Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Next Article in Journal
Erroneous Classification and Coding as a Limitation for Big Data Analyses: Causes and Impacts Illustrated by the Diagnosis of Clavicle Injuries
Previous Article in Journal
DeepGenMon: A Novel Framework for Monkeypox Classification Integrating Lightweight Attention-Based Deep Learning and a Genetic Algorithm
Previous Article in Special Issue
Validity of Wearable Inertial Sensors for Gait Analysis: A Systematic Review
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Diagnostics
Volume 15
Issue 2
10.3390/diagnostics15020128
diagnostics-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Evaluation of Non-Invasive Hemoglobin Monitoring in Perioperative Patients: A Retrospective Study of the Rad-67TM (Masimo)

by
Philipp Helmer
1,*,†,
Andreas Steinisch
1,†,
Sebastian Hottenrott
1,
Tobias Schlesinger
2,
Michael Sammeth
1,3,
Patrick Meybohm
1 and
Peter Kranke
1
1
Department of Anaesthesiology, Intensive Care, Emergency and Pain Medicine, University Hospital Würzburg, Oberdürrbacher Str. 6, 97080 Würzburg, Germany
2
Department of Anaesthesiology and Intensive Care, BG Murnau, Professor-Küntscher-Str. 8, 82418 Murnau, Germany
3
Department of Applied Sciences and Health, Coburg University of Applied Sciences and Art, Friedrich-Streib-Str. 2, 96450 Coburg, Germany
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Diagnostics 2025, 15(2), 128; https://doi.org/10.3390/diagnostics15020128
Submission received: 25 November 2024 / Revised: 2 January 2025 / Accepted: 3 January 2025 / Published: 8 January 2025
(This article belongs to the Special Issue Wearable Sensors for Health Monitoring and Diagnostics)
Browse Figures
Review Reports Versions Notes

Abstract

:
Background: Hemoglobin (Hb) is a crucial parameter in perioperative care due to its essential role for oxygen transport and tissue oxygenation. Accurate Hb monitoring allows for timely interventions to address perioperative anemia and, thus, prevent morbidity and mortality. Traditional Hb measurements rely on invasive blood sampling, which significantly contributes to iatrogenic anemia and poses discomfort and increased infection risks. The advent of non-invasive devices like Masimo’s Rad-67™, which measures Hb using pulse CO-oximetry (SpHb), offers a promising alternative. This study evaluates the accuracy of SpHb compared to clinical standard blood gas analysis (BGA) in perioperative patients. Methods: This retrospective study analyzed 335 paired Hb measurements with an interval <15 min between SpHb and BGA in the operating theater and post-anesthesia care unit of a university hospital. Patients experiencing hemodynamic instability, acute bleeding, or critical care were excluded. Statistical analysis included Bland–Altman plots and Pearson correlation coefficients (PCCs) to assess the agreement between SpHb and BGA. Potential confounders, e.g., patient age, skin temperature, sex, perfusion index (PI), and atrial fibrillation, were also analyzed. Results: The bias of the SpHb compared to BGA according to Bland–Altman was 0.00 g/dL, with limits of agreement ranging from −2.70 to 2.45 g/dL. A strong correlation was observed (r = 0.79). Overall, 57.6% of the paired measurements showed a deviation between the two methods of ≤±1 g/dL; however, this applied to only 33.3% of the anemic patients. Modified Clark’s Error Grid analysis showed 85.4% of values fell within clinically acceptable limits. Sex was found to have a statistically significant, but not clinically relevant, effect on accuracy (p = 0.02). Conclusions: The Rad-67TM demonstrates reasonable accuracy for non-invasive SpHb, but exhibits significant discrepancies in anemic patients with overestimating low values. While it offers potential for reducing iatrogenic blood loss, SpHb so far should not replace BGA in critical clinical decision-making.

1. Introduction

Hemoglobin plays a crucial role in oxygen transport and tissue oxygenation, making it a critical parameter in perioperative care [1]. Anemia, characterized by decreased hemoglobin levels, can impair tissue oxygenation and increase mortality, particularly in vulnerable perioperative patients [2]. These individuals face an increased risk of bleeding, which often leads to anemia and significantly worse patient outcomes [3]. Hemorrhagic events are, thus, a major contributor to perioperative mortality [4,5]. Consequently, the close monitoring of hemoglobin levels in perioperative settings is essential for timely interventions and optimizing patient safety.
Traditionally, hemoglobin is measured by invasive blood sampling methods for point-of-care diagnostics such as blood gas analysis (BGA). These methods require a blood sample to be drawn from the patient, which either requires invasive catheters or venous puncture associated with discomfort and increased infection risk [6,7]. Frequent invasive blood draws contribute significantly to iatrogenic blood loss and can exacerbate anemia, counteracting the principles of Patient Blood Management (PBM) [8]. Thereby, PBM is a patient-centered interdisciplinary approach that focuses on patient safety and improving patient outcomes [9]. It is based on three key pillars with more than 100 specific aspects focusing on (i) the management of anemia with optimizing coagulation, (ii) the minimization of iatrogenic and unnecessary blood loss, and (iii) patient-centered decision-making, especially with the use of blood transfusions [10]. Frequent invasive blood draws contribute significantly to iatrogenic blood loss and can exacerbate anemia, counteracting the principles of Patient Blood Management (PBM).
The development of non-invasive hemoglobin measurement devices, such as Masimo’s Rad-67™, offers a promising solution to reduce the need for invasive blood draws and, therefore, minimize iatrogenic blood loss. The device utilizes pulse CO-oximetry, a continuous, non-invasive method to measure total hemoglobin concentration (SpHb) in the blood. The Rad-67™ sensor, placed on the patient’s finger, emits various wavelengths of light through the tissue and detects changes in light absorption, allowing for real-time hemoglobin monitoring without the need for invasive blood drawings.
Non-invasive hemoglobin monitoring offers several advantages. Continuous SpHb monitoring is associated with reduced intraoperative transfusions, lower postoperative bleeding rates, and shorter ICU stays [11]. In a recent study, the percentage of patients needing a transfusion decreased by 7.4% when Hb was continuously monitored, and the number of transfused units per patient decreased by 12.6% [12]. Furthermore, pediatric patients undergoing cranioplasty operations with continuous SpHb monitoring experienced fewer transfusions and better postoperative outcomes [13,14]. Similarly, the implementation of SpHb monitoring has even reduced mortality in various surgical settings, including cesarean sections and neurosurgery [15,16]. Furthermore, SpHb monitoring has been shown to decrease the incidence of unnoticed anemia, adhere to an Hb target range, and reduce inappropriate blood transfusions, contributing to better patient care and reduced health care costs [17,18,19].
Given the potential benefits of non-invasive hemoglobin monitoring in specific areas of care, we conducted a retrospective study to evaluate the measurement accuracy of the Rad-67™ device in general in perioperative patients, as the Rad-67™ was routinely used in our department. Our primary outcome was to investigate the accuracy of SpHb measurements compared to the clinical standard BGA and, secondarily, to identify factors influencing the device’s performance in a real-world clinical setting.

