British Journal of Anaesthesia 112 (6): 1109–14 (2014)
`Advance Access publication 15 April 2014 . doi:10.1093/bja/aeu095
`
`TRANSLATIONAL RESEARCH
`
`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`Feasibility and accuracy of nasal alar pulse oximetry
`T. E. Morey1*, M. J. Rice1, T. Vasilopoulos2, D. M. Dennis1,3 and R. J. Melker1,3
`
`1 Department of Anesthesiology, University of Florida, PO 100254, 1600 SW Archer Road, Gainesville, FL 32601-0254, USA
`2 Department of Psychiatry and Behavioral Neuroscience, University of Chicago, 5841 S. Maryland Ave., MC 3077, CNPRU, Room L-603,
`Chicago, IL 60637, USA
`3 Xhale, Inc., 3630 SW 47th Ave., Suite 100, Gainesville, FL 32608, USA
`* Corresponding author. E-mail: morey@ufl.edu
`
`Editor’s key points
`† Nasal pulse oximetry is
`an attractive alternative
`when digital perfusion
`is poor or inaccessible.
`† Use of monitoring devices
`outside of their intended
`use requires specific
`validation.
`† Results obtained from
`volunteer studies may not
`translate to the clinical
`environment.
`† The performance of
`monitoring devices is
`most crucial at the
`margins of the normal
`range.
`
`Background. The nasal ala is an attractive site for pulse oximetry because of perfusion by
`branches of the external and internal carotid arteries. We evaluated the accuracy of a novel
`pulse oximetry sensor custom designed for the nasal ala.
`Methods. After IRB approval, healthy non-smoking subjects [n¼12; aged 28 (23–41) yr; 6M/6F]
`breathed hypoxic mixtures of fresh gas by a facemask to achieve oxyhaemoglobin saturations
`of 70–100% measured by traditional co-oximetry from radial artery samples. Concurrent alar
`and finger pulse oximetry values were measured using probes designed for these sites. Data
`were analysed using the Bland–Altman method for multiple observations per subject.
`Results. Bias, precision, and accuracy root mean square error (ARMS) over a range of 70–100%
`were significantly better for the alar probe compared with a standard finger probe. The mean
`bias for the alar and finger probes was 0.73% and 1.90% (P,0.001), respectively, with
`corresponding precision values of 1.65 and 1.83 (P¼0.015) and ARMS values of 1.78% and
`2.72% (P¼0.047). The coefficients of determination were 0.96 and 0.96 for the alar and
`finger probes, respectively. The within/between-subject variation for the alar and finger
`probes were 1.14/1.57% and 1.87/1.47%, respectively. The limits of agreement were
`3.96/22.50% and 5.48/21.68% for the alar and finger probes, respectively.
`
`Conclusions. Nasal alar pulse oximetry is feasible and demonstrates accurate pulse oximetry
`values over a range of 70–100%. The alar probe demonstrated greater accuracy compared
`with a conventional finger pulse oximeter.
