American Journal of Emergency Medicine 32 (2014) 356–362
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Brief Report
Cerebral oxygen saturation monitoring in pediatric altered mental status patients☆ Ian Kane, MD a,⁎, Thomas Abramo, MD b, Mark Meredith, MD a, Abby Williams, MD a, Kristen Crossman, MD a, Wei Wang, PhD c, Rameela Chandrasekhar, PhD c a b c
Pediatric Emergency Medicine, Vanderbilt Children's Hospital, Nashville, TN, USA Arkansas Children's Hospital, Little Rock, AR 72202, USA Department of Biostatistics, Vanderbilt University, Nashville, TN, USA
a r t i c l e
i n f o
Article history: Received 10 October 2013 Accepted 26 October 2013
a b s t r a c t Objectives: A pilot study assessing the potential utility of cerebral oximetry (local cerebral oxygen saturation [rcSO2]) in children presenting to the emergency department (ED) with altered mental status (AMS) and no history of trauma. Methods: Patients who presented to a tertiary pediatric ED with AMS were monitored with left and right cerebral near-infrared spectroscopy probes and the first 30 minutes of rcSO2 data was analyzed. Patients with a history of trauma were excluded. Patients with an abnormal head computed tomography (CT) (n = 146) were compared with those with a negative head CT (n = 45). Results: Mean rcSO2 values were consistent during each time period studied (5, 10, 20, and 30 minutes). In this study population, rcSO2 less than 50% or greater than 80% and increased absolute difference between the left and right rcSO2 measurements were associated with an abnormal CT scan. A difference of 12.2% between the left and right rcSO2 values had a 100% positive predictive value for an abnormal head CT among our patients. Cumulative graphical plots of rcSO2 trends showed that values b50% were associated with subdural hematomas (SDH) and values >80% were associated with epidural hematomas (EDH). Conclusions: This study demonstrated that cerebral oximetry can noninvasively detect altered cerebral physiology among a selected patient population. The difference between the left and right rcSO2 readings most reliably identified those subjects with altered cerebral physiology. In the future, rcSO2 monitoring has the potential to be used as a screening tool to identify, localize, and characterize intracranial injuries among children with AMS without a history of trauma. Published by Elsevier Inc.
1. Introduction Emergency department (ED) evaluation of children presenting with altered mental status (AMS) is challenging due to subtle signs of external trauma, unclear histories, and a lack of focal neurologic signs [1]. Rapid assessment of children with AMS is vital because delays in recognition of traumatic brain injuries (including subdural and epidural hematoma) are associated with significant morbidity and mortality. Studies have demonstrated with aggressive neurovascular resuscitation therapies such as hypertonic saline and mannitol to lower intracranial pressure. In the ED setting, computed tomography (CT) of the brain is the mainstay of diagnosis in these patients, but new evidence regarding the long-term risks of radiation has forced clinicians to become more judicious with their use of this imaging modality [2]. Emerging studies using near-infrared spectroscopy
☆ No grant support to report. ⁎ Corresponding author. Arkansas Children's Hospital, #1 Children's Way, Slot 51216, Little Rock, AR 72202. E-mail addresses: [email protected] (I. Kane), [email protected] (T. Abramo). 0735-6757/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.ajem.2013.10.047
(NIRS) suggest that it may be a useful tool for the evaluation of these children, and we report our experience with a pilot study of its use among a cohort of children in the pediatric ED. Cerebral oximetry, which relies on the differential absorption spectrum of oxygenated and deoxygenated hemoglobin, has emerged as a noninvasive method of monitoring changes in local cerebral oxygen saturation (rcSO2), oxygen delivery, oxygen extraction, and metabolism [3-6]. The rcSO2 value, expressed as a percentage, is an indirect measure of the relative oxygen supply and demand within the small area of tissue beneath the probe. rcSO2 readings, like pulse oximetry and end-tidal capnography, do not portray an absolute value but rather represent the aggregate of the local (direct) and global (indirect) physiological factors affecting that particular monitoring area. Low rcSO2 readings are thought to reflect increased oxygen extraction, increased metabolism, and/or decreased perfusion, whereas increased rcSO2 readings indicate decreased oxygen extraction and/or increased perfusion. Studies of cerebral oximetry during CPR have established that an rcSO2 value of 15% equates to the 25% of oxygen that is irreversibly bound to hemoglobin [7-12]. These rcSO2 readings like the pulse oximetry and capnography sampling can never portray an absolute sampling site value but represents more the
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362 Table 1 Demographics and rcSO2 of patients and controls
Age, y (SD) GCS SDH EDH SDH and EDH Unilateral rcSO2 b50% Bilateral rcSO2 b50% Unilateral rcSO2 N80% Bilateral rcSO2 N80% Left mean rcSO2 (SD) Right mean rcSO2 (SD) Difference between left and right mean rcSO2 (SD)
Abnormal head CT (n = 146)
Negative head CT P (n = 45)
4.1 (3.7) 9.4 (2.7) 28.0% 32.9% 39.7% 45.2% 24.0% 41.8% 17.8% 57.3 (28.7) 58.1 (28.3) 27.2 (21.8)
3.2 (3.3) 11.3 (1.8) – – – 0% 0% 4.4% 13.3% 70.5 (9.4) 69.6 (7.8) 4.2 (3)
.10 b.0001 – – – b.0001 b.0001 b.0001 .65 .11 .25 b.0001
Abbreviation: GCS, Glasgow Coma Scale.
