White blood cell count in birds: evaluation of a commercially available method

Hematological analysis is an essential diagnostic part of the clinical management of avian patients to evaluate their health, clinical disease progression, and response to therapy [1]. To analyze avian blood samples, standard automated cell counting used in mammalians is unreliable because all avian blood cells are nucleated. In impedance-based hematological instruments, the nuclei of the avian erythrocytes interfere with those of white blood cells after lysis of the cells, or the lysing solutions are unable to lyse the avian blood cells adequately [2]. Therefore, quantitative white blood cell counting in birds is still performed manually with a hemocytometer.

First activities in counting white blood cells in avian blood began in 1906 by Warthin [3]. After several attempts by other authors to optimize dyes and counting procedures, Natt and Herrick refined the protocol and developed an indirect method to count avian white blood cells using a stable buffered saline solution containing methyl violet 2B, a stain commonly referred to as Natt-Herrick solution; this method has the advantages using a single dye for all cells that is stabile for a long period of two years, and the shape of the cells is retained after staining [3]. A disadvantage of Natt-Herrick solution is the difficulty to differentiate thrombocytes from lymphocytes, thus creating significant counting errors. Therefore, in the author’s Laboratory a modification of the original Natt-Herrick protocol for avian WBC counting is in use for the last 25 years, which includes the counting of only the granulocyte population in the hemocytometer, followed by a manual blood smear differential of 200 WBC and adding the lymphocyte and monocyte counts to the granulocyte counts. With this technique, avian WBC count is easier to perform and much more reliable.

Another very common quantitative counting method was the Unopette method [2, 4]. This method was based on the principle that avian eosinophils and heterophils stain with phloxine B, a red dye. However, the Eosinophil Unopette 5877 stain kit from Becton Dickinson, Rutherford, New Jersey that used this dye was removed from the market in 2007. As a replacement method, the Leukopet kit from Vetlab (Florida, USA) [5] is now available.

Estimation of white blood cell numbers on a stained blood smear is a technique available to most practitioners because it requires no special equipment. Thus, the average number of white blood cells evaluated in at least 10 high power fields (40 x objective) is multiplied by 2000 to obtain an estimated total white blood cell count per microliter [6]. This technique is less precise than a hemocytometer-derived white blood cell count [7].

All manual methods including the Natt-Herrick method are time consuming and require highly trained technicians [7].

To overcome some of the inherent problems of avian blood analysis and to facilitate avian white blood cell counting in a clinical setting or private practice, a commercially available, prefilled and single test system for avian white blood cell counting, the Natt-Herricks-Tic® (Bioanalytic GmbH, Umkirch, Germany) has been developed, which is evaluated in the present study. In this study, agreement between the Natt-Herricks-Tic® and a laboratory reference method (Natt-Herrick solution) was determined. Furthermore, the within-run precision for the test and the reference method were assessed, and ultimately, the practicability of the Natt-Herricks-Tic® was judged.

Total leukocyte counts determined by the reference method ranged from 4167 cells/μL to 37,121 cells/μL. The results of the method agreement study are presented in Figs. 1 and 2. Correlation between the white blood cell counts determined with the Natt-Herrick Tic® test method and the Natt-Herrick reference method revealed a correlation coefficient of 0.95. Regression analysis by Passing-Bablok method showed very small systematic error between the test method and the reference method: intercept, 412 WBC/μL (95% confidence interval, 1.596 to 863); slope, 1.03 WBC/μL (95% confidence interval, 0.90 to 1.17) (Fig. 1). The Bland-Altman difference plot showed a very small positive bias due to systematic error (mean bias, 272 WBC/μL) but with wide limits of agreement (− 4682 to 5226) (Fig. 2), indicating random error. The Natt-Herrick Tic® test method slightly overestimated WBC counts compared to the reference method. The mean values and CV from the within-run precision study are presented in Table 1. Given that both test and reference methods are manual techniques, these CVs can be judged as good. Two samples out of 40 (5%) would have led to false clinical interpretations (Table 2). One sample from a gamefowl was falsely positive (leukopenia instead of a normal leukocyte count) with a TLC of 14,506 cells/μL with the reference method and 6035 cells/μL with the test method. A second sample from a buzzard was falsely negative (normal leukocyte count instead of leukocytosis) with a TLC of 10,823 cells/μL with the reference method and 9755 cells/μL with the test method.