`
`Keywords: alar nasal cartilages; lateral nasal cartilages; nasal cartilages; oximetry; pulse
`oximetry
`
`Accepted for publication: 22 November 2013
`
`Pulse oximetry remains a vital tool in healthcare facilities and a
`standard monitor to measure the oxygenation of patients re-
`ceiving anaesthesia. Fingers are a commonly used site for
`probe placement, but may be suboptimal because diminished
`perfusion from numerous causes may result in loss of sufficient
`pulse signal to provide an accurate pulse oximetry oxygen sat-
`).1 – 3 Additionally, finger use may be limited by
`uration (SpO2
`injury, presence on the surgical field, non-invasive arterial pres-
`sure cuff interruption, arm tucking, and shivering.4 – 6 Finally,
`digital pulse oximetry produces an unacceptably high inci-
`dence of failure (79%) during emergency pre-hospital airway
`management.7 Therefore, application to sites other than
`fingers may be useful. Several alternative probe sites on the
`head have been suggested (e.g. earlobe, forehead, cheek,
`tongue) because perfusion to these structures may provide ad-
`equate pulsatile signal even in the presence of significant
`pathophysiology and because the time to detect desaturation
`
`is less when measured from a site on the head, compared with
`a finger or toe.8 – 12
`Recently, Saban and colleagues13 demonstrated that
`branches of both the facial and ophthalmic arteries perfuse the
`nasal alae, the lateral walls of the nares (Fig. 1).13 The ophthalmic
`artery, which is a branch of the internal carotid artery, has been
`shown to autoregulate during experimental hypotension to pre-
`serve blood flow.14 15 Integrating these anatomical and physio-
`logical data, we hypothesize that the nasal ala should be a
`suitable pulse oximetry location. As a first step, the primary aim
`of this investigation was to validate the performance character-
`istics ofalarpulse oximetry probes similar to the seminal pulse ox-
`imeter evaluations of Yelderman and New16 in 1983. In healthy
`subjects receiving hypoxic mixtures of gas to generate oxygen
`saturations of 70–100%, we compared novel alar pulse oximetry
`and conventional finger pulse oximetry with the reference stand-
`ard, co-oximetry measurement of radial artery blood specimens.
`
`& The Author [2014]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
`For Permissions, please email: journals.permissions@oup.com
`
`1
`
`Masimo Ex. 2008
`Apple v. Masimo
`IPR2020-01722
`
`

`

`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`Morey et al.
`
`time-stamped data (SpO2
`, heart rate) were collected from the
`probes at a rate of 1 Hz and automatically logged onto a spread-
`sheet supported by a personal computer for later analysis.
`After instrumentation, subjects began the experimental
`protocol for desaturation similar to that previously conducted
`in this laboratory.17 In brief, subjects were in a 30–458 head-
`up position. Two arterial blood samples were obtained ,30 s
`apart while each subject breathed room air. After aspiration of
`blood (2 ml), the samples were immediately processed and ana-
`lysed to measure the SaO2. Investigators placed a tight-fitting
`full facemask to prevent entrainment of room air. Thereafter,
`hypoxia was induced by supplying a fresh gas mixture of nitro-
`gen, room air, and carbon dioxide. Each plateau level of oxy-
`haemoglobin saturation was maintained for at least 30 s until
`pulse oximeter readings stabilized. Then, two arterial blood
`samples were obtained (cid:3)30 s apart. Each stable plateau was
`therefore maintained for at least 60 s defined by SpO2 variation
`≤3%. The different plateaus were nominally at 98–100% (room
`air saturation), 93%, 90%, 87%, 85%, 82%, 80%, 77%, 75%, and
`70%. This procedure was performed twice with the second
`epoch completing any saturation values missed during the
`first run. Approximately 25 samples were obtained across this
`anticipated range of saturation (70–100%) for each subject
`that completed the protocol.
`
`Statistical analysis
`Descriptive data are reported as mean (SD). Data were analysed to
`calculate bias, precision, and the accuracy root mean square error
`(ARMS). ARMS is a composite value determined by both bias and
`precision and is used as an overall index of uncertainty of the
`SpO2
`value by regulatory authorities such as the US National Insti-
`tute of Standards in Technology with values ,2–3% considered
`(cid:3)
`
`bias2 + precision2
`to be acceptable.18 ARMS was calculated as
`.
`These values were calculated for oxyhaemoglobin saturation
`bands of 70–80%, 80–90%, and 90–100%, and 70–100% (all
`data).18 Differences in bias between the nasal alar probe and
`finger probe were assessed using a paired Student’s t-test, utiliz-
`ing the full sample of subject measurements across all saturation
`bands and within each band. A folded F-test for homogeneity of
`variances was similarly used to determine significant differences
`in precision between the probes.