aggregate of the local (direct) and global (indirect) physiological factors affecting that particular monitoring area [5-8]. A majority of healthcare provider still misinterpreted or perceive pulse oximetry or capnography readings as absolute sampling values not as a composite value of local and global physiological parameters which has been an issue with cerebral oximetry’s acceptance. A recent study among children with congenital heart disease suggests that rcSO2 measurements correlate with jugular venous saturations across a wide spectrum of oxygenation values [7]. Most pediatric studies of NIRS technology have addressed its utility as a marker of adequate cerebral perfusion among children undergoing cardiac surgery for congenital heart disease [3,8,9]. Numerous Cerebral Oximetry NIRS studies suggest it may be a useful noninvasive tool for cerebral physiology and pathology for adults and children [7-14]. Cerebral Oximetry NIRS technology studies have addressed its utility as a noninvasive neurological monitor in clinical setting outside the emergency department for cerebral physiology and pathology among adults and children in numerous clinical studies especially in neurological emergencies [5-7,12-16]. In military and civilian studies utilizing cerebral NIR technology in assessing trauma patients they have shown the ability for cerebral oximetry to identify cases of intracranial hemorrhage among adults and children [17-20]. In our pediatric ED, rcSO2 monitoring has been used as part of the evaluation of children with suspected neurologic emergencies. We present a pilot study of the utility of rcSO2 monitoring among children presenting with AMS without historical or physical examination
357
findings indicative of trauma. This pilot study’s primary objective is to investigate the potential for cerebral oximetry rcSO2 as a screening tool for altered children with no history of trauma and correlate it to their CT scan in the emergency department setting and not to correlate their rcSO2 readings to the degree, size, age or depth of their cerebral pathology. The ability to promptly recognize altered cerebral physiology among this vulnerable population could decrease the time to head CT and definitive treatment, thereby improving overall neurologic outcomes. 2. Methods We conducted a PED retrospective chart analysis from January 2008 to October 2012, of patients who presented with chief complaints, signs or symptoms for 1) altered mental status, 2) no history of trauma, 3) had cerebral oximetry rcSO2 monitoring and 4) had head CT scan as a part of their initial workup. Patients who were seizing, those with a history of trauma or obvious traumatic injuries, and those undergoing active cardiopulmonary resuscitation or with a history of cardiopulmonary resuscitation were excluded. Children who did not have head CT imaging performed as part of their AMS workup were also excluded. Children were also excluded if they had neurosurgical interventions performed as a part of their initial resuscitation. A total of 191 patients fulfilled inclusion criteria, and rcSO2 data was collected using an INVOS 5100C Cerebral Oximeter (Somanetics, Troy, MI) using either adult (N40 kg) or pediatric (b 40 kg) probes placed on the patient's left and right forehead. The rcSO2 data were collected from the left and right cerebrum every 30 seconds for a total of 30 minutes (60 total measurements) for each individual in the abnormal head CT (n = 146) and negative head CT (n = 45) group. This time interval was chosen so as to minimize the effects of any ED interventions on the cerebral physiology. The electronic medical record was reviewed for each patient, and demographic, clinical, and radiographic data were collected. Mean rcSO2 values among patients with a negative head CT were compared with those with an abnormal head CT at 5, 15, 20, and 30 minutes. The absolute difference between the left and right mean rcSO2 was the summary statistic with the greatest area under the curve as the most useful characteristic to identify patients with an abnormal head CT. Performance characteristics including sensitivity, specificity, positive predictive value, and negative predictive value were computed for the mean absolute difference between left and right rcSO2. Subgroup analysis was performed to determine if rcSO2 predicted injury type (subdural or epidural hematoma) and/or location (left or right cerebrum). Cumulative graphical spaghetti plots demonstrating the
Table 2 Performance characteristics using left and right mean rcSO2 difference to predict an abnormal head CT Left and right mean rcSO2 difference (%) 3.8 5.1 8.0 10.0 12.2
Sensitivity (95% CI)
Specificity (95% CI)
PPV (95% CI)
NPV (95% CI)
0.83 0.77 0.71 0.69 0.67
0.62 (0.48-0.75) 0.71 (0.57-0.82) 0.84 (0.71-0.92) 0.96 (0.85-0.99) 1 (0.92-1.0)
0.88 0.90 0.94 0.98 1
0.53 0.49 0.48 0.49 0.48
(0.76-0.88) (0.70-0.83) (0.63-0.78) (0.61-0.76) (0.59-0.74)
(0.81-0.92) (0.83-0.94) (0.88-0.97) (0.93-1.0) (0.96-1.0)
(0.40-0.66) (0.38-0.61) (0.37-0.58) (0.39-0.59) (0.39-0.58)
Abbreviations: CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value. Table 3 Local cerebral oxygen saturation among patients with an abnormal head CT (n = 146)
Left mean rcSO2 (SD) Right mean rcSO2 (SD)
Only left CT lesion (n = 42)
Only right CT lesion (n = 33)
Bilateral CT lesion (n = 72)
47.6 (28.4)
77.4 (13.6)
49.3 (30)
73.8 (15.2)
50.9 (26.9)
51 (32.1)
Bilateral SDH (n = 27)
Bilateral EDH (n = 9)
24 (6.9)
89.2 (4.2)
22.5 (9.5)
89.9 (3.5)
358
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
Table 4 Local cerebral oxygen saturation among patients with an abnormal head CT and bilateral rcSO2 values less than 50% or greater than 80%
L mean rcSO2 (SD) R mean rcSO2 (SD)
SDH or SDH/EDH (n = 57)
EDH only (n = 9)
P value
51.6 (32.3) 49 (32.9)
86.3 (9) 88 (6.2)
.014 .006
overall rcSO2 trend for each injury type and location were constructed for patients with an abnormal head CT.