2. Methods

2.1. Study Design

This retrospective study was conducted in the operating theater and post-anesthesia care unit at the Department of Anaesthesiology, Intensive Care, Emergency, and Pain Medicine at the University Hospital Würzburg, Germany. Data collection and analysis were performed in accordance with Article §27 of Bavarian hospital law and the local ethics committee of the University of Würzburg approved the analysis (ref. no. 20240216-01). All simultaneous hemoglobin (Hb) measurements with an interval of less than 15 min, using both the Rad-67™ Pulse CO-Oximeter® and the clinical standard BGA, GEM Premier 5000 (Instrumentation Laboratory Company, Bedford, MA, USA), were included in the analysis. Measurements from patients who were hemodynamically unstable, suffering from acute bleeding, or receiving critical care were excluded.
Both measurement techniques were employed as part of the clinical standard between 2021 and 2023 in our department. The primary objective of this study was to investigate the measurement accuracy of the non-invasive SpHb measurements obtained by the Rad-67™ compared to the invasive clinical standard, defined as BGA. The secondary objective was to investigate factors potentially affecting measurement accuracy. The study was designed, conducted, and analyzed without financial support or any other involvement from industrial partners in order to prevent conflicts of interest. All employed devices, including the Rad-67™ device and all sensors, were purchased through regular channels.

2.2. Measurements of the Hemoglobin

In this study, we compared the measurement accuracy of a non-invasive spot-method for hemoglobin (SpHb) with the established clinical standard of point-of-care invasive hemoglobin (Hb) measurement employing blood gas analysis (BGA). The non-invasive SpHb was measured using the Rad-67™ Pulse CO-Oximeter® (Masimo, Irvine, CA, USA), a certified Class B medical device (IEC 60601-1-2:2007) authorized according to ANSI/AAMI ES 60601-1:2005 and EN/ISO 80601-2-61:201. A Rainbow® Super DCI®-mini sensor was placed on the patient’s index finger, utilizing (i) spectrophotometry, (ii) photoplethysmography (PPG), and (iii) Signal Extraction Technology (SET®). The Rad-67™ device uses multiple light-emitting diodes (LEDs) that emit various wavelengths of visible and infrared light (500–1400 nm), which pass through the tissue to a detector diode on the opposite site [20].
(i)
Spectrophotometry is a method used to measure how much light is absorbed by different hemoglobin species within the blood. The principle is based on the fact that various hemoglobin types (oxyhemoglobin, deoxyhemoglobin, carboxyhemoglobin, methemoglobin) and other blood plasma components absorb light at distinct wavelengths. By passing visible and infrared light through the tissue and measuring the light absorption, the Rad-67™ distinguishes between these hemoglobin types. Each compound has a unique absorption spectrum, and the detector captures the amount of light that passes through the blood. The difference in light absorption at specific wavelengths allows for the identification and quantification of the various hemoglobin types.
(ii)
Photoplethysmography (PPG) is an optical technique that detects changes in blood volume in the tissue over time, based on light absorption related to the pulsatile nature of blood flow [21]. The PPG signal consists of a pulsatile component (‘AC’) due to cardiac activity and a slowly varying baseline (‘DC’) influenced by factors such as respiration and thermoregulation. By detecting the changes in light absorption caused by variations in arterial blood volume, the device can distinguish between arterial and venous blood.
(iii)
The signal extraction technique (SET®) is employed to separate the arterial signal from noise and interference, such as motion artifacts [22]. By improving the signal-to-noise ratio, SET® assists the device in providing more reliable measurements of SpHb.
This combination of (i) to (iii) and the use of parallel algorithms are employed in the Rad-67™ to analyze the raw data for outputting the SpHb.
For the invasive clinical standard point-of-care method, approximately 1 mL of arterial or venous blood was drawn via a placed 20G arterial line (Arrow, Teleflex Medical, Wayne, PA, USA or Insyte-W, BD Medical, Franklin Lakes, NJ, USA) or a 22-16G venous catheter (Vasofix Safety Cannulae, B Braun). The blood samples were analyzed using the GEM 5000 Premier (Instrumentation Laboratory Company, Werfen, Bedford, MA, USA), which uses a spectrophotometric method that operates over a broader range of wavelengths compared to the non-invasive device, making it less susceptible to interference and yielding a direct measurement of hemoglobin without the need for motion artifact correction or pulsatile flow differentiation.