`In review of the medical literature, we did not discover a
`method to statistically evaluate ARMS values. We did learn,
`however, of a potential approach for similar metrics that have
`been utilized in other fields by Nilsson and colleagues.19 We
`adapted these methods to evaluate statistical differences in
`ARMS between probes. ARMS values were calculated for each
`subject (n¼12). Specifically, subject-level bias and precision
`were first calculated using the multiple measurements obtained
`from each subject. Then, these bias and precision values were
`combined to determine ARMS for each subject. This resulted in
`two sets of ARMS values, with one set for each type of probe.
`The average ARMS values for each probe were then compared
`using a paired Student’s t-test. To the best of our knowledge,
`this is the first study to adapt this type of approach to examine
`statistical differences in ARMS.
`
`BJA
`
`Fig 1 Reproduction of the anastomotic polygonal system forming
`the arterial supply of the nose that includes the facial artery from
`the external carotid artery and the nasal branch of the ophthalmic
`artery from the internal carotid artery. Note the anastomosis of the
`external and internal carotid arterial system at the right and left
`ala. Reproduced from Saban and colleagues13 with permission.
`&(2012) American Medical Association. All rights reserved.
`
`Methods
`Subject recruitment
`This non-treatment investigation was approved by the Univer-
`sity of California at San Francisco Committee on Human Re-
`search under protocol H6301-01706-24 and was conducted
`at the HYPO2XIALAB at that institution (San Francisco, CA,
`USA). Written,
`informed consent was obtained from all
`recruited subjects, who were aged 18–50 yr. Subjects self-
`identified their own ethnic and racial categories as defined by
`the National Institutes of Health. Exclusionary criteria were
`history of tobacco smoking, hypertension, respiratory disease,
`or haemoglobin ≤10 g dl21.
`
`Protocol
`On the day of study, a catheter was inserted using aseptic tech-
`nique into a radial artery of each subject in order to aspirate
`blood for determination of arterial oxyhaemoglobin saturation
`(SaO2 ). Values for SaO2 were measured using a multi-wavelength
`oximeter (OSM3w Hemoximeter, Radiometer Medical A/S,
`Copenhagen, Denmark), which served as the reference standard
`for subsequent comparisons with SpO2
`readings. An alar pulse
`oximeter sensor
`(Assurance Biosense,
`Inc., Glastonbury,
`CT, USA) was placed on the right ala to measure SpO2
`. An add-
`itional, conventional pulse oximeter probe and monitor was
`placed on the index finger for reference purposes (N-595 Pulse
`Oximeter, Mallinckrodt Inc., St Louis, MO, USA). Date- and
`
`1110
`
`2
`
`

`

`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`BJA
`
`to calculate the bias and limits of agreement as suggested by
`Myles and Cui.22 P,0.05 was considered significant.
`
`Results
`Patient characteristics data
`Subjects (n¼12) aged 28 (23–41) yr were divided evenly
`between males (n¼6) and females (n¼6). The ethnic and
`racial categories were as follows: six Asians, four non-Hispanic
`whites, one Hispanic white, and one African-American. Skin
`tone was variable and characterized as light (n¼3), light-
`medium (n¼2), medium (n¼3), medium-dark (n¼2), and
`dark (n¼2). Eleven subjects completed the entire protocol of
`the study with 25 or more saturation observations. One
`subject complained of anxiety during the first episode of desat-
`uration and self-terminated participation after collection of
`eight saturation observations. All subjects’ data were included
`in this analysis.
`
`Oximetry data
`A representative study (Subject 1) of the alar SpO2
`values and
`heart rate over time are shown in an illustrative example
`(Fig. 2). A number of co-oximetry oxyhaemoglobin saturations
`were episodically collected as well. Using these data from
`Figure 2 and from all other subjects (n¼12 for all subjects), ana-
`lysis was performed to determine the bias, precision, and ARMS
`for oxyhaemoglobin saturation bands of 70–80%, 80–90%,
`and 90–100% (Table 1). The bias from the alar probe was sig-
`nificantly less than that from the finger probe for all bands of
`saturation. However, across the separate bands, there were
`not consistent significant differences in precision and ARMS
`between probes. Using the method comparison technique for
`multiple measurements, no magnitude effect on bias was
`evident. Further, the Bland–Altman method comparison ana-
`lysis was also performed with all saturation results (70–100%)
`as noted in Table 2.
