the correlation analysis. Outcomes measures Our primary outcomes measure was correlation and agreement between invasive and noninvasive spot check Hgb measurement. Secondary outcomes measures were correlation among 3 spot check Hgb measurements and predictive value (sensitivity, specificity, positive predictive value, negative predictive value, and accuracy) of the noninvasive Hgb measurement. Predictive value was calculated by grouping data based on the admission Hgb level as Hgb >8 mg/dL and Hgb 8 mg/dL. For our study, we defined normal Hgb value as Hgb >8 mg/dL. Statistical analysis Data are reported as the mean SD for continuous descriptive variables, the median (range) for ordinal descriptive variables, and the proportion for categorical variables. The intraclass correlation coefficient was used to assess correlation among the 3 noninvasive Hgb measurements. Spearman correlation test was performed to assess correlation among the invasive and the mean of 3 noninvasive spot check Hgb measurements. The agreement between invasive and noninvasive spot check Hgb was assessed using the Bland-Altman analysis. This method allows for quantification of the difference between the 2 techniques under consideration (invasive and noninvasive), and allows us to assess if the device (Masimo Pronto-7) has sufficient accuracy to allow acceptance of the device values in lieu of standard laboratory techniques.10,11 These data are summarized by the mean, the bias, and the 95% CI of the differences between techniques (limits of agreement). Bias was defined as the difference between the proposed method and the standard technique, and therefore estimated as the mean difference between Hgb readings from the Masimo and the laboratory blood draw (standard). The limits of agreement were calculated as (mean difference SD). Data were inserted into an Microsoft Excel 2007 work sheet and all statistical analysis was performed using Statistical Package for Social Sciences software (SPSS Inc).
RESULTS A total of 525 patients had attempted noninvasive spot check Hgb measurements with a success rate of 86%
Vol. 220, No. 1, January 2015
Table 1.
Joseph et al
Demographics of Study Population (n ¼ 525)
Variables
Age, y, mean SD Males, % (n) SBP, mmHg, mean SD HR, per min, mean SD GCS, mean SD ISS, mean SD AIS, median (range) Head, mean SD Chest, mean SD Abdomen, mean SD Extremity, mean SD
41 21 74 (333) 110 30 100 23 11 5 20 13 3 (1e5) 3.8 0.6 2.3 1 3.4 1.4 2.1 1.1
AIS, Abbreviate Injury Score; GCS, Glasgow Coma Score; HR, heart rate; ISS, Injury Severity Score; SBP, systolic blood pressure.
(n ¼ 450). A total of 450 invasive and 1,350 spot check Hgb measurements were obtained. Mean age was 41 21 years, 74% (n ¼ 333) were male, mean systolic blood pressure was 110 30 mmHg, and median Injury Severity Score was 17 (range 10 to 20). The main mechanism of injury was motor vehicle accidents (68%), followed by gunshot wounds (14%), and assault (8%). Table 1 highlights the demographics of the study population. The mean invasive Hgb was 11.5 4.36 g/dL (range 6 to 16 g/dL), mean noninvasive spot check Hgb was 11.1 3.60 g/dL (range 6.4 to 16.3 g/dL), and the difference in mean between invasive and noninvasive Hgb was 0.3 1.3 g/dL. There was no significant difference in the mean between invasive and noninvasive Hgb measurements (p ¼ 0.23). Thirty-eight percent (n ¼ 173) of patients had Hgb 8 g/dL on presentation based on invasive Hgb measurement. Twelve percent (n ¼ 54) received blood transfusion and 8% (n ¼ 36) underwent emergent surgical intervention for bleeding. After dichotomization of our patients into 2 groups, patient with Hgb 8 g/ dL and patients with Hgb >8 g/dL spot check Hgb measurements were found to have a sensitivity of 95.4% and an accuracy of 76%. Tables 2 and 3 demonstrate the dichotomization and predictive performance of the noninvasive spot check Hgb monitor. Noninvasive spot check Hgb measurement has a strong correlation with the invasive Hgb measurements (intraclass correlation ¼ 0.70; 95% CI, 0.57e0.80) and excellent correlation among the 3 consecutive noninvasive spot check Hgb measurements (intraclass correlation ¼ 0.90; 95% CI, 0.87e0.94). Spearman correlation plot revealed a strong correlation between invasive and noninvasive Hgb measurement (R2 ¼ 0.77; R ¼ 0.86; p ¼ 0.04). Figure 1 demonstrates the Spearman correlation plot between invasive and noninvasive Hgb measurement.