	Natt-Herrick Tic® method	Reference method
Mean (× 103/μL)	9.7	7.9
Standard deviation (× 103/μL)	1.54	1.8
Coefficient of variation (%)	16	23

Avian order	Number of samples	Leukopenia	Normal leukocyte count	Leukocytosis	False interpretation
Psittaciformes	17	2	12	3	0
Galliformes	9	4	4	1	1 (false positive)
Accipitriformes	7	0	2	5	1 (false negative)
Anseriformes	5	2	1	1	0
Passeriformes	0	0	0	0	0
Piciformes	1	0	0	1	0
Ciconiiformes	1	0	1	0	0
Total samples	40	0	1	1	2

The Natt-Herricks-Tic® test method was easy to learn and to perform. It also performed well in terms of staining quality and countability of the granulocytes with a hemocytometer. In a few cases, recognition of the white blood cells was easier with the reference method than with the test method. However, cell counting with the hemocytometer with the 10x objective as indicated by the manufacturer is not recommended. Instead, the 40x objective was used for the agreement and precision studies for the test and reference methods.

For the first time, the present study evaluated a new commercially available single test system for avian white blood cell counting, the Natt-Herricks-Tic®. According to the guidelines for manual hematology of nonmammalian species published by the American Society of Veterinary Clinical Pathology (ASVCP), prior to introducing a new method as a diagnostic tool, a method validation experiment should be performed [8]. This validation procedure should be modified from standard mammalian method validation procedures to be applicable to exotic manual hematology. Thus, it should be ensured that the new method is functioning satisfactorily to meet the laboratory’s requirements and the manufacturer’s specifications [8]. To follow the ASVCP requirements, a method agreement study was performed between the test method and a reference method used for thirty years in the authors’ laboratory with a reasonable number of avian blood samples of various avian species to identify systematic errors. The results indicated that the Natt-Herricks-Tic® test method slightly overestimated white blood cell counts compared to the reference method, with a very small mean bias. However, the limits of agreement were wide, indicating that discrepancies between the measurements were due to random error. Underlying causes for this random error are most likely different staining properties of the solutions, which affected cell identification. Furthermore, the protocols for the reference and test methods showed differences in blood dilutions, and different numbers of squares in the hemocytometer were counted. Moreover, inter-observer differences most likely influenced the counting results of the present study and contributed to the wide limits of agreement. According to ASVCP guidelines, an important analytical factor is the experience and knowledge of the laboratory analysts for identifying nonmammalian white blood cells with a hemocytometer and on a blood smear [8]. In the present study, the reference method was performed by trained technicians with several years of experience in manual chamber counting and blood smear evaluation of avian blood cells, whereas the test method was performed by a Master’s student who was trained intensively in avian hematology prior to the onset of this study.

Overall, this agreement study clearly showed that bias between the two methods due to systematic error was not of significance. More importantly, random error needs to be considered a disadvantage of analytical performance. In contrast, another very small study of avian white blood cell counting comparing Natt-Herrick solution and the Unopette technique showed considerable differences between the two methods (Dein et al., 1994).