`Revised Bland–Altman analysis of pooled data for repeated
`measurements showing bias plots for the alar (Fig. 3) and
`finger (Fig. 4) probes are shown.21 22 Overall, less bias across
`all saturations was noted with the alar probe compared with
`the finger probe (P,0.001). Additionally, the nasal alar probe
`
`Alar pulse oximetry accuracy
`
`Additionally, method comparisons were performed using
`revised Bland–Altman analysis using SaO2 as the reference
`standard. In the original Bland–Altman technique, only one
`measurement per subject was described under constant con-
`ditions.20 Some investigators, however, measure subject vari-
`ables (e.g. SpO2
`, SaO2) repeatedly under different conditions
`such as variable inspired concentrations of oxygen. Measure-
`ment of the subject more than once and under changing
`conditions violates assumptions of independent sampling.
`Therefore, they proposed a newer method that we have used
`herein.21 In the technique described in section 3 of their publi-
`cation [a required statistical analysis by the US Food and Drug
`Administration for 510(k) medical device clearance], Bland
`and Altman provide a way to determine within-subject vari-
`ance using analysis of variance (SigmaPlot 12.2, Systat Soft-
`ware, Inc., San Jose, CA, USA), between-subject variance,
`total variance, bias, and limits of agreement for subjects
`providing multiple samples under changing physiological con-
`ditions that would normally violate assumptions of independ-
`ent sampling.21 Additionally, we pooled data for each subject
`
`Pulse (min–1)
`
`125
`
`100
`
`75
`
`50
`
`SpO2
`SaO2
`Pulse
`
`0
`
`10
`
`20
`Time (min)
`
`30
`
`40
`
`100
`
`90
`
`80
`
`70
`
`60
`
`50
`
`Oxyhaemoglobin (%)
`
`Fig 2 Representative study of oxyhaemoglobin saturation and
`heart rate over time in a human subject receiving variable concen-
`trations of inspired oxygen by a facemask. Shown is the oxyhaemo-
`globin concentration measured by alar pulse oximetry (n¼2620)
`with some points also measured concurrently by co-oximetry
`(n¼25).
`
`Table 1 Bias and ARMS for alar and finger pulse oximetry data stratified by co-oximeter oxygen saturation bands from human subjects (n¼12)
`receiving variable fresh gas flow concentrations of inspired oxygen by a facemask. Oxygen saturation band grouping was determined by
`co-oximetry measurement
`
`Saturation (%)
`
`Observations
`
`Parameter
`
`70–80
`
`80–90
`
`90–100
`
`85
`
`97
`
`93
`
`Bias
`Precision
`ARMS
`Bias
`Precision
`ARMS
`Bias
`Precision
`ARMS
`
`Alar
`
`0.82
`1.77
`1.95
`1.13
`1.49
`1.87
`0.22
`1.51
`1.53
`
`Finger
`
`2.93
`1.98
`3.53
`2.06
`1.52
`2.56
`0.82
`1.24
`1.49
`
`P-value
`
`,0.001
`0.305
`0.023
`,0.001
`0.862
`0.155
`,0.001
`0.063
`0.396
`
`1111
`
`3
`
`

`

`BJA
`
`Morey et al.
`
`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`Upper limit=4.53
`
`Bias=2.02
`
`Lower limit=–0.50
`
`80
`
`85
`Mean [(S pO2
`+SaO2
`
`)/2] (%)
`
`90
`
`6 4
`
`2 0
`
`–2
`
`–4
`
`–6
`
`) (%)
`
`–SaO2
`
`Finger difference (SpO2
`
`Fig 4 Revised Bland–Altman analysis for multiple measurements
`of oxyhaemoglobin concentrations measured by finger pulse oxim-
`) and co-oximetry (SaO2) for human subjects (n¼12) re-
`etry (SpO2
`ceiving variable concentrations of inspired oxygen by a facemask.