2.3. Statistical Analysis

All routine data were extracted from the patient data management system (PDMS) “Copra” (COPRA System GmbH, Berlin, Germany, V6.84.20) and exported to an Excel file. The primary endpoint of this study was assessed using Bland–Altman plots [23] and linear regression analyses, employing the Pearson Correlation Coefficient (PCC). For secondary endpoints, scatter plots were used for continuous variables, and box plots were used for categorical variables. For the evaluation of possible effects of the investigated attributes and factors, linear regression analysis was applied to continuous variables, while differences in the categorical variables were evaluated using the Kolmogorov–Smirnov test.
Descriptive statistics were computed as both the mean with the corresponding standard deviation (SD) and the median with the interquartile range (IQR). All statistical analyses were performed using the R platform (v4.2.0). Bland–Altman plots were generated to assess the presence of systematic biases, with the values obtained by the reference method (BGA) subtracted from the corresponding Rad-67™ hemoglobin readings. Linear regression analysis was performed to evaluate the correlation between the two measurement methods, employing PCC.
For the enhanced interpretability of the results, we also applied a modified version of Clark’s error grid, originally developed to compare blood glucose monitoring devices [24]. The advantage of Clark’s error grid lies in its ability to present not only numerical deviation, but also the clinical relevance of the discrepancy. The grid categorizes differences between the test and reference measurements into three zones: green for clinically insignificant differences, yellow for potential clinical relevance, and red for significant clinical risk with potential patient harm. Notably, Mills et al. had already utilized this method for comparing different Hb measurement techniques [25]. To further account for Hb thresholds relevant to patient blood management (PBM), we self-designed a modified version of the Clark’s error grid with modified boundaries of these zones. Particular emphasis was placed on diagnostic thresholds for mild anemia, as these patients should be monitored perioperatively for factors such as iron deficiency. The early identification and timely treatment of modifiable factors, like iron deficiency, can improve postoperative recovery. Furthermore, we broaden the red zones, based on the likelihood of causing patient harm, because in our opinion, e.g., a measurement of Hb 7.5 g/dL by the clinical standard and a reading of 4.5 g/dL by a test device results in a relevant clinical risk with potential patient harm.
For data visualization, we employed the ggplot2 package (v3.3.6) to generate box plots and scatter plots. Descriptive statistics, including the arithmetic mean, median, quartiles, and IQR, were calculated using the R function summary(). The Pearson correlation and linear regression analyses were performed using cor.test() and lm(), respectively. Bland–Altman scatter plots were designed to plot the mean of each measurement pair against the real error, with bias estimated as the average of these errors. Limits of agreement (LoAs) were calculated by applying an offset of twice the corrected sample standard deviation to the mean errors. Confidence intervals for bias and LoAs were calculated assuming a Student’s t-distribution using the R function qt().

3. Results

3.1. Overview of the Cohort

A total of 335 simultaneous hemoglobin measurements were included in our analysis. The cohort consisted of 59.4% male and 40.6% female patients, with a median age of 65.3 years (range: 19 to 98 y.o.). Of our cohort, 39 patients (11.6%) presented with severe anemia (hemoglobin ≤ 9 g/dL), while 177 (52.8%) showed mild anemia (hemoglobin < 13 g/dL) [26], and 119 (35.5%) had no anemia (hemoglobin ≥ 13 g/dL). The median perfusion index (PI) was 4.1 (IQR: [1.6; 7.4]), and the median peripheral oxygen saturation (SpO2) was 97.0% (IQR: [94; 98]), ranging from a minimum of 79% to a maximum of 100%. The majority of patients had sinus rhythm (n = 303), while 22 patients were diagnosed with atrial fibrillation, and 10 patients with unknown heart rhythm. Hypothermia, defined as a body temperature < 36 °C, was observed in 19 patients, and the temperature was unrecorded in 14 patients. No patients suffered from hyperthermia, defined as a body temperature > 38.5 °C. Measurements were collected both intraoperatively (n = 54) and in the post-anesthesia care unit (n = 281).

3.2. Measurement Accuracy

We evaluated the measurement accuracy of the RAD-67™ in comparison to the clinical standard blood gas analysis (BGA) using Bland–Altman analysis and a linear regression model. Among the 335 paired measurements, the difference of the hemoglobin readings was greater than +1 g/dL in 75 pairs, less than or equal to ±1 g/dL in 193 pairs, and greater than −1 g/dL in 67 pairs. Therefore, 57.6% of all analyzed measurements showed a good clinical accuracy with a difference less than ±1 g/dL.
The bias calculated from the Bland–Altman plot was 0.00 g/dL (95% CI: −0.14 to 0.15 g/dL), with limits of agreement (LoAs) ranging from −2.70 g/dL (95%CI: −2.95 to −2.44 g/dL) to 2.45 g/dL (95%CI: 2.45 to 2.95 g/dL) (Figure 1). The correlation between RAD-67™ and BGA was strong, with a Pearson correlation coefficient of r = 0.79. The linear regression analysis yielded a slope of 0.63 and a shift of 4.35 g/dL (Figure 2).
However, when focusing on patients suffering from severe anemia defined as Hb ≤ 9 g/dL (n = 42), only 33.3% of the paired measurements showed a good clinical accuracy with a difference less than ±1 g/dL. The bias in the anemic cohort was −1.2 g/dL (95% CI: −1.65 to −1.04 g/dL).
To assess the clinical relevance of the observed differences, we employed a modified Clark’s Error Grid, adjusting for hemoglobin thresholds as recommended by patient blood management (PBM) guidelines (Figure 3). Based on this modified grid, 85.4% of the paired measurements fell into the green category (results within a clinically acceptable range), 14.6% into the yellow category (results within a clinically significant error), and 0% into the red category (results outside a clinically acceptable range). When applying the standard Clark’s Error Grid, 92.2% of the values were classified as green and 7.8% as yellow.

3.3. Potential Confounders

Furthermore, we investigated the influence of different attributes, including patient age, perfusion index (PI), skin temperature, sex, and atrial fibrillation on the measurement accuracy of the RAD-67™ compared to the clinical standard (Figure 4 and Figure 5). None of the analyzed attributes had a clinically significant impact on the measurement accuracy. However, we observed a statistically significant impact of sex (p = 0.02). The mean absolute error was 0.3 g/dL in men and −0.1 g/dL in women.

