27Despite the advances in canine medicine and the rapid gaining of attention of canine 28 models in biomedical field and particularly in hemoglobin genes research, the studies on canine 29 hemoglobin composition are sparse with ambiguous findings. Our aim was: i) to investigate 30 the electrophoretic pattern of canine hemoglobin and the possible effect of age, sex, and anemia 31 using a capillary electrophoresis assay, and ii) to validate this assay and calculate reference 32 intervals (RIs) for canine hemoglobin fractions. Blood samples were collected from 53 healthy 33 and 42 dogs with regenerative and non-regenerative anemias. The Sebia Capillarys 2 flex-34 piercing was used for hemoglobin analysis and it was validated using canine blood samples. R 35 statistical language was employed for the statistical analyses. A major hemoglobin fraction 36 (named HbA 0 ) and a minor one (named HbA 2 ) were identified in 100% and 47.4% of samples, 37 respectively. The within-run and between-run CV was 0.1% for HbA 0 and 9.1% and 11.2% for 38 HbA 2 , respectively. The extremely narrow range of HbA 0 and HbA 2 values hampered a 39 linearity study using canine blood samples. The RIs for HbA 0 and HbA 2 were 98.9-100% and 40 0-1.1%, respectively. HbA 0 and HbA 2 values were not correlated with age (P=0.866). No 41 differences were observed in the median HbA 0 and HbA 2 between the two sexes (P=0.823), 42 and healthy and anemic dogs (P=0.805). In conclusion, the capillary electrophoresis revealed 43 a major hemoglobin fraction and an inconsistently present minor fraction. No effect of age, 44 sex, or anemia was detected. The assay used was validated and RIs were generated, so as to be 45 suitable for use in future investigations. 46 47 48 49 50 3 51 6 125brief, the EDTA-treated whole blood sample is diluted with a hemolysing solution and the 126 resulting solution is then hydrodynamically injected at the anodic end of the capillary. A 127 constant, high voltage is applied for 8 min, which allows the migration and separation of the 128 hemoglobin variants. These are then directly detected by spectrophotometry (415 nm) and the 129 electrophoretograms are automatically generated. The total output time is approximately 20 130 min for the first run and 12 min for every other run.
131The validation of the analyzer was initially designed to include linearity, repeatability, 132 and reproducibility. The repeatability or within-run precision was evaluated using blood 133 samples from three dogs. Each sample was measured eight times in succession and the 134 coefficient of variation (CV) was calculated. Blood samples from the same three dogs were 135 used for the evaluation of reproducibility or between-run precision. Six aliquots were made 136 from each sample and were measured over a period of 3 days; then, the CV was calculated.
137The distribution of data was assessed using the Shapiro-Wilk test. The 95% reference 138 intervals (RIs) were calculated using the non-parametric method, while the 90% confidence 139 intervals (CIs) for ...