Accuracy of a portable hemoglobinometer (HemoCue) to measure fetal hemoglobin values during in utero transfusion

Abstract

Objectives

The current recommended treatment for severe fetal anemia is in utero transfusion (IUT). During this procedure, the evaluation of the necessary volume of transfused blood is based on regular measurement of fetal hemoglobin (FHb) concentration. The gold standard measurement is performed in the biology laboratory. A rapid medical test such as HemoCue^® is an effective way to predict FHb concentration. It would reduce the time to obtain results and therefore the procedure duration. To evaluate the accuracy of HemoCue^® to measure FHb during IUT, we compared Hb levels obtained by HemoCue^® and by our biology laboratory.

Methods

This retrospective study involved all pregnant women who had undergone an IUT in the university hospital of Clermont-Ferrand, France, during the period from 1 January 2010 to 6 June 2021. The FHb level was evaluated by two methods, a rapid medical test, HemoCue^®, and a standard method in the biology laboratory.

Results

We obtained 244 pairs of results from HemoCue^® and our laboratory, of 90 IUT procedures. The correlation between the two sets of results was excellent, with Lin’s concordance correlation coefficient of 0.979. However, we established that the measurements were not significantly modified by IUT number, puncture time, cause of fetal anemia, estimated fetal weight, gestational age, and delay between two IUT or middle cerebral artery peak systolic velocity values.

Conclusion

Our results allowed to extend the relevance of FHb measurements by HemoCue^® during IUT.

Keywords:

• Fetal transfusion
• HemoCue
• hemoglobinmeter
• fetal hemoglobin
• fetal anemia

Introduction

Fetal anemia is a potentially fatal complication and in utero treatment can avoid fetal death and adverse neonatal consequences [1–3]. Severe fetal anemia is defined as fetal hemoglobin (FHb) less than 5 g/dL, diagnosed by cordocentesis. The most common cause is maternal red blood cell alloimmunization, anti-D (RH1), and anti-Kell [4, 5]. The suspicion of severe fetal anemia was performed by a fetal middle cerebral artery peak systolic velocity (MCA-PSV) >1.5 MoM, evaluated by Doppler measurement [6–9].

The current treatment of severe fetal anemia is in utero transfusion (IUT), commonly performed in the umbilical cord under ultrasound guidance [10–13]. When intravascular transfusion is technically challenging, intraperitoneal fetal transfusion can be performed. The volume of blood transfused is evaluated from the FHb or hematocrit rate during the procedure. This measurement can be performed by a rapid medical test such as HemoCue^® (Ängelholm, Sweden). Such tests can be an effective replacement for traditional laboratory analyses, the current gold standard. However, few studies have been conducted on the accuracy of HemoCue^® in measuring FHb concentration [14–16].

The gold standard Hb measurement test is the complete blood count done on a hematology analyzer and typically requires a venous blood draw, trained laboratory technicians, and expensive analytical equipment and reagents [17]. To evaluate non-FHb value in critical care areas in health facilities, the HemoCue^® hemoglobin photometer has been widely used. HemoCue^® devices commonly reported high accuracy and precision when compared with hematology analyzers; however, in field settings, the HemoCue^® device has shown a greater bias and higher variability [18–20].

The primary objective of this study was to determine the feasibility of HemoCue^® to measure FHb levels during an IUT procedure by comparing the results of the test with those of a traditional complete blood count in a biology laboratory. The secondary objective was to study the factors influencing FHb measurement by HemoCue^®.

Material and methods

This retrospective study involved all pregnant women who had undergone an IUT in the university hospital of Clermont-Ferrand, France, during the period from 1 January 2010 to 6 June 2021.

The main data collected were the cause of fetal anemia, number of fetuses, number of procedures per pregnancy, gestational age, MCA-PSV values before IUT, estimated fetal weight (EFW), pairs of FHb values provided by HemoCue^® and by our laboratory for each sample. Crossway^®, ICOS^®, and Viewpoint^® medical software were used to collect the data.

The FHb level was evaluated by two methods, a rapid medical test, HemoCue^®, and a standard method in the biology laboratory. Two models of HemoCue^® were used, HemoCue Hb 201 DM^® and HemoCue Hb 201+^®. The reproducibility of the results from one machine to another had been studied beforehand. The laboratory used Sysmex-XN10^® automated systems.