3. Results The study population’s mean rcSO2 values were similar across each time period studied (5, 10, 15, 20, and 30 minutes) and the 5 minute values were chosen for further analysis due to this timeframe's earliest relevance and clinical significance. Demographic data from patients and controls are included in Table 1. Children with abnormal head CT imaging had significantly lower Glasgow Coma Scale score on presentation and were more likely to have rcSO2 values less than 50% and greater than 80%. The difference between the left and right rcSO2 values was significantly greater among children with an abnormal head CT compared with those with a negative CT scan. Among all patients, there was a trend toward lower rcSO2 values in the abnormal head CT group. Most children with a negative head CT had rcSO2 values between 60% and 80%, with a mean of 70%. Selected performance characteristics of the left and right mean rcSO2 difference are shown in Table 2. Sensitivity decreased and specificity increased the as mean difference in rcSO2 increased. A mean difference of 3.8 had a sensitivity of 0.83 for detecting an abnormal head CT, whereas a mean difference of 12.2 had a specificity and positive predictive value of 1.0. Analysis among patients with an abnormal head CT demonstrated lower rcSO2 readings which corresponded to the side of the CT lesion. Bilateral SDH patients had lower rcSO2 readings and patients with EDH had higher rcSO2 readings (Table 3). A subgroup
analysis was performed among patients with an abnormal head CT who also had left and right mean rcSO2 values which were outside the range of normal (predetermined based on literature review to be b50% or >80%). Within this group rcSO2 was significantly higher among patients with EDH compared to those with SDH or a mixed lesion (Table 4). Composite spaghetti plots showing each patient's individual rcSO2 tracing were constructed to provide a graphical view of the left and right rcSO2 values over the first 30 minutes of measurement. Among patients with a negative head CT, left and right rcSO2 values clustered around 70% (Fig. 1). Patients with isolated unilateral subdural hematoma (SDH) had decreased rcSO2 ipsilateral to the side of the lesion (Figs. 2 and 3), whereas those with bilateral SDH had decreased rcSO2 values in each hemisphere (Fig. 4). Patients with bilateral epidural hematoma (EDH) had elevated rcSO2 values, particularly within the first 10 minutes of measurement (Fig. 5). Among patients with mixed intracranial lesions on head CT, the hemisphere ipsilateral to the SDH was associated with a lower rcSO2 value, whereas the side ipsilateral to the EDH was associated with a higher rcSO2 value (Figs. 6 and 7).
4. Discussion This pilot study demonstrates that real time rcSO2 monitoring is clinically relevant in the evaluation of children who present with undifferentiated altered mental status in the emergency department. Although there was a degree of minute-to-minute variation in the rcSO2, in general, the rcSO2 trend for each individual patient remained stable during our 30-minute monitoring period. Our results show that the mean rcSO2 among children with an abnormal head CT was more likely to be less than 50% or greater than 80%. When the rcSO2 graphs from patients with bilateral or unilateral SDH on head CT were plotted together, a possible relationship between decreased rcSO2 and SDH was observed. This is of particular interest because we theorize that decreased rcSO2 ipsilateral to a SDH may be due to venous blood pooling and maximal oxygen extraction, leading to local tissue hypoxia and lower rcSO2. Although we had very few
Fig. 1. Local cerebral oxygen saturation among patients with a negative head CT (n = 45). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
359
Fig. 2. Decreased rcSO2 in the left cortex among patients with a left SDH (n = 26). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the grayshaded area, the 95% confidence interval of the means.
patients with isolated EDH, these patients tended to have elevated rcSO2 ipsilateral to the cortex with the EDH, which may be due to pooled arterial blood with ineffective oxygen extraction. This study also demonstrated that the absolute difference between the left and right mean rcSO2 has potential as a non-invasive monitor for identifying patients with abnormal cerebral physiology and head CT. Children with one or both rcSO2 values b50% were more likely to have an abnormality on head CT, as were those patients with one value >80%. This study primarily demonstrated with high significance
that the difference between the left and right mean rcSO2 has high potential as a noninvasive monitor for identifying patients with an abnormal cerebral physiology, pathology and head CT. Although this study provides promising initial data, several factors limit our study including its retrospective nature and the fact that it did not control for other clinical variables which could have impacted the rcSO2 readings. These factors such as the overall hemodynamic state the patient, the location of the intracranial injury relative to the NIRS probes, duration of the cerebral injury,
Fig. 3. Decreased rcSO2 in the right cortex among patients with a right SDH (n = 14). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
360
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
Fig. 4. Decreased rcSO2 in the left and right cortex among patients with bilateral SDH (n = 27). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
and the time elapsed between the cerebral injury, its secondary effect on the opposite cerebral side and the timeframe from injury to presentation to the emergency department can have significant effects on the rcSO2 readings as demonstrated in the study’s population rcSO2 readings especially in patients with various SDH and EDH findings. Despite the apparent differences in rcSO2 values between SDH and EDH, it remains unclear whether rcSO2 values remain reliable in acute and chronic SDH and EDH. Because the true injury time was unknown among our patients, our study likely
included a mix of chronic and acute intracranial lesions, which may bias our results. Our study also used NIRS probes on the left and right forehead only, limiting the area captured for analysis. Injuries to the occipital and parietal lobes of the brain would not have been detected directly by these probes, and any secondary hemodynamic impact of these injuries on the frontal rcSO2 values is not known. The study’s primary and secondary focus was not to investigate the utilization of cerebral oximetry diagnostically, prompting or the effects of interventions but rather potentially as an adjunct decision tool to assist
Fig. 5. Increased rcSO2 in the right and left cortex among patients with bilateral EDH (n = 9). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means. The three patients who's rcS02 dropped in the above graph had sudden deterioration of their neurological and cardiovascular system requiring interventions, anticonvulsants, 3% hypertonic saline 5ml/kg and intubation during their rcS02 monitoring.
