Non-invasive Hemoglobin Measurement (NIHM) devices have recently become available in clinical practice. These point-of-care devices use spectrophotometry to measure hemoglobin (Hb), providing real-time results and do not require blood draws. However, the validity of this technology has not been examined in patients with moderate to severe anemia (Hb< 90g/L).
The purpose of this study is to examine the validity and performance of non-invasive hemoglobin measurements (NIHM) among anemic patients in the real world setting.
Patients were recruited from two clinical settings that are highly applicable this technology.
(i) The intensive care setting, where frequent blood draws contribute to anemia.
Patients were recruited from three tertiary care critical care units
(ii) A developing country setting. Patients were recruited from the Internal Medicine and Paediatrics wards at Mulago hospital in Kampala, Uganda.
An additional cohort of patients was recruited from an outpatient Hematology clinic which served as a comparator group.
All patients were enrolled in a consecutive manner. In the ICU population, patients were recruited through a consecutive, convenience sample in order to maximize patients with hemoglobin values < 90 g/L.
The Pronto 7® NIHM device was compared to Coulter Counter Complete Blood Count (CBC) measurements. All CBC were processed at College of American Pathologist or Ontario Laboratory Accreditation approved labs; all CBC determinations were drawn within 4 hours of the Pronto determination in patients with stable hemoglobin values. Basic demographics along with suspected etiology of anemia were also collected. The consistency between CBC and NIHM was assessed by Pearson's Correlation, mean difference and linear regression. The study was reviewed by the Integrated REB at both St Joseph's Hospital, Hamilton and Mulago Hospital and Makerere University in Uganda. As a minimal risk study, the requirement for consent was waived at the Hamilton sites by the research ethics board. Consent was obtained from Ugandan participants.
188 patients were recruited.
a. ICU population. Seventy four patients were recruited with mean age of 61 years (range 16-89). The patients were of a medical (n=44; 69%), surgical (n=17; 23%), trauma (n=7; 18%) mixed population (n=6; 13%). The mean Hb on CBC was 97 g/L. Forty patients (53%) had an Hb less than 90 g/L. The mean absolute difference between the CBC and NIHM was 14.85 (range 0-53g/L). The overall Pearson correlation coefficient was 0.62 and the R2 value was 0.37.
b. Ugandan population. Eighty-three patients were recruited (31 pediatric and 52 adult) with a mean age of 27 (range 2-80). The mean Hb value on CBC was 80 g/L and 43% percent of patients had Hb values than 70 g/L. Seventeen percent of patients had sickle cell disease. NIHM were not obtainable in over half (n=55 (65%) of the study population) and in 30 of the 31 pediatric patients. In those who could be tested the mean absolute difference between the CBC and NIHM was 24 g/L (range 0-72) with Pearson coefficient of 0.84 and R2 value of 0.71.
c. Outpatient population. Thirty one patients were recruited with mean age of 66. The mean Hb value was 119 g/L with mean absolute difference between the CBC and NIHM of 14.6 g/L (range 1-43 g/L) with Pearson correlation coefficient of 0.69 and R2 of 0.47.
Figure 1 shows the scatter plot for the three patient groups. In all three groups, the majority of data points fall to the right of the line of equal value between NIHM and CBC measurements, indicating a tendency of the device towards overestimation of the hemoglobin.
This is the first study of NIHM to target the patients with moderate to severe anemia and to include pediatric patients and those with hemoglobinopathy. There is only moderate degree of correlation between the NIHM and CBC measurements with the NIHM demonstrating a consistent positive bias. The positive bias was seen in all patient groups, and both the inpatient and outpatient setting. The device was unable to provide measurements in children though this may have been due the probe size available to the investigators. The discrepancy in values between NIHM and venipuncture CBC values suggests that the non-invasive hemoglobin monitor will be of limited utility in the patient groups evaluated in this study.
Crowther:Asahi Kasai: Membership on an entity’s Board of Directors or advisory committees; Baxter: Membership on an entity’s Board of Directors or advisory committees, Speakers Bureau; Boehringer Ingelheim: Consultancy, Honoraria, Membership on an entity’s Board of Directors or advisory committees; CSL Behring: Speakers Bureau; Leo Pharma: Consultancy, Honoraria, Membership on an entity’s Board of Directors or advisory committees, Research Funding, Speakers Bureau; Merck: Consultancy; Octapharma: Consultancy, Membership on an entity’s Board of Directors or advisory committees; Pfizer: Consultancy, Honoraria, Research Funding; Sanofi-Aventis: Consultancy, Honoraria, Membership on an entity’s Board of Directors or advisory committees, Research Funding; Viropharma: Membership on an entity’s Board of Directors or advisory committees.
Asterisk with author names denotes non-ASH members.