All IUT procedures were performed under ultrasound guidance by a fetal medicine specialist, in the umbilical cord in most cases. Before transfusion, a cord blood test was performed with an etiological assessment, and FHb measurement by both methods. During the IUT, the FHb measurements were not performed on the first collection syringe to avoid measuring the Hb level of the transfused adult blood. IUT was maintained until a target Hb level between 14 and 16 g/dL was reached.

The study was approved by the local Ethics Committee (IRB00013412, “CHU de Clermont-Ferrand IRB #1”, IRB number 2022-CF009) in compliance with the French policy of individual data protection.

Calculation

Statistical analyses were performed with Stata^® software (version 15, StataCorp, College Station, TX, USA). Categorical data are expressed as numbers and percentages. Quantitative data are expressed as means and standard deviations. Normality was examined by the Shapiro–Wilk test. Concordance was assessed by Lin’s concordance coefficient for quantitative variables and by Kappa concordance coefficient test for categorical variables. Sensitivity, specificity, and negative and positive predictive values were expressed with a 95% confidence interval, and the hematology analyzer was considered as a gold standard Hb measurement test. A difference in FHb levels between the two methods of less than 1 g/dL was considered acceptable. Variations in the results obtained by the two methods were studied by analysis of variance (ANOVA) or the Kruskal–Wallis test when the assumptions for ANOVA were not met. Homoscedasticity was studied by Bartlett’s test. Relationships between quantitative variables were studied by Spearman’s correlation coefficient. All statistical tests were two-sided with a type one error set at 5%.

Results

Twenty-eight patients were included, and two had two pregnancies with IUT. Two others had a twin pregnancy, of whom one had IUT for both twins and the other only one fetus transfused, for a twin anemia polycythemia syndrome. The study involved 31 fetuses. The demographic data are given in Table 1.

Table 1. Demographic data.

Pregnant (n)	n = 28
Singleton (n (%))	n = 26 (92.9%)
Maternal age (years; mean ± SD)	30.7 ± 6
Parity (n)
Nulliparity (n (%))	n = 24 (85.7%)
Gestational age for the first IUT (WG; mean ± SD)	28.7 ± 4.3
EFW (grams; mean ± SD)	1244.03 ± 632.41
EFW at the first IUT (grams; mean ± SD)	974.34 ± 554.79
Fetal anemia etiology
Maternal immunization (n (%))	n = 23 (82.1%)
Rhésus (n (%))	n = 17 (60.7%)
Kell (n (%))	n = 6 (21.4%)
Parvovirus B19 infection (n (%))	n = 3 (10.7%)
Materno-fetal hemorrhage (n (%))	n = 1 (3.6%)
TTTS and/or TAPS (n (%))	n = 1 (3.6%)
Alpha-thalassemia (n (%))	n = 1 (3.6%)
Essential anemia (n (%))	n = 1 (7.1%)
Average time between IUT (days; mean ± SD)	18.4 ± 7.54
Number of IUT (mean ± SD)	2.6 ± 1.9

EFW: estimated fetal weight; IUT: in utero transfusion; SD: standard deviation; TAPS: twin anemia polycythemia syndrome; TTTS: twin-to-twin transfusion syndrome; WG: weeks of gestation.

A total of 310 samples were taken (1–9 per IUT) from over 90 different procedures (1–7 per pregnancy). Analysis was made of 244 pairs of FHb values obtained by the two methods. No significant statistical difference was found between FHb values within the pairs. Agreement was very good, with Lin’s concordance coefficient of 0.979 (Figures 1–3).

Figure 1. Comparison graphic of hemoglobin results due to biological laboratory and HemoCue^®.

hb_hemocue: hemoglobin results due to HemoCue^®; hb_labo: hemoglobin results due to biological laboratory.

Figure 2. Bland and Altman representation of the differences between hemoglobin levels obtained by HemoCue^® and in the laboratory.

hb_hemocue: hemoglobin results due to HemoCue^®; hb_labo: hemoglobin results due to biological laboratory.

Figure 3. Distribution of the results according to the deviations obtained by comparing the hemoglobin values in g/dL obtained by the HemoCue^® and the laboratory.