`Shown are individual subjects’ pooled observations, bias (solid
`blue line), and upper and lower limits of agreement (dashed blue
`line).
`
`Bland–Altman analysis using non-pooled data is shown in
`Supplementary Figs S1–S4.)
`
`Discussion
`We designed and tested the accuracy of a novel, alar pulse ox-
`imeter sensor in healthy volunteers during both normal (room
`air) and hypoxic conditions, using a standard method for
`testing pulse oximeters and sensors. The ARMS of the alar
`data set compares favourably with the accepted instrument
`specifications of a ≤4% ARMS for functional oxygen saturation
`promulgated by the International Organization for Standard-
`ization (IOS), ≤3% for the US Food and Drug Administration,
`and ≤2% ARMS for most other regulatory agencies.18 Moreover,
`the sensor demonstrated acceptable accuracy not only for all
`aggregated data (70–100%), but also at all incremental satur-
`ation bands tested. An ARMS of 2% does not mean that all points
`are within 2% of the reference standard co-oximetry, but rather
`that most points ((cid:3)67%) are within this range. Furthermore, a
`sensor could theoretically have a relatively poor bias or low pre-
`cision, but not both, and meet the ambiguity target of ≤4%
`ARMS. Manufacturers of pulse oximeters generally claim an ac-
`curacy of 2–3% over the saturation range of 70–100%, which
`would roughly correspond to an ARMS of 2%, but the Inter-
`national Standard Organization reminds us that ‘. . . only
`about two-thirds . . .’ of measurements will fall within the
`stated twice the ARMS.1 – 3 23 Likewise, Batchelder and Raley18
`point out that the ARMS represents a group of values and does
`not allow for consideration of temporary ‘pop-ups’ or assess
`dynamic performance conditions. Notwithstanding,
`the
`present alar probe possessed acceptable characteristics as cur-
`rently assessed by various agencies. Furthermore, via several
`significance testing approaches, the nasal alar probe demon-
`strated significantly less bias, better precision, and lower ARMS
`
`Table 2 Bland–Altman analysis of oxyhaemoglobin saturation for
`alar and finger oximetry probes with comparison with co-oximetry
`saturations as the reference standard method in human subjects
`(n¼12) receiving variable concentrations of inspired oxygen by a
`facemask
`
`Parameter
`
`Alar
`
`Finger
`
`n
`Regression
`Coefficient of determination
`(r2)
`Within-subject variation
`Between-subject variation
`Total variation
`Standard deviation
`Bias
`Upper limits of agreement
`Lower limit of agreement
`Overall ARMS
`
`275
`y¼0.96 x+3.92
`0.96
`
`275
`y¼0.89 x+10.97
`0.96
`
`1.14
`1.57
`2.71
`1.65
`0.73
`3.96
`22.50
`1.78
`
`1.87
`1.47
`3.34
`1.83
`1.90
`5.48
`21.68
`2.72
`
`Upper limit=3.44
`
`Bias=0.56
`
`6 4 2 0
`
`–2
`
`–4
`
`–6
`
`80
`
`Lower limit=–2.31
`
`85
`Mean [(S pO2
`+SaO2
`
`)/2] (%)
`
`90
`
`Fig 3 Revised Bland–Altman analysis for multiple measurements
`of oxyhaemoglobin concentrations measured by alar pulse oxim-
`) and co-oximetry (SaO2) for human subjects (n¼12) re-
`etry (SpO2
`ceiving variable concentrations of inspired oxygen by a facemask.
`Shown are individual subjects’ pooled observations, bias (solid
`blue line), and upper and lower limits of agreement (dashed blue
`line).