4. Discussion

We investigated the measurement accuracy of the Masimo RAD-67™ for non-invasive hemoglobin (SpHb) monitoring in comparison to the established clinical standard of invasive blood gas analysis (BGA) in a perioperative setting. Our cohort comprised 335 consecutive Hb measurements. While our findings indicate that the RAD-67™ device demonstrates reasonable overall accuracy, clinical caution is warranted due to inaccuracies in patients with severe anemia.
The correlation coefficient between SpHb and BGA was r = 0.79, with a bias of 0.0 g/dL and limits of agreement (LoAs) ranging from −2.70 g/dL to 2.45 g/dL. Although this suggests a reasonable level of agreement, it is crucial to note that low hemoglobin values tended to be overestimated, while high values were underestimated by the RAD-67™. This systematic error, reflected by a slope of 0.63 and a shift of 4.35 in the linear regression model, represents a clinical risk, particularly as overestimation of low hemoglobin levels may lead to delayed or insufficient therapy in patients at risk of anemia-related complications. Consequently, SpHb measurement should not be used in patients with high perioperative blood loss or known severe anemia. Additionally, the use of SpHb in patients at high risk of anemia should be carefully considered for critically ill patients, given the high prevalence of anemia in intensive care settings and the potential risks of Hb overestimation.
Furthermore, we observed that none of the investigated attributes had a clinically relevant impact on the measurement accuracy of the RAD-67™. While sex showed a statistical significant effect (p = 0.02), as previously described in another study, we declared the impact as clinically irrelevant [27]. The PI, representing the ratio of pulsatile to non-pulsatile blood flow, has been previously described as a factor influencing measurement accuracy [28]. However, in our cohort, variations in PI did not affect the performance of the device. Similarly, no significant differences were observed in patients with atrial fibrillation, despite the manufacturer’s advice not to use the device in patients with cardiac arrhythmias. Although patient age does not significantly impact measurement accuracy, particular caution is advised when using the device in elderly patients due to their higher prevalence of anemia, a condition in which the device demonstrates reduced performance.
Comparison with previous studies reveals a heterogeneous body of evidence (Table 1) [14,16,18,20,25,27,29,30,31,32,33,34,35,36,37]. Studies evaluating SpHb measurements with Masimo devices in different patient cohorts and clinical settings have reported bias ranging from 0.0 g/dL (intraoperative) to 2.2 g/dL (obstetric anesthesia), and correlations between r = 0.55 (children aged 1–5 years) and r = 0.95 (intraoperative) [20,25,30,34]. Further contextualizing our results, 7.8% of our data points fell within the “yellow zone” of the standard Clark’s error grid, with no measurements falling into the “red zone,” indicating a better performance compared to previous reports in obstetric anesthesia with up to 15.6% of data points in the yellow zone [25]. Comparing our proportions of the absolute error with a previous study reveals that 57.6% vs. 68.6% of the measurements showed a difference ≤±1 g/dL between SpHb and BGA Hb [27]. However, they reported that SpHb was over-reported with >1 g/dL in 18.6% vs. 22.4% of the measurements in our analysis. Furthermore, it is essential to acknowledge that even Hb measurements obtained by BGA are not entirely free from bias when compared to the laboratory gold standard, with one study reporting a BGA bias of −0.38 g/dL [38].
The manufacturer explicitly states that SpHb monitoring should not replace laboratory-based hemoglobin measurements and that clinical decisions should be based on confirmation by gold standard blood sample analysis. Our findings reinforce this guidance, as 40.9% of the SpHb readings in our study fell outside the manufacturer’s stated accuracy range of ±1 g/dL for hemoglobin levels between 8 and 17 g/dL. These deviations are clinically significant and could impact decision-making, especially in cases wherein prompt and accurate hemoglobin monitoring is crucial.

Limitations

Our study has several limitations, especially due to its retrospective design. First, we cannot draw any conclusions about the device performance in repeated measurements, as only one measurement per patient was available. As a result, no conclusions can be drawn regarding the reproducibility of measurements within individual patients. Additionally, our patient data management system (PDMS) only documented successful measurements, so we have no information on failed measurement attempts, including the number of attempts or their reasons for failure. Furthermore, we exclusively included stable patients who were not receiving vasoactive drugs, preventing us from drawing conclusions about critically ill patients. The cohort is also sex-unbalanced, with more men than women, and no structured data regarding skin color was available. This limitation precludes any assessment of potential influences of skin pigmentation on the measurement accuracy.
Another limitation concerns the unknown measurement side in our cohort. According to the manufacturer, certain factors, such as wearing a blood pressure cuff on the same arm, can affect accuracy, but we were unable to evaluate this, as the measurement side was not documented. Finally, the manufacturer indicates that the device may be inaccurate in patients with tricuspid regurgitation, a disease we cannot evaluate in our cohort due to the lack of systematic documentation of pre-existing diseases in our PDMS.

5. Conclusions

In conclusion, while the RAD-67™ offers a non-invasive alternative for hemoglobin monitoring, its accuracy is variable, especially at low hemoglobin values. Clinical use should be accompanied by confirmatory testing using established methods, particularly when hemoglobin levels are critical for treatment decisions. However, the RAD-67™ serves as a valuable supplementary tool for orientation in patients with unknown hemoglobin values, such as during urgent operations without recent laboratory results. Future research should continue to explore the utility of non-invasive hemoglobin monitoring in specific patient populations and refine the technology to improve its clinical reliability.

Author Contributions

Study concept and design: A.S., P.K. and T.S.; data acquisition: P.H., A.S., S.H. and T.S.; data analysis: P.H. and M.S.; data interpretation: P.H. and M.S.; supervision: P.K. and P.M.; writing of the initial manuscript draft: P.H.; critical revision: A.S., S.H., T.S., M.S., P.M. and P.K. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the Interdisciplinary Center for Clinical Research (IZKF) at the University of Wuerzburg (IZKF-Clinician Scientist Program CSP-19).