Hemoglobin levels were divided into four categories (<5, [5–15], ≥5 g/dL). The correlation percentages are reported in Table 2. The Kappa concordance coefficient was 0.87, or 90.2% agreement. The results of Lin’s concordance coefficient for quantitative variables and by Kappa concordance coefficient test for categorical variables in each category are shown in Table 3, as so sensitivities, specificities, and negative predictive values to compare the two FHb measurement methods with consideration as Coulter Counter^® device results were the gold standard. The overall Kappa concordance coefficient was 0.87, or 90.2% agreement.

Table 2. Correlation of the hemoglobin level, divided into four categories, obtained with HemoCue^® (hb_hemocue) compared to laboratory results (hb_labo).

		hb_labo
	g/dL	<5	[5–10	[10–15]	≥15
hb_hemocue	<5	92.86	1.35	0.00	0.00
	[5–10]	7.14	93.24	4.55	0.00
	[10–15]	0.00	5.41	92.42	25.00
	≥15	0.00	0.00	3.03	75.00

hb_hemocue: hemoglobin results due to HemoCue^®.

Table 3. Sensitivity, specificity, positive predictive value, and negative predictive value of results obtained with HemoCue^® compared to laboratory results, the hematology analyzer considered as a gold standard Hb measurement test.

g/dL	<5	[5–10]	[10–15]	≥15
Number of laboratory results	16	75	133	20
Lin’s concordance coefficient	0.96	0.91	0.88	0.83
Kappa concordance coefficient	0.92	0.89	0.90	0.75
Sensitivity [95%CI]	86.7 [59.5–98.3]	94.7 [87.1–98.5]	93.4 [87.9–97.0]	82.4 [56.6–96.2]
Specificity [95%CI]	99.1 [96.9–99.9]	95.2 [90.8–97.9]	94.4 [88.2–97.9]	98.7 [96.2–99.7]
PPV [95%CI]	86.7 [59.5–98.3]	90.0 [81.2–95.3]	95.5 [90.5–98.3]	82.4 [56.6–96.2]
NPV [95%CI]	99.1 [96.9–99.9]	97.6 [93.9–99.3]	91.8 [85.0–96.2]	98.7 [96.2–99.7]

CI: confidence interval; NPV: negative predictive value; PPV: positive predictive value.

The difference between the FHb values within the pairs did not vary significantly according to IUT number (p = .67), time of puncture (p = .47), cause of fetal anemia (p = .06), EFW (p = .32), gestational age (p = .21), time between IUTs (p = .13), or MCA-PSV value (p = .73).

Discussion

Our results showed excellent agreement between the Hb values obtained by HemoCue^® and our laboratory and no particular circumstances could be found that significantly altered the concordance of the values. They were also consistent with those in the literature. Laifer et al. found no significant difference in FHb concentrations obtained by HemoCue^® and the Coulter Counter^® device, on 39 fetuses [14]. The HemoCue^® accuracy was not significantly affected by the ratio of fetal to adult Hb, nor by extreme values of FHb (but only two fetuses had FHb less than 8 g/dL), nor by extreme gestational age in the study of Berry et al. on 58 fetal samples. None of the values obtained were from a fetus that had previously undergone IUT [15].

The limitations of our study include its retrospective nature, which, as an intrinsic consequence, means a lack of control over the procedure course and the protocol follow-up. In addition, numerous data, such as the presence of fetal movements or fetal blood coagulation, are missing.

Conclusion

Our results confirm those of the literature and allowed to extend the relevance of FHb measurements by HemoCue^® during IUT irrespective of the number of IUTs, time between IUTs, gestational age, EFW, and cause of anemia. Thus, HemoCue^® could be an alternative to Coulter laboratory devices to obtain rapid results and we can reasonably rely on HemoCue^® data collected during IUT procedures without waiting for the laboratory results, a delay that can lead to an increase in the duration of the procedure and hence an increase in the associated risks.

Author contributions

Amélie Delabaere: lead author, fetal medicine specialist to perform IUT. Maeva Guerard: coauthor, creation of the database, carrying out statistical analyses. Romain Cahierc: fetal medicine specialist to perform IUT. Bruno Pereira: bio-statistician. Damien Bouvier: biologist contributing to the development of techniques. Gallot Denis: fetal medicine specialist to perform IUT.

Disclosure statement

The authors report there are no competing interests to declare.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.