`
`was overall more precise than the finger probe (P¼0.017). For
`both Figures 3 and 4, the outlying point originates from the
`subject that completed only eight saturation values at the
`greater end of the desaturation protocol before withdrawing
`from the investigation. After calculating ARMS for each subject
`for each probe, the average ARMS for the alar probe (1.76%)
`and finger probe (2.54%) using this approach closely
`matched the overall ARMS values that were calculated using
`each subject measurement separately (Table 2). Using this ap-
`proach, the average ARMS for the alar probe was significantly
`less than that for the finger probe (P¼0.047). (Note: The
`
`1112
`
`) (%)
`
`–Sa
`
`O2
`
`O2
`
`Alar difference (Sp
`
`4
`
`

`

`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`BJA
`
`population was entirely healthy with a normal perfusion
`state. Specific testing is required with subjects in a low perfu-
`sion state and subjects with impaired circulation to determine
`if alar pulse oximetry is accurate in these subjects when digital
`pulse oximetry fails. Finally, the long-term effects of continu-
`ous alar oximetry over days and weeks at this anatomical loca-
`tion are not known.
`In summary, we have demonstrated acceptable accuracy of
`a novel nasal alar pulse oximetry sensor in subjects both with
`normal and low oxygen saturations and also have provided evi-
`dence that nasal alar pulse oximetry may be less biased, more
`precise, and more accurate than conventional finger pulse ox-
`imetry. Future studies should focus on the use and accuracy of
`this sensor in subjects under a variety of pathophysiological
`conditions such as trauma, cardiopulmonary bypass, and in
`subjects in critical care units.
`
`Supplementary material
`is available at British Journal of
`Supplementary material
`Anaesthesia online.
`
`Authors’ contributions
`The authors individually contributed to the study design and
`data acquisition (R.J.M., T.E.M.), data analysis (T.E.M., D.M.D.,
`T.V.), and early drafts of the manuscript (T.E.M., M.J.R., R.J.M.)
`and critical revisions (T.E.M., M.J.R., T.V., R.J.M., D.M.D.). All
`authors have approved the final version of the manuscript.
`
`Acknowledgements
`The authors appreciate the time and careful effort of the staff
`of the HYPO2XIALAB located at the Department of Anesthesi-
`ology, University of California, San Francisco, San Francisco,
`CA, USA.
`
`Declaration of interest
`R.J.M. and D.M.D. are employees of Xhale, Inc., the company
`that sponsored the study. T.E.M. is a consultant for Xhale, Inc.
`R.J.M., D.M.D., and T.E.M. also own equity in Xhale, Inc. In add-
`ition, the University of Florida owns equity in Xhale, Inc. If a
`device is sold commercially, then the authors listed above
`and the University of Florida could benefit financially.
`
`Funding
`The project was supported by Xhale, Inc., Gainesville, FL, USA.
`
`References
`1 Brimacombe J, Keller C. Successful pharyngeal pulse oximetry in
`low perfusion states. Can J Anaesth 2000; 47: 907–9
`2 Secker C, Spiers P. Accuracy of pulse oximetry in patients with low
`systemic vascular resistance. Anaesthesia 1997; 52: 127–30
`3 Wilson BJ, Cowan HJ, Lord JA, Zuege DJ, Zygun DA. The accuracy of
`pulse oximetry in emergency department patients with severe
`sepsis and septic shock: a retrospective cohort study. BMC Emerg
`Med 2010; 10: 9
`
`1113
`
`Alar pulse oximetry accuracy
`
`when compared with a finger probe when all saturations
`(70–100%) were analysed. These differences in bias were also
`evident when analyses were performed across oxyhaemoglo-
`bin saturation bands of 70–80%, 80–90%, and 90–100%.
`However, the significant differences in precision and ARMS were
`not similarly consistent across each saturation band. These
`inconsistent results could be due to the reduced sample sizes
`of the analyses using the separate bands, which would reduce
`power to detect group differences.