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
361
Fig. 6. Increased rcSO2 in the left cortex and decreased rcSO2 in the right cortex among patients with left EDH and right SDH (n = 11). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
in earlier detection of abnormal cerebral physiology and pathology and should not be misinterpreted as negating the patient’s need for the CT scan. Because of these limitations, at this time, rcSO2 monitoring should not be used diagnostically but rather as an adjunct decision tool to assist in earlier detection of brain injury. Numerous clinical studies have demonstrated the necessity of early identification of brain injury in an effort to prevent secondary brain injury and improve neurologic outcomes. The utilization of rcSO2 monitoring to detect occult abnormal cerebral physiology and pathol-
ogy in AMS patients is an attractive noninvasive tool for the bedside clinician. Cerebral oximetry technology offers many benefits to the acute care physician in that it is portable, painless, and can produce results within minutes. This real-time information could prompt immediate interventions that improve patient outcomes. By using the mean difference between the rcSO2 from the left and right cerebrum, the study was able to identify those subjects with altered cerebral physiology and an abnormal head CT. We also demonstrated an association between decreased rcSO2 and SDH and increased rcSO2 and
Fig. 7. Decreased rcSO2 in the left cortex and increased rcSO2 in the right cortex among patients with left SDH and right EDH (n = 10). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
362
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
EDH. Our results highlight the need for further randomized studies of rcSO2 monitoring among children who present to the ED with suspicion of intracranial injury. Prospective studies are also needed to establish normative rcSO2 values among healthy pediatric patients in the ED. In summary, Cerebral Oximetry rcSO2 monitoring can immediately detect altered cerebral physiology and pathology among a selected population. Cerebral Oximetry has demonstrated its potentials as a screening tool to identify, localize, and characterize intracranial injuries among children with AMS without trauma. In this study population the difference between either Left or Right cerebral rcSO2 readings most reliably identified those subjects with altered cerebral physiology and intracranial pathology and abnormal head CT. Either side or Both cerebral rcSO2 readings that were less than 50% or unilateral readings greater than 80% were significant for Subdural(b 50% cerebral rcSO2 mean) or Epidural (N 80% cerebral rcSO2 mean). Cerebral Oximetry in the emergency department has the potential to screen for occult intracranial injuries among this vulnerable population but further studies are warranted. References [1] Salonia R, Bell MJ, Kochanek PM, et al. The utility of near infrared spectroscopy in detecting intracranial hemorrhage in children. J Neurotrauma 2012;29:1047–53. [2] Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet 2009;374:160–1170. [3] Drayna PC, Abramo TJ, Estrada CE. Near-infrared spectroscopy in the critical setting. Pediatr Emerg Care 2011;27:432–42. [4] Ghanayem NS, Wernovsky G, Hoffman GM. Near infrared spectroscopy as a hemodynamic monitor in critical illness. Pediatr Crit Care Med 2011;12: S27–32. [5] Murkin JM, Arango A. Near-infrared spectroscopy as an index of brain and tissue oxygenation. Br J Anaesth 2009;103:i3–i13. [6] Henson LC, Calalang C, Temp JA, et al. Accuracy of a cerebral oximeter in healthy volunteers under conditions of isocapnic hypoxia. Anesthesiology 1998;88:58–65.
[7] Kreeger RN, Ramamoorthy C, Nicolson SC, et al. Evaluation of pediatric nearinfrared cerebral oximeter for cardiac disease. Ann Thorac Surg 2012;94:1527–33. [8] Uebling A, Furck AK, Hansen JH, et al. Perioperative cerebral and somatic oxygenation in neonates with hypoplastic left heart syndrome or transposition of the great arteries. J Thorac Cardiovasc Surg 2011;142:523–30. [9] Hansen JH, Schlangen J, Armbrust S, et al. Monitoring of regional tissue oxygenation with near-infrared spectroscopy during the early postoperative course after superior cavopulmonary anastomosis. Eur J Cardiothorac Surg 2013;43:e37–43. [10] Alderliesten T, Lemmers PM, Smarius JJ, et al. Cerebral oxygenation, extraction, and autoregulation in very preterm infants who develop peri-intraventricular hemorrhage. J Pediatr 2013;162:698–704. [11] Tina LG, Frigiola A, Abella R, et al. Near Infrared Spectroscopy in healthy preterm and term newborns: correlation with gestational age and standard monitoring parameters. Curr Neurovasc Res 2009;6:148–54. [12] Bernal NP, Hoffman GM, Ghanayem NS, et al. Cerebral and somatic near-infrared spectroscopy in normal newborns. J Pediatr Surg 2010;45:1306–10. [13] Dunham CM, Ransom KJ, Flowers LL, Siegal JD, Kohli CM. Cerebral hypoxia in severely brain-injured patients is associated with admission Glasgow coma scale score, computed tomographic severity, cerebral perfusion pressure, and survival. J Trauma 2004;56:482–91. [14] Kim MB, Ward DS, Cartwright CR, Kolano J, Chlebowski S, Henson LC, et al. Estimation of jugular venous O2 saturation from cerebral oximetry or arterial O2 saturation during isocapnic hypoxia. J Clin Monit Comput 2000; 16:191–9. [15] Myburgh JA. Quantifying cerebral autoregulation in health and disease. Crit Care Resusc 2004;6(1):59–67. [16] Obrig H. NIRS in clinical neurology - a 'promising' tool?Neuroimage 2013 pii: S1053-8119(13)00295-4. [17] Francis SJ, Ravindran G, Visvanathan K, et al. Screening for unilateral intracranial abnormalities using near infrared spectroscopy. A preliminary report. J Clin Neurosci 2005;12:291–5. [18] Gopinath S, Robertson C, Contant C, et al. Early detection of delayed traumatic intracranial hematomas using near infrared spectroscopy. J Neurosurg 1995;83: 438–44. [19] Kahraman S, Kayali H, Atabey C, et al. The accuracy of near infrared spectroscopy in detection of subdural and epidural hematomas. J Trauma 2006;61:1480–3. [20] Robertson C, Zager E, Narayan R, et al. Clinical evaluation of a portable nearinfrared device for detection of traumatic intracranial hematomas. J Neurotrauma 2010;27:1597–604.