Noninvasive Hemoglobin Measurement in Trauma
95
Table 2. Dichotomy of Hemoglobin 8 mg/dL and >8 mg/dL
Variable
Spot check noninvasive Hgb measurement
8 mg/dL (n ¼ 265) >8 mg/dL (n ¼ 185)
Invasive Hgb measurement 8 mg/dL >8 mg/dL (n ¼ 173) (n ¼ 277)
165* 8
100 177
*Numbers in the table demonstrate number of patients (total n ¼ 450).
On subanalysis after stratifying patients into 2 groups (Hgb 8 g/dL and Hgb >8 g/dL), we found strong correlation between invasive and noninvasive Hgb measurement in patients with Hgb >8 g/dL (R2 ¼ 0.79; R ¼ 0.84; p ¼ 0.03). In patients with Hgb 8 g/dL, we found a good correlation between invasive and noninvasive Hgb measurement in patients with (R2 ¼ 0.71; R ¼ 0.80; p ¼ 0.04). Also, on assessing correlation in patients with admission Hgb between 10 and 13 mg/dL, we found good correlation (R2 ¼ 0.76; R ¼ 0.79; p ¼ 0.04) between invasive and noninvasive Hgb measurements. In our study, 20.9% (n ¼ 94) of patients were hypotensive (systolic blood pressure 90 mmHg) on admission. On assessing correlation in these patients, we found that there was good correlation between noninvasive spot check Hgb and invasive Hgb measurements (R2 ¼ 0.71; R ¼ 0.78; p ¼ 0.042). Figure 2 demonstrates the Bland-Altman analysis for invasive and noninvasive spot check Hgb measurements. The absolute mean difference (bias) was 0.096 g/dL, with an upper agreement limit at 4.016 g/dL and a lower agreement limit at 3.84 g/dL. Of all the measurements, 98.7% fell within 2 SDs of the mean difference. Mean hospital lengths of stay were 5 3 days, and mean ICU length of stay was 9 9 days. Overall mortality rate was 3% (16 of 525). The spot check Hgb probe failed to record the Hgb in 14% (n ¼ 75) of patients. Of the 75 patients in which Hgb could not be measured, 36 patients had nail polish or soot or tar on their fingers; 21 patients had difficulty in the sensor fitting, as we had only one size sensor for Table 3. Oximeter
Predictive Performance of Noninvasive Pulse
Variable
Sensitivity Specificity PPV NPV Accuracy NPV, negative predictive value; PPV, positive predictive value.
Value, %
95.4 63.8 62.2 95.6 76
96
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
Figure 1. Spearman correlation plot demonstrating strong correlation between invasive and continuous noninvasive Hgb monitoring.
our study; 10 patient had radiologic interference because of external factors, such as x-ray and ultrasound devices; and the remaining 8 patients’ Hgb could not be recorded due to anxious or agitated patients. On subanalysis of the 14% (n ¼ 75) of patients in whom the device could record the reading, the mean invasive Hgb was 9.1 2.30 g/dL, mean systolic blood pressure was 95.8 20.5 mmHg, and mean Injury Severity Score was 23 (range 21e28); 37.3% (n ¼ 28) were hypotensive and 11% (n ¼ 8) required operative intervention.
J Am Coll Surg
DISCUSSION The efficacy and accuracy of the Pronto-7 spot check pulse CO-oximeter (Masimo Rainbow SET; Masimo) in a trauma population has never been described. This is the first study to our knowledge to assess the use of a portable spot check pulse CO-oximeter in trauma patients. In our study, we found that the mean difference in Hgb values between spot check Hgb and invasive Hgb was 0.3 1.3 g/dL. We found excellent correlation (R2 ¼ 0.77) between invasive and spot check noninvasive Hgb measurements. We conclude that this novel technology allows for immediate and accurate Hgb measurements in trauma patients. Blood loss in trauma patients is routinely assessed based on Hgb and hematocrit measurement. Admission Hgb measurement has been shown to correlate with outcomes in trauma patients.12-14 Thorson et al14 showed that initial hematocrit and the change in hematocrit level have valuable information in transfusion requirements and are predictors of adverse outcomes in trauma patients. In our study, we assessed the reliability of noninvasive Hgb measurement compared with the invasive method, which is currently the standard of care in most trauma centers. We believe that implementing this new technology is going to benefit the patients by eliminating an unnecessary painful puncturing of the veins and unwanted complications. Also, the rapid measurement of Hgb level and ability to quickly repeat the measurement enables the trauma specialist to have an eye on patient’s hemodynamic stability. Point of care testing has recently expanded in many fields of laboratory medicine.8 In our study, we found
Figure 2. Bland-Altman agreement plot demonstrating strong agreement with 98.7% of reading within 2 SDs of agreement limits.