`We specifically selected the nasal ala as a potential new site
`for pulse oximetry because of its unique arterial perfusion
`characteristics. Recently, Saban and colleagues13 used ultrason-
`ography on 40 living subjects and anatomical dissection on 20
`cadavers to demonstrate that the nose is perfused by a four
`component, multidirectional, arterial arcade to form a polygon-
`al system of vessels that anastomose the external and internal
`carotid arteries (Fig. 1). In contrast, the arterial supply of the ear,
`pharynx/cheek, and tongue originates solely from the external
`carotid artery. During periods of severe pathophysiology (e.g.
`shock, hypothermia, stress) that potentially require vasopressor
`therapy, pulse oximeter sensors on the fingers may cease to
`function because of reduction in systemic blood flow, particular-
`ly to the periphery. During these conditions, however, cerebral
`perfusion may be maintained via the internal carotid artery
`with autoregulation of blood flow to the central nervous
`system. Although not answered in the present feasibility in-
`vestigation, we hypothesize that alar pulse oximetry will be a
`superior site for pulse oximetry during periods of low systemic
`blood flow because of preserved arterial perfusion to the
`central nervous system. Since the nasal branch of the ophthal-
`mic artery provides a sample of internal carotid artery flow
`and partially perfuses the ala, an adequate pulse oximetry
`signal could be preserved to provide accurate SpO2
`values.
`The only other region of the head with a blood supply origin-
`ating from the internal carotid artery and assessable non-
`invasively for oximetry is the forehead which is perfused by a
`branch of the internal carotid artery (supraorbital artery).
`This oximetry site has been demonstrated to have acceptable
`bias and limits of agreement for patients undergoing vascular
`surgery, is more accurate than finger SpO2
`measurement in crit-
`ically ill patients receiving high-dose vasopressor therapy, and
`monitor.8 – 12 24 – 26
`responds more rapidly than a finger SpO2
`Monitoring at this site requires reflectance oximetry, however,
`wherein precise localization of the sensor and the presence of
`a corresponding vein in close proximity may cause erroneously
`values.13 27 To prevent erroneous values, the
`decreased SpO2
`use of a circumferential headband is suggested to dampen
`venous pulsations.14 15 28 Because these data demonstrate
`that monitoring from the nasal ala is feasible, accurate SpO2
`monitoring using transmittance oximetry may be useful for
`patients where traditional finger placement fails and is an alter-
`native approach to the supraorbital artery.
`The major limitation of this study was subject selection. That
`is, only healthy adult volunteers were observed over a short
`time period. In this initial investigation, no paediatric or neo-
`natal subjects were included that would have required a
`smaller alar pulse oximetry probe. Moreover, the subject
`
`5
`
`

`

`Downloaded from https://academic.oup.com/bja/article/112/6/1109/2919827 by guest on 02 August 2021
`
`BJA
`
`4 Gehring H, Hornberger C, Matz H, Konecny E, Schmucker P. The
`effects of motion artifact and low perfusion on the performance
`of a new generation of pulse oximeters in volunteers undergoing
`hypoxemia. Respir Care 2002; 47: 48–60
`5 Barker SJ, Shah NK. Effects of motion on the performance of pulse
`oximeters in volunteers. Anesthesiology 1996; 85: 774–81
`6 Barker SJ. ‘Motion-resistant’ pulse oximetry: a comparison of new
`and old models. Anesth Analg 2002; 95: 967–72
`7 Davis DP, Aguilar S, Sonnleitner C, Cohen M, Jennings M. Latency
`and loss of pulse oximetry signal with the use of digital probes
`during prehospital rapid-sequence intubation. Prehosp Emerg
`Care 2011; 15: 18–22
`8 Yu H, Liu B. Successful use of pharyngeal pulse oximetry with the
`oropharyngeal airway in severely shocked patients. Anaesthesia
`2007; 62: 734–6
`9 Rogers SR, Gan TJ. Pulse oximeter probe sheath for buccal use.