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Brief Report
Cerebral oxygen saturation monitoring in pediatric altered mental status patients☆ Ian Kane, MD a,⁎, Thomas Abramo, MD b, Mark Meredith, MD a, Abby Williams, MD a, Kristen Crossman, MD a, Wei Wang, PhD c, Rameela Chandrasekhar, PhD c a b c
Pediatric Emergency Medicine, Vanderbilt Children's Hospital, Nashville, TN, USA Arkansas Children's Hospital, Little Rock, AR 72202, USA Department of Biostatistics, Vanderbilt University, Nashville, TN, USA
a r t i c l e
i n f o
Article history: Received 10 October 2013 Accepted 26 October 2013
a b s t r a c t Objectives: A pilot study assessing the potential utility of cerebral oximetry (local cerebral oxygen saturation [rcSO2]) in children presenting to the emergency department (ED) with altered mental status (AMS) and no history of trauma. Methods: Patients who presented to a tertiary pediatric ED with AMS were monitored with left and right cerebral near-infrared spectroscopy probes and the first 30 minutes of rcSO2 data was analyzed. Patients with a history of trauma were excluded. Patients with an abnormal head computed tomography (CT) (n = 146) were compared with those with a negative head CT (n = 45). Results: Mean rcSO2 values were consistent during each time period studied (5, 10, 20, and 30 minutes). In this study population, rcSO2 less than 50% or greater than 80% and increased absolute difference between the left and right rcSO2 measurements were associated with an abnormal CT scan. A difference of 12.2% between the left and right rcSO2 values had a 100% positive predictive value for an abnormal head CT among our patients. Cumulative graphical plots of rcSO2 trends showed that values b50% were associated with subdural hematomas (SDH) and values >80% were associated with epidural hematomas (EDH). Conclusions: This study demonstrated that cerebral oximetry can noninvasively detect altered cerebral physiology among a selected patient population. The difference between the left and right rcSO2 readings most reliably identified those subjects with altered cerebral physiology. In the future, rcSO2 monitoring has the potential to be used as a screening tool to identify, localize, and characterize intracranial injuries among children with AMS without a history of trauma. Published by Elsevier Inc.
1. Introduction Emergency department (ED) evaluation of children presenting with altered mental status (AMS) is challenging due to subtle signs of external trauma, unclear histories, and a lack of focal neurologic signs [1]. Rapid assessment of children with AMS is vital because delays in recognition of traumatic brain injuries (including subdural and epidural hematoma) are associated with significant morbidity and mortality. Studies have demonstrated with aggressive neurovascular resuscitation therapies such as hypertonic saline and mannitol to lower intracranial pressure. In the ED setting, computed tomography (CT) of the brain is the mainstay of diagnosis in these patients, but new evidence regarding the long-term risks of radiation has forced clinicians to become more judicious with their use of this imaging modality [2]. Emerging studies using near-infrared spectroscopy
☆ No grant support to report. ⁎ Corresponding author. Arkansas Children's Hospital, #1 Children's Way, Slot 51216, Little Rock, AR 72202. E-mail addresses: [email protected] (I. Kane), [email protected] (T. Abramo). 0735-6757/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.ajem.2013.10.047
(NIRS) suggest that it may be a useful tool for the evaluation of these children, and we report our experience with a pilot study of its use among a cohort of children in the pediatric ED. Cerebral oximetry, which relies on the differential absorption spectrum of oxygenated and deoxygenated hemoglobin, has emerged as a noninvasive method of monitoring changes in local cerebral oxygen saturation (rcSO2), oxygen delivery, oxygen extraction, and metabolism [3-6]. The rcSO2 value, expressed as a percentage, is an indirect measure of the relative oxygen supply and demand within the small area of tissue beneath the probe. rcSO2 readings, like pulse oximetry and end-tidal capnography, do not portray an absolute value but rather represent the aggregate of the local (direct) and global (indirect) physiological factors affecting that particular monitoring area. Low rcSO2 readings are thought to reflect increased oxygen extraction, increased metabolism, and/or decreased perfusion, whereas increased rcSO2 readings indicate decreased oxygen extraction and/or increased perfusion. Studies of cerebral oximetry during CPR have established that an rcSO2 value of 15% equates to the 25% of oxygen that is irreversibly bound to hemoglobin [7-12]. These rcSO2 readings like the pulse oximetry and capnography sampling can never portray an absolute sampling site value but represents more the
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362 Table 1 Demographics and rcSO2 of patients and controls
Age, y (SD) GCS SDH EDH SDH and EDH Unilateral rcSO2 b50% Bilateral rcSO2 b50% Unilateral rcSO2 N80% Bilateral rcSO2 N80% Left mean rcSO2 (SD) Right mean rcSO2 (SD) Difference between left and right mean rcSO2 (SD)
Abnormal head CT (n = 146)
Negative head CT P (n = 45)
4.1 (3.7) 9.4 (2.7) 28.0% 32.9% 39.7% 45.2% 24.0% 41.8% 17.8% 57.3 (28.7) 58.1 (28.3) 27.2 (21.8)
3.2 (3.3) 11.3 (1.8) – – – 0% 0% 4.4% 13.3% 70.5 (9.4) 69.6 (7.8) 4.2 (3)
.10 b.0001 – – – b.0001 b.0001 b.0001 .65 .11 .25 b.0001
Abbreviation: GCS, Glasgow Coma Scale.