Vol. 220, No. 1, January 2015
Joseph et al
that spot check Hgb rapidly measured Hg, which accurately correlated to invasive Hgb. Laboratory measures for Hgb, although accurate, have limitations. Laboratory measures require blood samples, transport of the sample to a laboratory, and analysis, all of which are timeconsuming steps and can delay patient care; spot check Hgb monitoring can allow for an efficient and more rapid means for detection of hemorrhage, requirements of blood transfusion, and appropriate surgical management. We believe this device has the potential to improve clinical care, patient safety, and cost of care. In our study, we found excellent correlation between spot check Hgb and invasive Hgb. In a case series published in 2007, Macknet and colleagues15 compared noninvasive CO-Oximeter (Masimo Inc.) with invasive Hgb testing and concluded good correlation between the Hgb values obtained during times of rapidly changing Hgb concentrations due to surgical blood loss and transfusions.15 Similarly, in a study by Miller and colleagues,16 pulse COoximeter based spot check Hgb results were accurate to 1 g/dL (1 SD) when compared with laboratory measurements in healthy subjects undergoing hemodilution.16 Causey and colleagues17 demonstrated significant correlation between noninvasive Hgb monitoring with laboratory values in ICU patients and in general surgery patients undergoing elective operations.17 In another study on patients who underwent complex spine surgery, Berkow and colleagues8 demonstrated clinically acceptable accuracy of Hgb measurement within 1.5 g/dL between continuous noninvasive Hgb measurement via pulse CO-oximetry and standard laboratory reference device.8 We believe that, the new spot check Hgb monitor can play an important for quick assessment of Hgb in trauma patients. Despite excellent correlation with invasive Hgb measurement, one of the findings of concern was that spot check Hgb did not record data points 14% of the time. The inconsistency of spot check Hgb might be due to several factors, namely, low perfusion states, motion artifact, and disconnection of the sensor. Accordingly, the inability of the spot check Hgb monitor to recognize the trauma patient with life-threatening bleeding is the failure of this device, which needs to be further evaluated before implementation of this device in the clinical decision-making process. The results of this study should be interpreted in the context of its limitations. Through all invasive and spot check Hgb measurements were drawn simultaneously; however, there was no set protocol for the use of spot check Hgb probe. Second, we did not assess the effect of prehospital IV fluids administered during resuscitation on the accuracy of the device. Third, we did not perform a cost analysis. Despite these limitations, we demonstrate excellent correlation and
Noninvasive Hemoglobin Measurement in Trauma
97
agreement between noninvasive Hgb measurement technology and the invasive Hgb measurements.
CONCLUSIONS Spot check Hgb monitoring has excellent correlation with invasive Hgb measurements. Application of spot check has more clinical use as compared with previous continuous Hgb monitoring. This novel technology allows immediate and accurate Hgb measurements in trauma patients. Author Contributions Study conception and design: Joseph, O’Keeffe, Snyder, Rhee Acquisition of data: Joseph, Pandit, Aziz, Zangbar, Jehangir Analysis and interpretation of data: Joseph, Pandit, Kulvatunyou, Zangbar, Tang Drafting of manuscript: Joseph, Pandit, Aziz, Kulvatunyou, Tang Critical revision: Zangbar, O’Keeffe, Snyder, Rhee REFERENCES 1. Kauvar DS, Lefering R, Wade CE. Impact of hemorrhage on trauma outcome: an overview of epidemiology, clinical presentations, and therapeutic considerations. J Trauma 2006;60[Suppl]:S3eS11. 2. Acosta JA, Yang JC, Winchell RJ, et al. Lethal injuries and time to death in a level I trauma center. J Am Coll Surg 1998;186:528e533. 3. Joseph B, Hadjizacharia P, Aziz H, et al. Continuous noninvasive hemoglobin monitor from pulse ox: ready for prime time? World J Surg 2013;37:525e529. 4. Shapiro NI, Kociszewski C, Harrison T, et al. Isolated prehospital hypotension after traumatic injuries: a predictor of mortality? J Emerg Med 2003;25:175e179. 5. Gehring H, Hornberger C, Dibbelt L, et al. Accuracy of point of care testing (POCT) for determining hemoglobin concentrations. Acta Anaesthesiol Scand 2002;46:980e986. 6. Mokken FC, van der Waart FJ, Henny CP, et al. Differences in peripheral arterial and venous hemorheologic parameters. Ann Hematol 1996;73:135e137. 7. Yang ZW, Yang SH, Chen L, et al. Comparison of blood counts in venous, fingertip, and arterial blood and their measurement variation. Clin Lab Haematol 2001;23:155e159. 8. Berkow L, Rotolo S, Mirski E. Continuous noninvasive hemoglobin monitoring during complex spine surgery. Anesth Analg 2011;113:1396e1402. 9. Masimo. Pronto 7. Available at: http://www.masimo.com/pdf/ pronto-7/LAB6400G_Brochure_Pronto-7.pdf. Accessed August 21, 2014. 10. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8: 135e160. 11. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346[8982]:1085e1087.