Institutional Review Board Statement

The local ethics committee of the University of Würzburg in Germany approved the retrospective analysis (ref. no. 20240216-01) on 22 February 2024.

Informed Consent Statement

Patient consent was waived due to the retrospective design of the study, exclusively analyzing data from routine care.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request and for research purposes only. Data will not be shared for commercial use.

Acknowledgments

The Open Access Publication Fund of the University of Wuerzburg supported the publication. Generative AI tools, specifically “ChatGPT” (Versions 4.0) and “DeepL” were employed for linguistic revisions. It is essential to clarify that no content was created by the employed generative AI. These tools were exclusively utilized for linguistic refinement and revision including gramma and spelling checks. The final review of the manuscript was conducted by the authors. Moreover, the use of these tools adhered to established scientific guidelines, e.g., DFG-guideline regarding the ethical use of artificial intelligence in research, promoting transparency and compliance with established research practices.

Conflicts of Interest

A.S., S.H., T.S., and M.S. declare no conflicts of interest. P.H. received a research award from Vogel-Foundation and is a member of the Clinician Scientist Programme, Wuerzburg. P.M. received honoraria for scientific lectures from CSL Behring GmbH, Pharmacosmos GmbH, and CSL Vifor GmbH. P.K. received lecturing fees from TEVA, Sintetica, CSL Behring GmbH, Senzyme, Vifor Pharma GmbH, Pharmacosmos, and Grünenthal and consulted for TEVA, Milestone Scientific Inc, Sintetica, and Amicus Ltd. All mentioned funders and especially the manufacturers of the investigated devices had no role in the design of the study; collection, analyses, or interpretation of data; writing of the manuscript; or decision to publish the results.