aggregate of the local (direct) and global (indirect) physiological factors affecting that particular monitoring area [5-8]. A majority of healthcare provider still misinterpreted or perceive pulse oximetry or capnography readings as absolute sampling values not as a composite value of local and global physiological parameters which has been an issue with cerebral oximetry’s acceptance. A recent study among children with congenital heart disease suggests that rcSO2 measurements correlate with jugular venous saturations across a wide spectrum of oxygenation values [7]. Most pediatric studies of NIRS technology have addressed its utility as a marker of adequate cerebral perfusion among children undergoing cardiac surgery for congenital heart disease [3,8,9]. Numerous Cerebral Oximetry NIRS studies suggest it may be a useful noninvasive tool for cerebral physiology and pathology for adults and children [7-14]. Cerebral Oximetry NIRS technology studies have addressed its utility as a noninvasive neurological monitor in clinical setting outside the emergency department for cerebral physiology and pathology among adults and children in numerous clinical studies especially in neurological emergencies [5-7,12-16]. In military and civilian studies utilizing cerebral NIR technology in assessing trauma patients they have shown the ability for cerebral oximetry to identify cases of intracranial hemorrhage among adults and children [17-20]. In our pediatric ED, rcSO2 monitoring has been used as part of the evaluation of children with suspected neurologic emergencies. We present a pilot study of the utility of rcSO2 monitoring among children presenting with AMS without historical or physical examination
357
findings indicative of trauma. This pilot study’s primary objective is to investigate the potential for cerebral oximetry rcSO2 as a screening tool for altered children with no history of trauma and correlate it to their CT scan in the emergency department setting and not to correlate their rcSO2 readings to the degree, size, age or depth of their cerebral pathology. The ability to promptly recognize altered cerebral physiology among this vulnerable population could decrease the time to head CT and definitive treatment, thereby improving overall neurologic outcomes. 2. Methods We conducted a PED retrospective chart analysis from January 2008 to October 2012, of patients who presented with chief complaints, signs or symptoms for 1) altered mental status, 2) no history of trauma, 3) had cerebral oximetry rcSO2 monitoring and 4) had head CT scan as a part of their initial workup. Patients who were seizing, those with a history of trauma or obvious traumatic injuries, and those undergoing active cardiopulmonary resuscitation or with a history of cardiopulmonary resuscitation were excluded. Children who did not have head CT imaging performed as part of their AMS workup were also excluded. Children were also excluded if they had neurosurgical interventions performed as a part of their initial resuscitation. A total of 191 patients fulfilled inclusion criteria, and rcSO2 data was collected using an INVOS 5100C Cerebral Oximeter (Somanetics, Troy, MI) using either adult (N40 kg) or pediatric (b 40 kg) probes placed on the patient's left and right forehead. The rcSO2 data were collected from the left and right cerebrum every 30 seconds for a total of 30 minutes (60 total measurements) for each individual in the abnormal head CT (n = 146) and negative head CT (n = 45) group. This time interval was chosen so as to minimize the effects of any ED interventions on the cerebral physiology. The electronic medical record was reviewed for each patient, and demographic, clinical, and radiographic data were collected. Mean rcSO2 values among patients with a negative head CT were compared with those with an abnormal head CT at 5, 15, 20, and 30 minutes. The absolute difference between the left and right mean rcSO2 was the summary statistic with the greatest area under the curve as the most useful characteristic to identify patients with an abnormal head CT. Performance characteristics including sensitivity, specificity, positive predictive value, and negative predictive value were computed for the mean absolute difference between left and right rcSO2. Subgroup analysis was performed to determine if rcSO2 predicted injury type (subdural or epidural hematoma) and/or location (left or right cerebrum). Cumulative graphical spaghetti plots demonstrating the
Table 2 Performance characteristics using left and right mean rcSO2 difference to predict an abnormal head CT Left and right mean rcSO2 difference (%) 3.8 5.1 8.0 10.0 12.2
Sensitivity (95% CI)
Specificity (95% CI)
PPV (95% CI)
NPV (95% CI)
0.83 0.77 0.71 0.69 0.67
0.62 (0.48-0.75) 0.71 (0.57-0.82) 0.84 (0.71-0.92) 0.96 (0.85-0.99) 1 (0.92-1.0)
0.88 0.90 0.94 0.98 1
0.53 0.49 0.48 0.49 0.48
(0.76-0.88) (0.70-0.83) (0.63-0.78) (0.61-0.76) (0.59-0.74)
(0.81-0.92) (0.83-0.94) (0.88-0.97) (0.93-1.0) (0.96-1.0)
(0.40-0.66) (0.38-0.61) (0.37-0.58) (0.39-0.59) (0.39-0.58)
Abbreviations: CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value. Table 3 Local cerebral oxygen saturation among patients with an abnormal head CT (n = 146)
Left mean rcSO2 (SD) Right mean rcSO2 (SD)
Only left CT lesion (n = 42)
Only right CT lesion (n = 33)
Bilateral CT lesion (n = 72)
47.6 (28.4)
77.4 (13.6)
49.3 (30)
73.8 (15.2)
50.9 (26.9)
51 (32.1)
Bilateral SDH (n = 27)
Bilateral EDH (n = 9)
24 (6.9)
89.2 (4.2)
22.5 (9.5)
89.9 (3.5)
358
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
Table 4 Local cerebral oxygen saturation among patients with an abnormal head CT and bilateral rcSO2 values less than 50% or greater than 80%
L mean rcSO2 (SD) R mean rcSO2 (SD)
SDH or SDH/EDH (n = 57)
EDH only (n = 9)
P value
51.6 (32.3) 49 (32.9)
86.3 (9) 88 (6.2)
.014 .006
overall rcSO2 trend for each injury type and location were constructed for patients with an abnormal head CT.