98
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
12. Thorson CM, Van Haren RM, Ryan ML. Admission hematocrit and transfusion requirements after trauma. J Am Coll Surg 2013;216:65e73. 13. Ryan ML, Thorson CM, Otero CA. Initial hematocrit in trauma: a paradigm shift? J Trauma Acute Care Surg 2012; 72:54e59; discussion 59e60. 14. Thorson CM, Ryan ML, Van Haren RM, et al. Change in hematocrit during trauma assessment predicts bleeding even with ongoing fluid resuscitation. Am Surg 2013;79:398e406.
J Am Coll Surg
15. Macknet M, Norton S, Kimball-Jones P, et al. Continuous noninvasive measurement of hemoglobin via pulse CO-oximetry. Anesth Analg 2007;105:S-108. 16. Miller RD, Ward TA, Shiboski SC, et al. Comparison of three methods of hemoglobin monitoring in patients undergoing spine surgery. Anesth Analg 2011;112:858e863. 17. Causey MW, Miller S, Foster A, et al. Validation of noninvasive hemoglobin measurements using the Masimo Radical-7 SpHb Station. Am J Surg 2011;201:590e596.
MD,
Technological advances now allow for noninvasive Hbg measurements. Previous studies have reported on the efficacy of continuous noninvasive Hgb devices. Recently, a new device, Pronto-7, a spot check pulse CO-oximeter has become available. The aim of our study was to assess noninvasive Hgb measurement in trauma patients. METHODS: We performed a prospective cohort analysis of all trauma patients presenting to our Level I trauma center. Invasive Hgb and spot check Hgb measurements were obtained simultaneously at presentation. Spot check was measured 2 times with each invasive Hgb value. Normal Hgb was defined as >8 mg/dL. Spearman correlation and Bland-Altman analysis was performed. RESULTS: A total of 525 patients had attempted spot check Hgb measurements with a success rate of 86% (n ¼ 450). A total of 450 invasive and 1,350 spot check Hgb measurements were obtained. Mean SD age was 41 21 years, 74% were male, and mean Injury Severity Score was 21 13. Thirty-eight percent (n ¼ 173) of patients had Hgb 8 mg/dL at presentation. Mean invasive Hgb was 11.5 4.36 g/dL, mean spot check Hgb 11.1 3.60 g/dL, and mean difference was 0.3 1.3 g/dL. Spot check Hgb values had strong correlation with invasive Hgb measurements (R2 ¼ 0.77; R ¼ 0.86; p ¼ 0.04) with 76% accuracy and 95.4% sensitivity. CONCLUSIONS: Spot check Hgb monitoring had excellent correlation with invasive Hgb measurements. Application of spot check has more clinical use as compared with previous continuous Hgb monitoring. This novel technology allows immediate and accurate Hgb measurements in trauma patients. (J Am Coll Surg 2015;220:93e98. 2015 by the American College of Surgeons)
BACKGROUND:
Hemorrhage remains a major cause of potentially preventable deaths.1-3 Early recognition of hypovolemia due to hemorrhage can be challenging, especially in trauma patients, as the clinical signs and symptoms are nonspecific and can be confounded by multiple factors, such as traumatic brain injury, anxiety, or drug intoxication.3
Additionally, hypotension is a late sign, and delaying therapy until this phenomenon occurs is associated with an increase in mortality.1-4 Hemoglobin measurements are one of the most frequently ordered laboratory tests in the trauma setting, as these can help guide therapeutic plan of management.5 Hemoglobin measurements have traditionally required an invasive blood draw and although the invasive method is accurate, it has its own limitations, such as delay in therapeutic plan.6,7 The ability to measure Hgb noninvasively can allow for a more rapid assessment of a patient’s condition and need for transfusion, enabling prompt and appropriate clinical management (requirement of blood transfusion, requirement of operative intervention).8 Our institution has recently shown a weak correlation between the Hbg measurements obtained from continuous noninvasive pulse oximeter and invasive blood draws.3 Recently, a new device, Pronto-7, a spot check pulse CO-oximeter, was developed. Pronto-7 spot check
CME questions for this article available at http://jacscme.facs.org Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. Abstract presented at the 27th Annual Scientific Assembly of the Eastern Association for the Surgery of Trauma, Naples, FL, January 2014. Received July 23, 2014; Revised September 9, 2014; Accepted September 25, 2014. From the Division of Trauma, Emergency Surgery, Critical Care, and Burns, Department of Surgery, University of Arizona, Tucson, AZ. Correspondence address: Bellal Joseph, MD, FACS, Division of Trauma, Critical Care and Emergency Surgery, Department of Surgery, University of Arizona, 1501 N Campbell Ave, Rm 5411, PO Box 245063, Tucson, AZ 85724. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
93
http://dx.doi.org/10.1016/j.jamcollsurg.2014.09.022 ISSN 1072-7515/14
94
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
pulse CO-oximeter has a spectrophotometric sensor (Rainbow DCI) that senses multiple wavelengths of light.9 The pulse CO-oximetry method discerns the distinctive light-absorption characteristics of different Hgb species and applies proprietary algorithms to determine Hgb levels.9 The efficacy and accuracy of a spot check Hgb measuring device in trauma patients has not been established. The aim of this study was to evaluate the accuracy of the spot check noninvasive Hgb device compared with invasive laboratory Hgb measurements in trauma patients.