References

  1. Null, N. Practice guidelines for perioperative blood management: An updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Management. Anesthesiology 2015, 122, 241–275. [Google Scholar] [CrossRef]
  2. Rumpf, F.; Hof, L.; Old, O.; Friederich, P.; Friedrich, J.; Thoma, J.; Wittmann, M.; Zacharowski, K.; Choorapoikayil, S.; Meybohm, P. Preoperative hemoglobin thresholds for survival equity in women and men. Front. Med. 2024, 11, 1334773. [Google Scholar] [CrossRef] [PubMed]
  3. Blum, L.V.; Schmitt, E.; Choorapoikayil, S.; Baumhove, O.; Bayer, A.; Friederich, P.; Friedrich, J.; Geisen, C.; Gruenewald, M.; Gutjahr, M.; et al. Association of anaemia, co-morbidities and red blood cell transfusion according to age groups: Multicentre sub-analysis of the German Patient Blood Management Network Registry. BJS Open 2022, 6, zrac128. [Google Scholar] [CrossRef] [PubMed]
  4. Smilowitz, N.R.; Ruetzler, K.; Berger, J.S. Perioperative bleeding and outcomes after noncardiac surgery. Am. Heart J. 2023, 260, 26–33. [Google Scholar] [CrossRef]
  5. Turan, A.; Rivas, E.; Devereaux, P.J.; Bravo, M.; Mao, G.; Cohen, B.; Maheshwari, K.; Pu, X.; Ruetzler, K.; Li, K.; et al. Association between postoperative haemoglobin concentrations and composite of non-fatal myocardial infarction and all-cause mortality in noncardiac surgical patients: Post hoc analysis of the POISE-2 trial. Br. J. Anaesth. 2021, 126, 87–93. [Google Scholar] [CrossRef]
  6. Rüsch, D.; Koch, T.; Spies, M.; Hj Eberhart, L. Pain During Venous Cannulation. Dtsch. Arztebl. Int. 2017, 114, 605–611. [Google Scholar] [CrossRef]
  7. Zhang, L.; Cao, S.; Marsh, N.; Ray-Barruel, G.; Flynn, J.; Larsen, E.; Rickard, C.M. Infection risks associated with peripheral vascular catheters. J. Infect. Prev. 2016, 17, 207–213. [Google Scholar] [CrossRef]
  8. Helmer, P.; Hottenrott, S.; Steinisch, A.; Röder, D.; Schubert, J.; Steigerwald, U.; Choorapoikayil, S.; Meybohm, P. Avoidable Blood Loss in Critical Care and Patient Blood Management: Scoping Review of Diagnostic Blood Loss. J. Clin. Med. 2022, 11, 320. [Google Scholar] [CrossRef]
  9. Althoff, F.C.; Neb, H.; Herrmann, E.; Trentino, K.M.; Vernich, L.; Füllenbach, C.; Freedman, J.; Waters, J.H.; Farmer, S.; Leahy, M.F.; et al. Multimodal Patient Blood Management Program Based on a Three-pillar Strategy: A Systematic Review and Meta-analysis. Ann. Surg. 2019, 269, 794–804. [Google Scholar] [CrossRef]
  10. Meybohm, P.; Richards, T.; Isbister, J.; Hofmann, A.; Shander, A.; Goodnough, L.T.; Muñoz, M.; Gombotz, H.; Weber, C.F.; Choorapoikayil, S.; et al. Patient Blood Management Bundles to Facilitate Implementation. Transfus. Med. Rev. 2017, 31, 62–71. [Google Scholar] [CrossRef]
  11. Cros, J.; Dalmay, F.; Yonnet, S.; Charpentier, M.; Tran-Van-Ho, J.; Renaudeau, F.; Drouet, A.; Guilbaut, P.; Marin, B.; Nathan, N. Continuous hemoglobin and plethysmography variability index monitoring can modify blood transfusion practice and is associated with lower mortality. J. Clin. Monit. Comput. 2020, 34, 683–691. [Google Scholar] [CrossRef] [PubMed]
  12. Ribed-Sánchez, B.; González-Gaya, C.; Varea-Díaz, S.; Corbacho-Fabregat, C.; Pérez-Oteyza, J.; Belda-Iniesta, C. Economic Analysis of the Reduction of Blood Transfusions during Surgical Procedures While Continuous Hemoglobin Monitoring Is Used. Sensors 2018, 18, 1367. [Google Scholar] [CrossRef]
  13. Saracoglu, A.; Abdullayev, R.; Sakar, M.; Sacak, B.; Girgin Incekoy, F.; Aykac, Z. Continuous hemoglobin measurement during frontal advancement operations can improve patient outcomes. J. Clin. Monit. Comput. 2022, 36, 1689–1695. [Google Scholar] [CrossRef]
  14. Tang, B.; Yu, X.; Xu, L.; Zhu, A.; Zhang, Y.; Huang, Y. Continuous noninvasive hemoglobin monitoring estimates timing for detecting anemia better than clinicians: A randomized controlled trial. BMC Anesthesiol. 2019, 19, 80. [Google Scholar] [CrossRef] [PubMed]
  15. Elsakka, A.; Hossam, W.; Helmy, G.; Helmy, N. Monitoring of plethysmography variability index and total hemoglobin levels during cesarean sections with antepartum hemorrhage for early detection of bleeding. Egypt. J. Anaesth. 2017, 33, 5–8. [Google Scholar] [CrossRef]
  16. Awada, W.N.; Mohmoued, M.F.; Radwan, T.M.; Hussien, G.Z.; Elkady, H.W. Continuous and noninvasive hemoglobin monitoring reduces red blood cell transfusion during neurosurgery: A prospective cohort study. J. Clin. Monit. Comput. 2015, 29, 733–740. [Google Scholar] [CrossRef]
  17. Nakamori, E.; Shigematsu, K.; Higashi, M.; Yamaura, K. Postoperative Noninvasive Hemoglobin Monitoring Is Useful to Prevent Unnoticed Postoperative Anemia and Inappropriate Blood Transfusion in Patients Undergoing Total Hip or Knee Arthroplasty: A Randomized Controlled Trial. Geriatr. Orthop. Surg. Rehabil. 2021, 12, 21514593211060575. [Google Scholar] [CrossRef]
  18. Martin, J.R.; Camp, C.L.; Stitz, A.; Young, E.Y.; Abdel, M.P.; Taunton, M.J.; Trousdale, R.T. Noninvasive Hemoglobin Monitoring: A Rapid, Reliable, and Cost-Effective Method Following Total Joint Replacement. J. Bone Jt. Surg. 2016, 98, 349–355. [Google Scholar] [CrossRef]
  19. Park, S.K.; Hur, C.; Kim, Y.W.; Yoo, S.; Lim, Y.J.; Kim, J.T. Noninvasive hemoglobin monitoring for maintaining hemoglobin concentration within the target range during major noncardiac surgery: A randomized controlled trial. J. Clin. Anesth. 2024, 93, 111326. [Google Scholar] [CrossRef]
  20. Arai, Y.; Shoji, H.; Awata, K.; Inage, E.; Ikuse, T.; Shimizu, T. Evaluation of the use of non-invasive hemoglobin measurement in early childhood. Pediatr. Res. 2023, 93, 1036–1040. [Google Scholar] [CrossRef]
  21. Allen, J. Photoplethysmography and its application in clinical physiological measurement. Physiol. Meas. 2007, 28, R1–R39. [Google Scholar] [CrossRef] [PubMed]
  22. Barker, S.J.; Wilson, W.C. Racial effects on Masimo pulse oximetry: A laboratory study. J. Clin. Monit. Comput. 2023, 37, 567–574. [Google Scholar] [CrossRef] [PubMed]
  23. Bland, J.M.; Altman, D.G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986, 1, 307–310. [Google Scholar] [CrossRef] [PubMed]