3. Results The study population’s mean rcSO2 values were similar across each time period studied (5, 10, 15, 20, and 30 minutes) and the 5 minute values were chosen for further analysis due to this timeframe's earliest relevance and clinical significance. Demographic data from patients and controls are included in Table 1. Children with abnormal head CT imaging had significantly lower Glasgow Coma Scale score on presentation and were more likely to have rcSO2 values less than 50% and greater than 80%. The difference between the left and right rcSO2 values was significantly greater among children with an abnormal head CT compared with those with a negative CT scan. Among all patients, there was a trend toward lower rcSO2 values in the abnormal head CT group. Most children with a negative head CT had rcSO2 values between 60% and 80%, with a mean of 70%. Selected performance characteristics of the left and right mean rcSO2 difference are shown in Table 2. Sensitivity decreased and specificity increased the as mean difference in rcSO2 increased. A mean difference of 3.8 had a sensitivity of 0.83 for detecting an abnormal head CT, whereas a mean difference of 12.2 had a specificity and positive predictive value of 1.0. Analysis among patients with an abnormal head CT demonstrated lower rcSO2 readings which corresponded to the side of the CT lesion. Bilateral SDH patients had lower rcSO2 readings and patients with EDH had higher rcSO2 readings (Table 3). A subgroup
analysis was performed among patients with an abnormal head CT who also had left and right mean rcSO2 values which were outside the range of normal (predetermined based on literature review to be b50% or >80%). Within this group rcSO2 was significantly higher among patients with EDH compared to those with SDH or a mixed lesion (Table 4). Composite spaghetti plots showing each patient's individual rcSO2 tracing were constructed to provide a graphical view of the left and right rcSO2 values over the first 30 minutes of measurement. Among patients with a negative head CT, left and right rcSO2 values clustered around 70% (Fig. 1). Patients with isolated unilateral subdural hematoma (SDH) had decreased rcSO2 ipsilateral to the side of the lesion (Figs. 2 and 3), whereas those with bilateral SDH had decreased rcSO2 values in each hemisphere (Fig. 4). Patients with bilateral epidural hematoma (EDH) had elevated rcSO2 values, particularly within the first 10 minutes of measurement (Fig. 5). Among patients with mixed intracranial lesions on head CT, the hemisphere ipsilateral to the SDH was associated with a lower rcSO2 value, whereas the side ipsilateral to the EDH was associated with a higher rcSO2 value (Figs. 6 and 7).
4. Discussion This pilot study demonstrates that real time rcSO2 monitoring is clinically relevant in the evaluation of children who present with undifferentiated altered mental status in the emergency department. Although there was a degree of minute-to-minute variation in the rcSO2, in general, the rcSO2 trend for each individual patient remained stable during our 30-minute monitoring period. Our results show that the mean rcSO2 among children with an abnormal head CT was more likely to be less than 50% or greater than 80%. When the rcSO2 graphs from patients with bilateral or unilateral SDH on head CT were plotted together, a possible relationship between decreased rcSO2 and SDH was observed. This is of particular interest because we theorize that decreased rcSO2 ipsilateral to a SDH may be due to venous blood pooling and maximal oxygen extraction, leading to local tissue hypoxia and lower rcSO2. Although we had very few
Fig. 1. Local cerebral oxygen saturation among patients with a negative head CT (n = 45). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
359
Fig. 2. Decreased rcSO2 in the left cortex among patients with a left SDH (n = 26). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the grayshaded area, the 95% confidence interval of the means.
patients with isolated EDH, these patients tended to have elevated rcSO2 ipsilateral to the cortex with the EDH, which may be due to pooled arterial blood with ineffective oxygen extraction. This study also demonstrated that the absolute difference between the left and right mean rcSO2 has potential as a non-invasive monitor for identifying patients with abnormal cerebral physiology and head CT. Children with one or both rcSO2 values b50% were more likely to have an abnormality on head CT, as were those patients with one value >80%. This study primarily demonstrated with high significance
that the difference between the left and right mean rcSO2 has high potential as a noninvasive monitor for identifying patients with an abnormal cerebral physiology, pathology and head CT. Although this study provides promising initial data, several factors limit our study including its retrospective nature and the fact that it did not control for other clinical variables which could have impacted the rcSO2 readings. These factors such as the overall hemodynamic state the patient, the location of the intracranial injury relative to the NIRS probes, duration of the cerebral injury,
Fig. 3. Decreased rcSO2 in the right cortex among patients with a right SDH (n = 14). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
360
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
Fig. 4. Decreased rcSO2 in the left and right cortex among patients with bilateral SDH (n = 27). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
and the time elapsed between the cerebral injury, its secondary effect on the opposite cerebral side and the timeframe from injury to presentation to the emergency department can have significant effects on the rcSO2 readings as demonstrated in the study’s population rcSO2 readings especially in patients with various SDH and EDH findings. Despite the apparent differences in rcSO2 values between SDH and EDH, it remains unclear whether rcSO2 values remain reliable in acute and chronic SDH and EDH. Because the true injury time was unknown among our patients, our study likely
included a mix of chronic and acute intracranial lesions, which may bias our results. Our study also used NIRS probes on the left and right forehead only, limiting the area captured for analysis. Injuries to the occipital and parietal lobes of the brain would not have been detected directly by these probes, and any secondary hemodynamic impact of these injuries on the frontal rcSO2 values is not known. The study’s primary and secondary focus was not to investigate the utilization of cerebral oximetry diagnostically, prompting or the effects of interventions but rather potentially as an adjunct decision tool to assist
Fig. 5. Increased rcSO2 in the right and left cortex among patients with bilateral EDH (n = 9). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means. The three patients who's rcS02 dropped in the above graph had sudden deterioration of their neurological and cardiovascular system requiring interventions, anticonvulsants, 3% hypertonic saline 5ml/kg and intubation during their rcS02 monitoring.