METHODS After approval from the IRB at the University of Arizona, we performed a prospective cohort analysis of all trauma patients presenting to our Level I trauma center between April 2012 and October 2012. For this study, we obtained a waiver of consent for enrollment in the study, however, consent was obtained from the patients for assessing patients’ personal health information. Study population We included all trauma patients with trauma team activation presenting to our trauma center. Patients transferred from other centers and patients dead on arrival were excluded from the study. Data points collected We prospectively collected the following data points: patient demographics (age, sex, race, and ethnicity); mechanism of injury (blunt or penetrating); vital parameters on presentation, including heart rate, systolic blood pressure, temperature, and Glasgow Coma Scale score; invasive and noninvasive Hgb measurement on admission; fluids and blood products transfused in the emergency department; need for operative intervention; hospital and intensive care unit length of stay; and in-hospital mortality. The Abbreviated Injury Scale score and Injury Severity Score were obtained from the trauma registry. Study protocol and hemoglobin measurements Trauma patients were approached by a single investigator on admission. Invasive and noninvasive Hgb was recorded simultaneously in each patient on presentation to the trauma center. The noninvasive Hgb measurement probe was placed on the patient’s finger and a nurse collected a venous blood sample in an ethylenediamine tetraacetic acid tube, which was sent immediately to the hematology laboratory for Hgb measurement by the Advia 2120 (Siemens Medical Solutions Diagnostics) method, which is considered to be the be the gold standard measure for recording invasive Hgb.3 Noninvasive Hgb was measured 3 times in each patient using the Masimo Pronto-7 (version
J Am Coll Surg
2.1.9).9 The first read was obtained on presentation to the trauma unit; second and third reads were obtained 5 minutes after the first and second measurements, respectively. Three reading were recorded to assess the reliability of noninvasive Hgb device to measure Hgb in trauma patients. Intraclass correlation test was performed to confirm the correlation between the 3 readings. Mean of the 3 readings was calculated and used for the correlation analysis. Outcomes measures Our primary outcomes measure was correlation and agreement between invasive and noninvasive spot check Hgb measurement. Secondary outcomes measures were correlation among 3 spot check Hgb measurements and predictive value (sensitivity, specificity, positive predictive value, negative predictive value, and accuracy) of the noninvasive Hgb measurement. Predictive value was calculated by grouping data based on the admission Hgb level as Hgb >8 mg/dL and Hgb 8 mg/dL. For our study, we defined normal Hgb value as Hgb >8 mg/dL. Statistical analysis Data are reported as the mean SD for continuous descriptive variables, the median (range) for ordinal descriptive variables, and the proportion for categorical variables. The intraclass correlation coefficient was used to assess correlation among the 3 noninvasive Hgb measurements. Spearman correlation test was performed to assess correlation among the invasive and the mean of 3 noninvasive spot check Hgb measurements. The agreement between invasive and noninvasive spot check Hgb was assessed using the Bland-Altman analysis. This method allows for quantification of the difference between the 2 techniques under consideration (invasive and noninvasive), and allows us to assess if the device (Masimo Pronto-7) has sufficient accuracy to allow acceptance of the device values in lieu of standard laboratory techniques.10,11 These data are summarized by the mean, the bias, and the 95% CI of the differences between techniques (limits of agreement). Bias was defined as the difference between the proposed method and the standard technique, and therefore estimated as the mean difference between Hgb readings from the Masimo and the laboratory blood draw (standard). The limits of agreement were calculated as (mean difference SD). Data were inserted into an Microsoft Excel 2007 work sheet and all statistical analysis was performed using Statistical Package for Social Sciences software (SPSS Inc).
RESULTS A total of 525 patients had attempted noninvasive spot check Hgb measurements with a success rate of 86%
Vol. 220, No. 1, January 2015
Table 1.
Joseph et al
Demographics of Study Population (n ¼ 525)
Variables
Age, y, mean SD Males, % (n) SBP, mmHg, mean SD HR, per min, mean SD GCS, mean SD ISS, mean SD AIS, median (range) Head, mean SD Chest, mean SD Abdomen, mean SD Extremity, mean SD
41 21 74 (333) 110 30 100 23 11 5 20 13 3 (1e5) 3.8 0.6 2.3 1 3.4 1.4 2.1 1.1
AIS, Abbreviate Injury Score; GCS, Glasgow Coma Score; HR, heart rate; ISS, Injury Severity Score; SBP, systolic blood pressure.
(n ¼ 450). A total of 450 invasive and 1,350 spot check Hgb measurements were obtained. Mean age was 41 21 years, 74% (n ¼ 333) were male, mean systolic blood pressure was 110 30 mmHg, and median Injury Severity Score was 17 (range 10 to 20). The main mechanism of injury was motor vehicle accidents (68%), followed by gunshot wounds (14%), and assault (8%). Table 1 highlights the demographics of the study population. The mean invasive Hgb was 11.5 4.36 g/dL (range 6 to 16 g/dL), mean noninvasive spot check Hgb was 11.1 3.60 g/dL (range 6.4 to 16.3 g/dL), and the difference in mean between invasive and noninvasive Hgb was 0.3 1.3 g/dL. There was no significant difference in the mean between invasive and noninvasive Hgb measurements (p ¼ 0.23). Thirty-eight percent (n ¼ 173) of patients had Hgb 8 g/dL on presentation based on invasive Hgb measurement. Twelve percent (n ¼ 54) received blood transfusion and 8% (n ¼ 36) underwent emergent surgical intervention for bleeding. After dichotomization of our patients into 2 groups, patient with Hgb 8 g/ dL and patients with Hgb >8 g/dL spot check Hgb measurements were found to have a sensitivity of 95.4% and an accuracy of 76%. Tables 2 and 3 demonstrate the dichotomization and predictive performance of the noninvasive spot check Hgb monitor. Noninvasive spot check Hgb measurement has a strong correlation with the invasive Hgb measurements (intraclass correlation ¼ 0.70; 95% CI, 0.57e0.80) and excellent correlation among the 3 consecutive noninvasive spot check Hgb measurements (intraclass correlation ¼ 0.90; 95% CI, 0.87e0.94). Spearman correlation plot revealed a strong correlation between invasive and noninvasive Hgb measurement (R2 ¼ 0.77; R ¼ 0.86; p ¼ 0.04). Figure 1 demonstrates the Spearman correlation plot between invasive and noninvasive Hgb measurement.
Noninvasive Hemoglobin Measurement in Trauma
95
Table 2. Dichotomy of Hemoglobin 8 mg/dL and >8 mg/dL
Variable
Spot check noninvasive Hgb measurement
8 mg/dL (n ¼ 265) >8 mg/dL (n ¼ 185)
Invasive Hgb measurement 8 mg/dL >8 mg/dL (n ¼ 173) (n ¼ 277)
165* 8
100 177
*Numbers in the table demonstrate number of patients (total n ¼ 450).
On subanalysis after stratifying patients into 2 groups (Hgb 8 g/dL and Hgb >8 g/dL), we found strong correlation between invasive and noninvasive Hgb measurement in patients with Hgb >8 g/dL (R2 ¼ 0.79; R ¼ 0.84; p ¼ 0.03). In patients with Hgb 8 g/dL, we found a good correlation between invasive and noninvasive Hgb measurement in patients with (R2 ¼ 0.71; R ¼ 0.80; p ¼ 0.04). Also, on assessing correlation in patients with admission Hgb between 10 and 13 mg/dL, we found good correlation (R2 ¼ 0.76; R ¼ 0.79; p ¼ 0.04) between invasive and noninvasive Hgb measurements. In our study, 20.9% (n ¼ 94) of patients were hypotensive (systolic blood pressure 90 mmHg) on admission. On assessing correlation in these patients, we found that there was good correlation between noninvasive spot check Hgb and invasive Hgb measurements (R2 ¼ 0.71; R ¼ 0.78; p ¼ 0.042). Figure 2 demonstrates the Bland-Altman analysis for invasive and noninvasive spot check Hgb measurements. The absolute mean difference (bias) was 0.096 g/dL, with an upper agreement limit at 4.016 g/dL and a lower agreement limit at 3.84 g/dL. Of all the measurements, 98.7% fell within 2 SDs of the mean difference. Mean hospital lengths of stay were 5 3 days, and mean ICU length of stay was 9 9 days. Overall mortality rate was 3% (16 of 525). The spot check Hgb probe failed to record the Hgb in 14% (n ¼ 75) of patients. Of the 75 patients in which Hgb could not be measured, 36 patients had nail polish or soot or tar on their fingers; 21 patients had difficulty in the sensor fitting, as we had only one size sensor for Table 3. Oximeter
Predictive Performance of Noninvasive Pulse
Variable
Sensitivity Specificity PPV NPV Accuracy NPV, negative predictive value; PPV, positive predictive value.
Value, %
95.4 63.8 62.2 95.6 76
96
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
Figure 1. Spearman correlation plot demonstrating strong correlation between invasive and continuous noninvasive Hgb monitoring.
our study; 10 patient had radiologic interference because of external factors, such as x-ray and ultrasound devices; and the remaining 8 patients’ Hgb could not be recorded due to anxious or agitated patients. On subanalysis of the 14% (n ¼ 75) of patients in whom the device could record the reading, the mean invasive Hgb was 9.1 2.30 g/dL, mean systolic blood pressure was 95.8 20.5 mmHg, and mean Injury Severity Score was 23 (range 21e28); 37.3% (n ¼ 28) were hypotensive and 11% (n ¼ 8) required operative intervention.
J Am Coll Surg
DISCUSSION The efficacy and accuracy of the Pronto-7 spot check pulse CO-oximeter (Masimo Rainbow SET; Masimo) in a trauma population has never been described. This is the first study to our knowledge to assess the use of a portable spot check pulse CO-oximeter in trauma patients. In our study, we found that the mean difference in Hgb values between spot check Hgb and invasive Hgb was 0.3 1.3 g/dL. We found excellent correlation (R2 ¼ 0.77) between invasive and spot check noninvasive Hgb measurements. We conclude that this novel technology allows for immediate and accurate Hgb measurements in trauma patients. Blood loss in trauma patients is routinely assessed based on Hgb and hematocrit measurement. Admission Hgb measurement has been shown to correlate with outcomes in trauma patients.12-14 Thorson et al14 showed that initial hematocrit and the change in hematocrit level have valuable information in transfusion requirements and are predictors of adverse outcomes in trauma patients. In our study, we assessed the reliability of noninvasive Hgb measurement compared with the invasive method, which is currently the standard of care in most trauma centers. We believe that implementing this new technology is going to benefit the patients by eliminating an unnecessary painful puncturing of the veins and unwanted complications. Also, the rapid measurement of Hgb level and ability to quickly repeat the measurement enables the trauma specialist to have an eye on patient’s hemodynamic stability. Point of care testing has recently expanded in many fields of laboratory medicine.8 In our study, we found
Figure 2. Bland-Altman agreement plot demonstrating strong agreement with 98.7% of reading within 2 SDs of agreement limits.
Vol. 220, No. 1, January 2015
Joseph et al
that spot check Hgb rapidly measured Hg, which accurately correlated to invasive Hgb. Laboratory measures for Hgb, although accurate, have limitations. Laboratory measures require blood samples, transport of the sample to a laboratory, and analysis, all of which are timeconsuming steps and can delay patient care; spot check Hgb monitoring can allow for an efficient and more rapid means for detection of hemorrhage, requirements of blood transfusion, and appropriate surgical management. We believe this device has the potential to improve clinical care, patient safety, and cost of care. In our study, we found excellent correlation between spot check Hgb and invasive Hgb. In a case series published in 2007, Macknet and colleagues15 compared noninvasive CO-Oximeter (Masimo Inc.) with invasive Hgb testing and concluded good correlation between the Hgb values obtained during times of rapidly changing Hgb concentrations due to surgical blood loss and transfusions.15 Similarly, in a study by Miller and colleagues,16 pulse COoximeter based spot check Hgb results were accurate to 1 g/dL (1 SD) when compared with laboratory measurements in healthy subjects undergoing hemodilution.16 Causey and colleagues17 demonstrated significant correlation between noninvasive Hgb monitoring with laboratory values in ICU patients and in general surgery patients undergoing elective operations.17 In another study on patients who underwent complex spine surgery, Berkow and colleagues8 demonstrated clinically acceptable accuracy of Hgb measurement within 1.5 g/dL between continuous noninvasive Hgb measurement via pulse CO-oximetry and standard laboratory reference device.8 We believe that, the new spot check Hgb monitor can play an important for quick assessment of Hgb in trauma patients. Despite excellent correlation with invasive Hgb measurement, one of the findings of concern was that spot check Hgb did not record data points 14% of the time. The inconsistency of spot check Hgb might be due to several factors, namely, low perfusion states, motion artifact, and disconnection of the sensor. Accordingly, the inability of the spot check Hgb monitor to recognize the trauma patient with life-threatening bleeding is the failure of this device, which needs to be further evaluated before implementation of this device in the clinical decision-making process. The results of this study should be interpreted in the context of its limitations. Through all invasive and spot check Hgb measurements were drawn simultaneously; however, there was no set protocol for the use of spot check Hgb probe. Second, we did not assess the effect of prehospital IV fluids administered during resuscitation on the accuracy of the device. Third, we did not perform a cost analysis. Despite these limitations, we demonstrate excellent correlation and
Noninvasive Hemoglobin Measurement in Trauma
97
agreement between noninvasive Hgb measurement technology and the invasive Hgb measurements.