  24. Clarke, W.L.; Cox, D.; Gonder-Frederick, L.A.; Carter, W.; Pohl, S.L. Evaluating clinical accuracy of systems for self-monitoring of blood glucose. Diabetes Care 1987, 10, 622–628. [Google Scholar] [CrossRef]
  25. Mills, K.; Vermeer, J.M.; Berry, W.E.; Karreman, E.; Lett, C.D. Determining the validity of non-invasive spot-check hemoglobin co-oximetry testing to detect anemia in postpartum women at a tertiary care centre, A prospective cohort study. BMC Pregnancy Childbirth 2023, 23, 479. [Google Scholar] [CrossRef]
  26. Hands, K.; Daru, J.; Evans, C.; Kotze, A.; Lewis, C.; Narayan, S.; Richards, T.; Taylor, C.; Timmins, S.; Wilson, A. Identification and management of preoperative anaemia in adults: A British Society for Haematology Guideline update. Br. J. Haematol. 2024, 205, 88–99. [Google Scholar] [CrossRef]
  27. Ke, Y.H.; Hwang, K.Y.; Thin, T.N.; Sim, Y.E.; Abdullah, H.R. The usefulness of non-invasive co-oximetry haemoglobin measurement for screening pre-operative anaemia. Anaesthesia 2021, 76, 54–60. [Google Scholar] [CrossRef]
  28. Schneider, L.; Clark, A.; Bailey, M. Predictive value of perfusion index (PI) on the discrepancy between oxygen saturation measurement via pulse oximetry and blood gas analysis. J. Emerg. Crit. Care Med. 2024, 8, 21. [Google Scholar] [CrossRef]
  29. Hornedo-González, K.D.; Jacob, A.K.; Burt, J.M.; Higgins, A.A.; Engel, E.M.; Hanson, A.C.; Belch, L.; Kor, D.J.; Warner, M.A. Non-invasive hemoglobin estimation for preoperative anemia screening. Transfusion 2023, 63, 315–322. [Google Scholar] [CrossRef]
  30. Beleta, M.I.; Abdallah, S.R.; Hammad, Y.M.; Yahia, H.E.M.; Ali, A.A.E.; Mohamed, A.A.; AbdElhameed, B.M. Noninvasive continuous hemoglobin monitoring of blood transfusion in obstetric procedures. Egypt. J. Anaesth. 2022, 38, 701–708. [Google Scholar] [CrossRef]
  31. Al Aseri, Z.A.A.; Aldawood, B.A.; Altamimi, A.A.; Mosleh, H.I.; Qaw, A.S.; Albatran, H.I.; AlMasri, M.M.; AlRihan, T.M.; AlAithan, F.B.; AlAssaf, L.S. Accuracy of non-invasive hemoglobin level measurement in the emergency department. Am. J. Emerg. Med. 2023, 64, 200–203. [Google Scholar] [CrossRef] [PubMed]
  32. Bıcılıoğlu, Y.; Bal, A.; Demir Yenigürbüz, F.; Ergonul, E.; Geter, S.; Kazanasmaz, H.; Bal, U. Noninvasive Hemoglobin Measurement Reduce Invasive Procedures in Thalassemia Patients. Hemoglobin 2022, 46, 95–99. [Google Scholar] [CrossRef] [PubMed]
  33. Welker, E.; Novak, J.; Jelsma, L.; Koehler, T.; Davis, A.; DeCou, J.; Durkin, E. Continuous hemoglobin monitoring in pediatric trauma patients with solid organ injury. J. Pediatr. Surg. 2018, 53, 2055–2058. [Google Scholar] [CrossRef] [PubMed]
  34. Adel, A.; Awada, W.; Abdelhamid, B.; Omar, H.; Abd El Dayem, O.; Hasanin, A.; Rady, A. Accuracy and trending of non-invasive hemoglobin measurement during different volume and perfusion statuses. J. Clin. Monit. Comput. 2018, 32, 1025–1031. [Google Scholar] [CrossRef]
  35. Gamal, M.; Abdelhamid, B.; Zakaria, D.; Dayem, O.A.E.; Rady, A.; Fawzy, M.; Hasanin, A. Evaluation of Noninvasive Hemoglobin Monitoring in Trauma Patients with Low Hemoglobin Levels. Shock 2018, 49, 150–153. [Google Scholar] [CrossRef]
  36. García-Soler, P.; Camacho Alonso, J.M.; González-Gómez, J.M.; Milano-Manso, G. Noninvasive hemoglobin monitoring in critically ill pediatric patients at risk of bleeding. Med. Intensiv. 2017, 41, 209–215. [Google Scholar] [CrossRef]
  37. Caulfield, K.C.; McMahon, O.; Dennis, A.T. Baseline haemoglobin variability by measurement technique in pregnant people on the day of childbirth. Anaesthesia 2024, 79, 178–185. [Google Scholar] [CrossRef]
  38. Tanner, L.; Lindau, S.; Velten, M.; Schlesinger, T.; Wittmann, M.; Kranke, P.; Berg, K.; Piekarski, F.; Füllenbach, C.; Choorapoikayil, S.; et al. Factors influencing the bias between blood gas analysis versus central laboratory hemoglobin testing. A secondary analysis of a randomized controlled trial. PLoS ONE 2020, 15, e0240721. [Google Scholar] [CrossRef]
Diagnostics 15 00128 g001
Figure 1. Bland–Altman plot comparing Hb measurements by the Rad-67™ (Masimo) with the clinical standard method (BGA). Scatterplots show the real errors of the measurements (y axis: Rad-67™ measurements minus BGA reference) stratified by the mean of each measurement pair (x axis). Dashed horizontal lines mark the bias, i.e., the arithmetic average of all real errors with the limits of agreement (LoAs) as determined by an offset of ±2 times the standard deviation (SD). Error bars show the 95% confidence interval (CI) for the bias and both LoAs.
Figure 1. Bland–Altman plot comparing Hb measurements by the Rad-67™ (Masimo) with the clinical standard method (BGA). Scatterplots show the real errors of the measurements (y axis: Rad-67™ measurements minus BGA reference) stratified by the mean of each measurement pair (x axis). Dashed horizontal lines mark the bias, i.e., the arithmetic average of all real errors with the limits of agreement (LoAs) as determined by an offset of ±2 times the standard deviation (SD). Error bars show the 95% confidence interval (CI) for the bias and both LoAs.
Diagnostics 15 00128 g001
Diagnostics 15 00128 g002
Figure 2. Linear correlation assessment of the Hb measurements comparing the Rad-67™ to BGA. Scatterplots localize each of the paired measurements by the Hb reference BGA (x axis) and the corresponding Hb measurement of the benchmarked device (y axis). The black solid line depicts the linear regression model, with the 95% confidence interval shaded in gray. The black dotted line depicts the perfect regression.
Figure 2. Linear correlation assessment of the Hb measurements comparing the Rad-67™ to BGA. Scatterplots localize each of the paired measurements by the Hb reference BGA (x axis) and the corresponding Hb measurement of the benchmarked device (y axis). The black solid line depicts the linear regression model, with the 95% confidence interval shaded in gray. The black dotted line depicts the perfect regression.
Diagnostics 15 00128 g002
Diagnostics 15 00128 g003