`Anesth Analg 1997; 85: 1414
`10 Reynolds LM, Nicolson SC, Steven JM, Escobar A, McGonigle ME,
`Jobes DR. Influence of sensor site location on pulse oximetry
`kinetics in children. Anesth Analg 1993; 76: 751–4
`11 Cote´ CJ, Daniels AL, Connolly M, Szyfelbein SK, Wickens CD. Tongue
`oximetry in children with extensive thermal injury: comparison
`with peripheral oximetry. Can J Anaesth 1992; 39: 454–7
`12 Phillips JP, Kyriacou PA, Jones DP, Shelley KH, Langford RM. Pulse ox-
`imetry and photoplethysmographic waveform analysis of the
`esophagus and bowel. Curr Opin Anaesthesiol 2008; 21: 779–83
`13 Saban Y, Andretto Amodeo C, Bouaziz D, Polselli R. Nasal arterial
`vasculature: medical and surgical applications. Arch Facial Plast
`Surg 2012; 14: 429–36
`14 Kolodjaschna J, Berisha F, Lung S, Schima H, Polska E, Schmetterer L.
`Comparison of the autoregulatory mechanisms between middle
`cerebral artery and ophthalmic artery after thigh cuff deflation in
`healthy subjects. Invest Ophthalmol Vis Sci 2005; 46: 636–40
`15 Ne´meth J, Kne´zy K, Tapaszto´ B, Kova´cs R, Harka´nyi Z. Different auto-
`regulation response to dynamic exercise in ophthalmic and central
`retinal arteries: a color Doppler study in healthy subjects. Graefes
`Arch Clin Exp Ophthalmol 2002; 240: 835–40
`
`Morey et al.
`
`16 Yelderman M, New W. Evaluation of pulse oximetry. Anesthesiology
`1983; 59: 349–52
`17 Feiner JR, Severinghaus JW, Bickler PE. Dark skin decreases the
`accuracy of pulse oximeters at low oxygen saturation: the effects
`of oximeter probe type and gender. Anesth Analg 2007; 105:
`S18–23
`18 Batchelder PB, Raley DM. Maximizing the laboratory setting for
`testing devices and understanding statistical output in pulse oxim-
`etry. Anesth Analg 2007; 105: S85–94
`19 Nilsson H, Olsson H, Juslin P. The cognitive substrate of subjective
`probability. J Exp Psychol Learn Mem Cogn 2005; 31: 600–20
`20 Bland JM, Altman DG. Statistical methods for assessing agreement
`between two methods of clinical measurement. Lancet 1986; 1:
`307–10
`21 Bland JM, Altman DG. Agreement between methods of measure-
`ment with multiple observations per individual. J Biopharm Stat
`2007; 17: 571–82
`22 Myles PS, Cui J. Using the Bland–Altman method to measure agree-
`ment with repeated measures. Br J Anaesth 2007; 99: 309–11
`23 Milner QJW, Mathews GR. An assessment of the accuracy of pulse
`oximeters. Anaesthesia 2012; 67: 396–401
`24 Wax DB, Rubin P, Neustein S. A comparison of transmittance and
`reflectance pulse oximetry during vascular surgery. Anesth Analg
`2009; 109: 1847–9
`25 Nesseler N, Fre´nel J-V, Launey Y, Morcet J, Malle´dant Y, Seguin P.
`Pulse oximetry and high-dose vasopressors: a comparison
`between forehead reflectance and finger transmission sensors.
`Intensive Care Med 2012; 38: 1718–22
`26 Choi SJ, Ahn HJ, Yang MK, et al. Comparison of desaturation
`and resaturation response times between transmission and re-
`flectance pulse oximeters. Acta Anaesthesiol Scand 2010; 54:
`212–7
`27 Dassel AC, Graaff R, Sikkema M, Meijer A, Zijlstra WG, Aarnoudse JG.
`Reflectance pulse oximetry at the forehead improves by pressure
`on the probe. J Clin Monit 1995; 11: 237–44
`28 Covidien. NellcorTM Forehead SpO2 Sensor with OxiMax Technology
`Application Guide. 2011
`
`Handling editor: P. S. Myles
`
`1114
`
`6
`
`