Nearly eight decades ago, manual methods used for blood cell counts in birds and reptiles were reported to be imprecise [9]. A more recent study confirmed large variations in manual counts without reporting CV [10]. For manual WBC counts, CVs were reported to range between 20 and 40% [10, 11]. In another study using Natt and Herrick solution, CVs were reported between 5.5 and 17.2% (Dein, 1994). The within-assay precision determined in the present study revealed that the CVs for the reference method and the test method were comparable to those reported in previous studies. It is very likely that the high degree of analytical imprecision contributed to the wide limits of agreement observed in the comparison study. Therefore, it is crucial to apply non-statistical quality control in daily routines by performing hemocytometer counting in duplicate, evaluating blood smears for plausibility and reducing inter-observer bias by frequent internal quality control audits of the staff.

Reference intervals are of key value in result interpretation, but interval generation is fraught with difficulties, especially concerning sample size in avian species [12]. Recommendations from the ASVCP provide allowances for sample sets of 20 or 40 samples in addition to the preferred standard of n ≥ 120 used to calculate reference intervals [13]. However, most published intervals for avian species do not meet the ASVCP criteria [12]; they are extremely wide and vary among species. To the best of our knowledge, lab-specific reference intervals for avian species are unavailable, and clinicians and exotics experts in zoos usually work with individual reference intervals. When interpreting the Natt-Herricks-tic® test methods results using published reference intervals for avian species, only two of 40 samples would have resulted in different clinical interpretations. Therefore, it can be concluded that the observed variations between the two methods within the present study are of minor importance, as clinical interpretation would have led to the same outcome in 95% of the cases. Clearly, this is not an assessment of diagnostic performance addressing the test methods ability to discriminate diseased and non-diseased birds. Rather, it confirms the good agreement of the test method with the reference method, considering that the diagnostic value of using published hematological reference intervals for avian patients remains questionable.

Advantageously, the test kit contains most of the consumables as prefilled sample vials filled with Natt-Herrick solution with long-term stability (2 years), end-to-end-volume capillary tubes (5 μL), and chamber filling capillary tubes. The required equipment includes a phase-contrast microscope, a hemocytometer (Neubauer improved) and a petri dish. Additional expensive pipettes are not needed. The overall handling of the test kit, ease of use, staining quality of the granulocytes and countability of the cells in the chamber were judged to be good. Counting of all leukocytes as indicated by the manufacture was not applied in this study as a clear differentiation between lymphocytes and thrombocytes was impossible. In a few cases, recognition of the granulocytes was easier with the reference method than with the test method. This may be due to different staining properties of the cells of the various avian species and to a lower degree due to differences in the Natt-Herrick solutions used for the reference method and the test method, although both method use methyl-violet 2B.

The Natt-Herricks-Tic® method performed well in terms of the ease of use, staining quality and countability of the granulocytes with a hemocytometer. This method allows easy in-house analysis by clinicians or technicians. The assessment of the analytical performance of the Natt-Herricks-Tic® method showed that the test was precise enough for a manual method and demonstrated good agreement with the reference method. It can be recommended as a reliable and suitable method for determining white blood cell counts in avian EDTA blood if nonstatistical quality control measures are used in daily routines. The application of individual reference intervals for interpreting white blood cell counts in birds may improve the diagnostic performance of this important analytic tool in a clinical setting.

EDTA blood samples from 40 birds with various clinical signs that were admitted between May 2014 and July 2015 to the Clinical Laboratory, Vetsuisse Faculty, University of Zurich, were included in the study. All samples were analyzed within 24 h after collection with the reference method and the test method. The study population consists of 22 different species: 7 Gy parrots, 2 gamefowl, 5 common buzzards, 1 capercaillie, 1 lovebird, 3 keas, 1 stork, 1 scarlet macaw, 1 bar-headed goose, 1 African pygmy goose, 1 toucan, 4 cockatoos, 1 blue-fronted amazon, 1 cockatiel, 1 domestic chicken, 1 crow, 1 Virginia quail, 1 peacock, 2 geese, 3 guinea fowl, 1 parrot, and 1 mute swan.

The within-assay precision of one EDTA-blood sample from a kea was analyzed 10 consecutive times by a skilled technician with the reference method and the Natt-Herrick-Tic® test method.