Figure 3. Self-modified Clark’s error grid to quantify clinical accuracy of the Hb measurements. The presented plot is adapted to Hb threshold values of the PBM. The dashed lines show the original Clark’s error grid. On the x axis of the scatter plot, the absolute Hb values of the reference method are shown, and on the y axis, the investigated device. Values within the green zone are considered clinically acceptable, measurement pairs in the yellow zone indicate a clinically relevant deviation, and those in the red zone correspond to severe clinically relevant deviations.
Figure 3. Self-modified Clark’s error grid to quantify clinical accuracy of the Hb measurements. The presented plot is adapted to Hb threshold values of the PBM. The dashed lines show the original Clark’s error grid. On the x axis of the scatter plot, the absolute Hb values of the reference method are shown, and on the y axis, the investigated device. Values within the green zone are considered clinically acceptable, measurement pairs in the yellow zone indicate a clinically relevant deviation, and those in the red zone correspond to severe clinically relevant deviations.
Diagnostics 15 00128 g003
Diagnostics 15 00128 g004
Figure 4. Scatterplots showing the absolute difference between the reference method subtracted from the investigated device (y axis) in the context of different continuous variables (x axis) to assess potential influences. (A): age [y.o.], (B): PI (Perfusion Index), and (C): temperature [°C]. The linear correlation is shown as a solid line, and the baseline is shown as a dotted line.
Figure 4. Scatterplots showing the absolute difference between the reference method subtracted from the investigated device (y axis) in the context of different continuous variables (x axis) to assess potential influences. (A): age [y.o.], (B): PI (Perfusion Index), and (C): temperature [°C]. The linear correlation is shown as a solid line, and the baseline is shown as a dotted line.
Diagnostics 15 00128 g004
Diagnostics 15 00128 g005
Figure 5. Box plot visualizations of the absolute errors binned by categorical classifications of the patient attributes. (A): gender; (B): atrial fibrillation.
Figure 5. Box plot visualizations of the absolute errors binned by categorical classifications of the patient attributes. (A): gender; (B): atrial fibrillation.
Diagnostics 15 00128 g005
Table 1. Overview of studies investigating the measurement accuracy of Co-Oximetry by Masimo for SpHb. The first author, the year of publication, the employed device, the number of analyzed patients, the description of the study cohort, the bias with the corresponding limits of agreement (if available), and the correlation are given.
Table 1. Overview of studies investigating the measurement accuracy of Co-Oximetry by Masimo for SpHb. The first author, the year of publication, the employed device, the number of analyzed patients, the description of the study cohort, the bias with the corresponding limits of agreement (if available), and the correlation are given.
AuthorYearDeviceNumberCohortBiasCorrelation
Hornedo-González, K.D. [29]2023Masimo Pronto Pulse Co-Oximeter 112preoperative elective surgical patients0.8 g/dL (LoA −2.2; 3.9 g/dL)NA
Al Aseri, Z.A.A. [31]2023Masimo Rad-67™650emergency department0.146 g/dL (LoA −2.58; 2.87 g/dL)r = 0.812
Arai, Y. [20]2023Masimo Rad-67™102early childhood0.188 (LoA −1.61; 1.99 g/dL)r = 0.548
Caulfield, KC. [37]2023Masimo Rad-67™180neurosurgery patients−1.4 g/dL (LoA −3.59; 0.79 g/dL)NA
Mills, K. [25]2023Masimo Rad-67™301pregnant people on the day of childbirth2.2 g/dL (LoA 0.08; 4.30 g/dL)ICC 0.4
Beleta, M.I. [30]2022Masimo pulse co-oximetry60elective CS under general anesthesia0.348 g/dLr = 0.946
Bıcılıoğlu, Y. [32]2022-110patients aged 1–5 years0.3 g/dL r = 0.675
Ke, Y.H. [27]2021Masimo Rad-67™392screening pre-operative anaemia0.14 g/dL (LoA 2.24; −1.95 g/dL)r = 0.76
Tang, B. [14]2019Masimo Radical-7™28children with thalassemia−0.29 g/dL (LoA −2.30; 1.72 g/dL)r = 0.69
Welker, E. [33]2018Masimo Radical-7™21spine surgery on multiple spinal segments and cytoreductive surgery0.80 g/dL (LoA 3.94; −2.33)NA
Adel, A. [34]2018Masimo Radical-7™210pediatric trauma patients0.01 g/dL (LoA −1.33; 1.34 g/dL)r = 0.938
Gamal, M. [35]2018Masimo Radical-7™184primary or revision total hip or total knee arthroplasty0.12 g/dL (LoA −0.56; 0.79 g/dL)r = 0.872
García-Soler, P. [36]2017Masimo Radical-7™284trauma patients with low hemoglobin levels0.07 g/dL (LoA −2.26; 3.59 g/dL)r = 0.72
Martin, J. [18]2016Masimo Pronto Pulse Co-Oximeter100general anesthesia with complete muscle relaxation, adult−0,2 g/dL (LoA −3.4; 4.4 g/dL)CCC = 0.69
Awada, W.N. [16]2015Masimo Radical-7™83pediatric Intensive Care Unit0.0 g/dL (LoA −1.6; 1.5 g/dL)NA
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Helmer, P.; Steinisch, A.; Hottenrott, S.; Schlesinger, T.; Sammeth, M.; Meybohm, P.; Kranke, P. Evaluation of Non-Invasive Hemoglobin Monitoring in Perioperative Patients: A Retrospective Study of the Rad-67TM (Masimo). Diagnostics 2025, 15, 128. https://doi.org/10.3390/diagnostics15020128

AMA Style

Helmer P, Steinisch A, Hottenrott S, Schlesinger T, Sammeth M, Meybohm P, Kranke P. Evaluation of Non-Invasive Hemoglobin Monitoring in Perioperative Patients: A Retrospective Study of the Rad-67TM (Masimo). Diagnostics. 2025; 15(2):128. https://doi.org/10.3390/diagnostics15020128

Chicago/Turabian Style

Helmer, Philipp, Andreas Steinisch, Sebastian Hottenrott, Tobias Schlesinger, Michael Sammeth, Patrick Meybohm, and Peter Kranke. 2025. "Evaluation of Non-Invasive Hemoglobin Monitoring in Perioperative Patients: A Retrospective Study of the Rad-67TM (Masimo)" Diagnostics 15, no. 2: 128. https://doi.org/10.3390/diagnostics15020128

APA Style

Helmer, P., Steinisch, A., Hottenrott, S., Schlesinger, T., Sammeth, M., Meybohm, P., & Kranke, P. (2025). Evaluation of Non-Invasive Hemoglobin Monitoring in Perioperative Patients: A Retrospective Study of the Rad-67TM (Masimo). Diagnostics, 15(2), 128. https://doi.org/10.3390/diagnostics15020128

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop