Comparison between the HemoCue and an automated counter for measuring hemoglobin

Correspondence to: Adriana de A Paiva Departamento de Nutrição Faculdade de Saúde Pública USP Av. Dr. Arnaldo, 715 Cerqueira César 01246-904 São Paulo, SP, Brasil E-mail: aapaiva@yahoo.com.br Supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (Fapesp Process n. 98/15796-3). Received on 29/2/2003. Reviewed on 7/1/2004. Approved on 18/3/2004.


INTRODUCTION
Anemia is a Public Health problem worldwide, especially in pregnant women, because of their increased requirements for hematopoietic nutrients to achieve augmented erythropoiesis during pregnancy and for storage by the fetus.Hemoglobin (Hb) concentration is the most used parameter to detect anemia, usually determined by portable hemoglobinometers, such us the HemoCue  (Blood Hemoglobin Photometer -HemoCue AB, Angelhoim, Sweden).Some advantages of this equipment are simplicity and speed in measuring Hb, requiring just a single drop of blood collected in a special cuvette.It is claimed by the manufacturer that the HemoCue  is suitable for Hb estimation using capillary, venous and arterial blood samples.

www.fsp.usp.br/rsp
[4] The aims of this study were to determine the precision and agreement of hemoglobin measurements in capillary and venous blood samples by both HemoCue  and a counter in a group of pregnant women.

METHODS
Twenty-nine (29) pregnant women participated in the study and were enrolled when they came to one of the hospitals for delivery.All of them had to sign a consent form which included permission to collect blood samples.The Ethics Committee of the Faculdade de Saúde Pública approved the research protocol.
Paired capillary blood samples were collected by finger prick in the middle finger of left hand, using two of the HemoCue  specific cuvettes, after cleaning and massaging the finger to stimulate blood flow.The first and second blood drops were collected and immediately analyzed.A trained nurse performed sampling, according to the manufacturer's recommendation. 1nous blood samples were collected into Vacutainer  tubes with EDTAK 3 (Becton Dickinson Vacutainer  System), by the same nurse.The tubes were sent to the medical laboratory of the Hospital São Vicente de Paula to measure Hb in duplicate on a counter Cell Dyn 3000.Hb was also measured in venous blood by the HemoCue  .
Since the volume of capillary blood samples was small, it was not possible to measure Hb in capillary blood by the Cell Dyn 3000.
The precision of the two equipments was assessed using the coefficient of repeatability (CR) and coefficient of variation (CV).Paired Student's t-test was used to compare the means of Hb obtained by both HemoCue  and counter.The intra-class correlation coefficient (r icc ) was determined to evaluate the agree-ment between the Hb measurements obtained by the two equipments.To evaluate the agreement between the diagnoses of anemia obtained by the HemoCue  and by the Cell Dyn 3000, Kappa's statistics was used.The Hb cutoff point to classify pregnant women as anemic was <11 g/dL.

RESULTS
The results showed that HemoCue  has a lower precision than Cell Dyn 3000, for analyses of both biological samples since because of the low repeatability (CR -Cell Dyn 3000 venous blood =0.08;HemoCue  venous blood =0.53;HemoCue  capillary blood =0.54) and a high variability (CV -Cell Dyn 3000 venous blood = 11.8;HemoCue  venous blood =17.0;HemoCue  capillary blood =13.6).Paired Hb measurements presented a higher variation when performed in venous blood, compared with capillary blood.
The Table shows the means and the r icc for the Hb measurements in capillary and venous blood by the HemoCue  and in venous blood by the counter Cell Dyn 3000.There was no significant difference between the Hb measurements determined in venous blood by the HemoCue  and by the counter (p=0.768).However, the r icc value was not so high, although significant (r icc =0.60; p<0.01).
The mean value of Hb obtained in capillary blood by the HemoCue  was higher than the mean value of Hb obtained in venous blood by both HemoCue  (p=0.021) and Cell Dyn (p=0.002).However, the r icc was high between the Hb measurements determined by the HemoCue  in capillary blood and by the counter (r icc =0.86; p<0.01).
The prevalences of anemia in pregnant women obtained by the HemoCue  using capillary blood (24%) and by the Cell Dyn 3000 (24%) were similar, and there was a good agreement between the diagnoses (k=0.81;p<0.01).quick assessment of Hb concentrations.In this study, it was analyzed the precision and agreement of the HemoCue  , a portable hemoglobinometer widely used for analysis of both capillary and venous blood samples.

Portable photometers have recently been used for
The Hb measurement performed by HemoCue  is based on a method described by Vanzetti. 6The sodium desoxicholate present in the cuvette hemolyze red cells, and the free Hb is converted into methemoglobin by sodium nitrite.Then, methemoglobin is converted into azide-methemoglobin by sodium azide.Transmitted light absorbance is measured at 565 and 880 nm to compensate for any turbidity in the sample.
According to the study results, the HemoCue  has a lower precision than the Cell Dyn 3000 for analyses of both blood samples.Similarly, some studies have also found a high variation between paired Hb measurements obtained by the HemoCue  when compared to a counter.
Chen et al, 2 studying forty-two (42) intensive care unit patients, compared the Hb measurements obtained in capillary, arterial, and venous blood samples, analyzed by the HemoCue  and an automated counter.According to the authors, the HemoCue  was found to give repeatable results when Hb estimates were made using either venous or arterial samples, but capillary estimates were found to be significantly less repeatable.
Conway et al, 4 investigating a group of forty (40) health adults, observed a wide variation between paired Hb measurements determined in capillary samples by the HemoCue  .The differences were higher than 10.0 g/L (1.0 mg/dL), comparing the measure-ments conducted by health visitors with the ones by biomedical scientists, who received identical training on the use of the HemoCue  .
According to Cogswell et al, 3 the collection of capillary samples for the diagnosis of anemia and iron deficiency anemia is not appropriate, rather than an inadequacy of the HemoCue  .It is important to point out that the measurements in capillary blood samples have been reported to be significantly higher than the corresponding values in venous samples. 5though the Hb measurements obtained by HemoCue  using capillary samples were higher than the ones obtained by the counter, the correlation between these measurements was good (r icc =0.86; p<0.01).In addition, the prevalence of anemic pregnant women was similar using both equipments (24%), and there was a good agreement in the diagnoses (k=0.81;p<0.01).
These results suggest that the HemoCue  show a higher variability between measurements in duplicate, especially using capillary blood samples.The Hb measurements obtained by the HemoCue  in capillary blood were always higher than those obtained in venous blood, regardless of the equipment.Nevertheless, the Hb measured by the HemoCue  in capillary blood showed a good correlation with the Hb measurements by the counter.
Undoubtedly, the HemoCue Ò is useful in both clinical and epidemiological settings, especially in situations of emergency.However, it is more appropriate to use capillary blood samples, considering the sampling error.In addition, it is necessary to have a thorough training to minimize poor repeatability of this equipment.

Table -
Comparison between capillary and venous hemoglobin determinations by HemoCue  and Cell Dyn 3000.