CONCLUSIONS Spot check Hgb monitoring has excellent correlation with invasive Hgb measurements. Application of spot check has more clinical use as compared with previous continuous Hgb monitoring. This novel technology allows immediate and accurate Hgb measurements in trauma patients. Author Contributions Study conception and design: Joseph, O’Keeffe, Snyder, Rhee Acquisition of data: Joseph, Pandit, Aziz, Zangbar, Jehangir Analysis and interpretation of data: Joseph, Pandit, Kulvatunyou, Zangbar, Tang Drafting of manuscript: Joseph, Pandit, Aziz, Kulvatunyou, Tang Critical revision: Zangbar, O’Keeffe, Snyder, Rhee REFERENCES 1. Kauvar DS, Lefering R, Wade CE. Impact of hemorrhage on trauma outcome: an overview of epidemiology, clinical presentations, and therapeutic considerations. J Trauma 2006;60[Suppl]:S3eS11. 2. Acosta JA, Yang JC, Winchell RJ, et al. Lethal injuries and time to death in a level I trauma center. J Am Coll Surg 1998;186:528e533. 3. Joseph B, Hadjizacharia P, Aziz H, et al. Continuous noninvasive hemoglobin monitor from pulse ox: ready for prime time? World J Surg 2013;37:525e529. 4. Shapiro NI, Kociszewski C, Harrison T, et al. Isolated prehospital hypotension after traumatic injuries: a predictor of mortality? J Emerg Med 2003;25:175e179. 5. Gehring H, Hornberger C, Dibbelt L, et al. Accuracy of point of care testing (POCT) for determining hemoglobin concentrations. Acta Anaesthesiol Scand 2002;46:980e986. 6. Mokken FC, van der Waart FJ, Henny CP, et al. Differences in peripheral arterial and venous hemorheologic parameters. Ann Hematol 1996;73:135e137. 7. Yang ZW, Yang SH, Chen L, et al. Comparison of blood counts in venous, fingertip, and arterial blood and their measurement variation. Clin Lab Haematol 2001;23:155e159. 8. Berkow L, Rotolo S, Mirski E. Continuous noninvasive hemoglobin monitoring during complex spine surgery. Anesth Analg 2011;113:1396e1402. 9. Masimo. Pronto 7. Available at: http://www.masimo.com/pdf/ pronto-7/LAB6400G_Brochure_Pronto-7.pdf. Accessed August 21, 2014. 10. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8: 135e160. 11. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346[8982]:1085e1087.
98
Joseph et al
Noninvasive Hemoglobin Measurement in Trauma
12. Thorson CM, Van Haren RM, Ryan ML. Admission hematocrit and transfusion requirements after trauma. J Am Coll Surg 2013;216:65e73. 13. Ryan ML, Thorson CM, Otero CA. Initial hematocrit in trauma: a paradigm shift? J Trauma Acute Care Surg 2012; 72:54e59; discussion 59e60. 14. Thorson CM, Ryan ML, Van Haren RM, et al. Change in hematocrit during trauma assessment predicts bleeding even with ongoing fluid resuscitation. Am Surg 2013;79:398e406.
J Am Coll Surg
15. Macknet M, Norton S, Kimball-Jones P, et al. Continuous noninvasive measurement of hemoglobin via pulse CO-oximetry. Anesth Analg 2007;105:S-108. 16. Miller RD, Ward TA, Shiboski SC, et al. Comparison of three methods of hemoglobin monitoring in patients undergoing spine surgery. Anesth Analg 2011;112:858e863. 17. Causey MW, Miller S, Foster A, et al. Validation of noninvasive hemoglobin measurements using the Masimo Radical-7 SpHb Station. Am J Surg 2011;201:590e596.