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
361
Fig. 6. Increased rcSO2 in the left cortex and decreased rcSO2 in the right cortex among patients with left EDH and right SDH (n = 11). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
in earlier detection of abnormal cerebral physiology and pathology and should not be misinterpreted as negating the patient’s need for the CT scan. Because of these limitations, at this time, rcSO2 monitoring should not be used diagnostically but rather as an adjunct decision tool to assist in earlier detection of brain injury. Numerous clinical studies have demonstrated the necessity of early identification of brain injury in an effort to prevent secondary brain injury and improve neurologic outcomes. The utilization of rcSO2 monitoring to detect occult abnormal cerebral physiology and pathol-
ogy in AMS patients is an attractive noninvasive tool for the bedside clinician. Cerebral oximetry technology offers many benefits to the acute care physician in that it is portable, painless, and can produce results within minutes. This real-time information could prompt immediate interventions that improve patient outcomes. By using the mean difference between the rcSO2 from the left and right cerebrum, the study was able to identify those subjects with altered cerebral physiology and an abnormal head CT. We also demonstrated an association between decreased rcSO2 and SDH and increased rcSO2 and
Fig. 7. Decreased rcSO2 in the left cortex and increased rcSO2 in the right cortex among patients with left SDH and right EDH (n = 10). Red triangles represent the mean for all subjects; the green triangles, the fitted mean; and the gray-shaded area, the 95% confidence interval of the means.
362
I. Kane et al. / American Journal of Emergency Medicine 32 (2014) 356–362
EDH. Our results highlight the need for further randomized studies of rcSO2 monitoring among children who present to the ED with suspicion of intracranial injury. Prospective studies are also needed to establish normative rcSO2 values among healthy pediatric patients in the ED. In summary, Cerebral Oximetry rcSO2 monitoring can immediately detect altered cerebral physiology and pathology among a selected population. Cerebral Oximetry has demonstrated its potentials as a screening tool to identify, localize, and characterize intracranial injuries among children with AMS without trauma. In this study population the difference between either Left or Right cerebral rcSO2 readings most reliably identified those subjects with altered cerebral physiology and intracranial pathology and abnormal head CT. Either side or Both cerebral rcSO2 readings that were less than 50% or unilateral readings greater than 80% were significant for Subdural(b 50% cerebral rcSO2 mean) or Epidural (N 80% cerebral rcSO2 mean). Cerebral Oximetry in the emergency department has the potential to screen for occult intracranial injuries among this vulnerable population but further studies are warranted. References [1] Salonia R, Bell MJ, Kochanek PM, et al. The utility of near infrared spectroscopy in detecting intracranial hemorrhage in children. J Neurotrauma 2012;29:1047–53. [2] Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet 2009;374:160–1170. [3] Drayna PC, Abramo TJ, Estrada CE. Near-infrared spectroscopy in the critical setting. Pediatr Emerg Care 2011;27:432–42. [4] Ghanayem NS, Wernovsky G, Hoffman GM. Near infrared spectroscopy as a hemodynamic monitor in critical illness. Pediatr Crit Care Med 2011;12: S27–32. [5] Murkin JM, Arango A. Near-infrared spectroscopy as an index of brain and tissue oxygenation. Br J Anaesth 2009;103:i3–i13. [6] Henson LC, Calalang C, Temp JA, et al. Accuracy of a cerebral oximeter in healthy volunteers under conditions of isocapnic hypoxia. Anesthesiology 1998;88:58–65.
[7] Kreeger RN, Ramamoorthy C, Nicolson SC, et al. Evaluation of pediatric nearinfrared cerebral oximeter for cardiac disease. Ann Thorac Surg 2012;94:1527–33. [8] Uebling A, Furck AK, Hansen JH, et al. Perioperative cerebral and somatic oxygenation in neonates with hypoplastic left heart syndrome or transposition of the great arteries. J Thorac Cardiovasc Surg 2011;142:523–30. [9] Hansen JH, Schlangen J, Armbrust S, et al. Monitoring of regional tissue oxygenation with near-infrared spectroscopy during the early postoperative course after superior cavopulmonary anastomosis. Eur J Cardiothorac Surg 2013;43:e37–43. [10] Alderliesten T, Lemmers PM, Smarius JJ, et al. Cerebral oxygenation, extraction, and autoregulation in very preterm infants who develop peri-intraventricular hemorrhage. J Pediatr 2013;162:698–704. [11] Tina LG, Frigiola A, Abella R, et al. Near Infrared Spectroscopy in healthy preterm and term newborns: correlation with gestational age and standard monitoring parameters. Curr Neurovasc Res 2009;6:148–54. [12] Bernal NP, Hoffman GM, Ghanayem NS, et al. Cerebral and somatic near-infrared spectroscopy in normal newborns. J Pediatr Surg 2010;45:1306–10. [13] Dunham CM, Ransom KJ, Flowers LL, Siegal JD, Kohli CM. Cerebral hypoxia in severely brain-injured patients is associated with admission Glasgow coma scale score, computed tomographic severity, cerebral perfusion pressure, and survival. J Trauma 2004;56:482–91. [14] Kim MB, Ward DS, Cartwright CR, Kolano J, Chlebowski S, Henson LC, et al. Estimation of jugular venous O2 saturation from cerebral oximetry or arterial O2 saturation during isocapnic hypoxia. J Clin Monit Comput 2000; 16:191–9. [15] Myburgh JA. Quantifying cerebral autoregulation in health and disease. Crit Care Resusc 2004;6(1):59–67. [16] Obrig H. NIRS in clinical neurology - a 'promising' tool?Neuroimage 2013 pii: S1053-8119(13)00295-4. [17] Francis SJ, Ravindran G, Visvanathan K, et al. Screening for unilateral intracranial abnormalities using near infrared spectroscopy. A preliminary report. J Clin Neurosci 2005;12:291–5. [18] Gopinath S, Robertson C, Contant C, et al. Early detection of delayed traumatic intracranial hematomas using near infrared spectroscopy. J Neurosurg 1995;83: 438–44. [19] Kahraman S, Kayali H, Atabey C, et al. The accuracy of near infrared spectroscopy in detection of subdural and epidural hematomas. J Trauma 2006;61:1480–3. [20] Robertson C, Zager E, Narayan R, et al. Clinical evaluation of a portable nearinfrared device for detection of traumatic intracranial hematomas. J Neurotrauma 2010;27:1597–604.
