Aqueous lithium heparin is a superior anticoagulant to solid heparin for blood collection from the retro-orbital sinus of rats

Transcription

1 272 Laboratory Animals (1991) 25, Aqueous lithium heparin is a superior anticoagulant to solid heparin for blood collection from the retro-orbital sinus of rats M. R. SLAUGHTER & J. S. MOEN SmithKline Beecham Pharmaceuticals, The Fry the, Welwyn, Herts AL6 9AR, UK Summary Blood specimens from the retro-orbital sinus of 80 Sprague Dawley rats were collected into tubes containing lithium heparin either as a solid or an aqueous solution. Plasma was separated for blood chemistry analysis. Twenty-eight blood specimens collected into tubes containing solid heparin were clotted and eight specimens were partially clotted making these samples unsuitable for some analyses. None of the specimens collected into heparin solution showed any evidence of clotting. The variances of lactate dehydrogenase and a-hydroxy butyrate dehydrogenase activities in plasma prepared with solid heparin were significantly greater than those prepared with heparin solution. Lithium heparin solution is now used routinely in our laboratory. Keywords: Liquid lithium heparin anticoagulant; retro-orbital sinus blood collection; plasma biochemistry The measurement of blood chemistry and haematological parameters in toxicological studies in the rat serves a vital role in the early detection of toxic effects of test chemicals. Plasma is preferable to serum for blood chemistry analysis since, on centrifugation, more supernatant is obtained, an important factor when routine toxicological studies demand assay of numerous analytes on a limited sample. During toxicology studies, we obtain plasma by bleeding rats from the retro-orbital sinus (Wechsler, 1983), collecting approximately 4ml blood into standard tubes containing anti- Received 6 April 1990; accepted 21 January 1991 coagulant. This includes 2 ml collected into a heparinized tube for plasma biochemistry. However, despite strict adherence to sampling procedures, partial or complete clotting of blood often occurs, particularly with heparinized specimens from male rats. This renders the specimen unsuitable for a number of assays, necessitating a rebleed of the animals or resulting in a loss of data. This has not been a problem with blood collected from the dorsal aorta into 5 ml tubes containing solid heparin. Furthermore, the laboratory's reference ranges for lactate dehydrogenase (LDH) and aspartate transaminase (AST) in blood collected from the retro-orbital sinus are markedly wider than those for blood collected from the dorsal aorta, suggesting that undetected partial clotting sometimes occurs with the former samples. During the clotting process, platelet disintegration occurs, with the liberation of intracellular contents such as LDH and AST into the surrounding medium. LDH activity is particularly high in platelets whereas AST activity is much lower (Kluge & Haug, 1970);thus the plasma activity of LDH and a-hydroxybutyrate dehydrogenase (HBDH) is more sensitive to clotting than AST. Increases in plasma potassium may also occur if clotting is significant. The effect of undetected clotting cannot be the only explanation, however, since different sampling sites and collection methods also affect values for LDH and AST in serum (Neptun et al. 1985). In an investigation carried out in our laboratory using double the normal quantity of heparin, the incidence of clotting was not reduced (Moen & Slaughter, unpublished data). We therefore decided to evaluate the use of heparin solution on the premise that the anticoagulant would be

2 Aqueous lithium heparin as an anticoagulant 273 more readily available in aqueous form as the dissolution step is omitted. Materials and methods Animals Eighty Sprague-Dawley (Crl: CD(SD)BR) male rats of 10 to 20 weeks of age, obtained from Charles River UK, Ltd were used. The rats were housed, 5 per cage, in polypropylene cages suspended over trays containing absorbent crape paper, in a barriered room maintained at 21 C ± 2 C and 55± relative humidity, with a 12 h light/12 h dark cycle. Filtered, nonrecirculated air was supplied at a rate of 14 to 16 room changes per hour. A standard commerciallaboratory diet for rats (SQC Rat and Mouse Maintenance Diet No.1, SDS Ltd, UK) and mains drinking water were freely available, except food was removed at approximately 1700 h on the day prior to blood sampling. The animals were allowed a period of at least 2 weeks' acclimatization prior to allocation to study. Blood collection Standard 9 x 44 mm plastic tubes containing 32 units of solid lithium heparin were obtained from Teklab (Medical Laboratories Ltd, Sacriston, Durham, UK). Tubes containing heparin solution were prepared by dissolving lithium heparin (porcine intestinal mucosa, Sigma Chemical Co. Ltd, Dorset, UK) in deionized water to give a final concentration of 16 units/ml and dispensing 20 ILlof this solution (i.e 32 units of heparin) into 9 X 44 mm plastic tubes immediately before use. Blood specimens (4 ml) were obtained from the retro-orbital sinus under light ether anaesthesia using flat ended capillary tubes not coated with anticoagulant (Hawksley & Sons Ltd, West Sussex, UK). Two millilitres of blood was dispensed into each type of sample tube, mixed immediately by gentle inversion and then placed on a roller mixer. The animals were bled by two operators in three batches, for each of which, half the samples were added to the tubes containing solid heparin before the tubes containing liquid heparin, the procedure being reversed for the other half. Specimens were inspected for clotting and discarded if affected. The remaining specimens were centrifuged at 1500g for 10 min and inspected for haemolysis and further evidence of clotting. Plasma was transferred into specimen containers and stored at 4 C for assay the same day. This preparation was completed within 1 h of blood collection. Paired specimens were processed simultaneously throughout the study to minimize procedural and analytical variance. Instrumentation Alanine aminotransferase (ALT), AST, HBDH, LDH, bilirubin, calcium, creatinine, phospholipids and triglycerides were assayed on a Cobas Bio centrifugal analyser (Roche Diagnostics, Hertfordshire, UK) and alkaline phosphatase (ALP), albumin, total protein, cholesterol, phosphate and urea were assayed on a Kone Progress Analyser (Lab medics Ltd, Cheshire, UK). Sodium and potassium were assayed on an I.L. 943 flame photometer (Instrumentation Laboratory Ltd, Cheshire, UK) and chloride was assayed by coulometry on a Corning 925 chloride meter (Ciba-Corning Diagnostics Ltd, Essex, UK). Standard routine assay methods were used throughout. Enzymes were assayed at 30 C. Statistical procedures All results were treated as paired data. Prior to statistical analysis, HBDH, LDH and AST results were excluded for samples which were retained for analysis but showed haemolysis (albeit minimal) and/or any evidence of focal clotting. Results for potassium were also excluded in the latter case. Where bilirubin concentration fell below the detection limit of the assay, the sample was assigned the detection limit value of O 5(.tM. The data were tested for normality and statistical outliers by the use of normal probability plots. Where required, logarithmic transformations were carried out and the data retested for normality. The 95% confidence interval for the observed difference between means was used to test the null hypothesis that the true difference

3 274 Slaughter & Moen between means was equal to a 1<Vo dilution error in the aqueous heparin samples. Therefore, if this predicted difference falls outside the confidence interval then the observed difference will be significant at the 5% level. As the sample variances are correlated, comparison of variances was made using Pitman's method, the F-test being inappropriate in this case (Snedecor & Cochran, 1980). Linearity between samples was assessed by the use of scatter diagrams and the line of best fit was calculated by a variation of Deming's method, where Deming's slope was derived from the least squares slope (Cornbleet & Gochman, 1979). Although this approach requires the calculation of the Pearson correlation coefficient, the latter was not used to assess correlation between samples as it is dependent upon the range covered by the data, thereby rendering it unreliable (Westgard & Hunt, 1973). Statistical significance was tested at the 5% level. Results Twenty-eight specimens collected into tubes contaiuing solid heparin were discarded because of severe clotting and 8 specimens showed evidence of focal clotting which was reflected by significant increases in LDH, HBDH, AST and potassium compared with their respective liquid heparin pairs. It was found that the order in which the tubes were filled with blood did not affect the degree of clotting. Statistical analysis of results is summarized in Table 1. There were statistically significant differences between means for ALP, total protein, phosphate and bilirubin. The differences were all slight, however, and for the first 3 analytes, probably due to effects of focal clotting. Thus, leakage of phosphates from red blood cells and ALP from platelets and leucocytes will produce increases in their plasma concentrations and a loss of fibrinogen will result in a decrease in plasma total protein (Ciuti & Rinaldi, 1989). The cause of higher bilirubin concentration seen in plasma prepared with solid heparin is not known but may be the result of protection from photolysis by the presence of a paper label on the tubes containing solid heparin. Owing to the very narrow range of data obtained Table 1. Statistical summary of plasma biochemistry values for blood collected from the retro-orbital sinus of the rat using liquid or solid lithium heparin as an anticoagulant Regression Mean Mean Analyte n equation (solid heparin) (heparin solution) ALP lull 52 y= 1'05x-18' m ALT lull 51 y=0'98x AST lull 33 y=0'87x LDH lull 33a Nonlinear '6 v HBDH lull 33a Nonlinear '3 v Total protein gil 52 y= 1'07x- 4, '1 m Albumin gil 51 y= 1'02x- 0, ,8 Urea mmolll 47 y= l'oox Sodium mmolll 48 See text '4 Potassium mmolll 45 y=0'89x+ 0, Chloride mmolll 52 See text ,6 Calcium mmolll 52 See text '52 Inorganic phosphate mmolll 49 y=0'96x m Cholesterol mmolll 51 y=0'95x+ 0, Triglycerides mmolll 51 y=i'ox - 0,007 0,54 0'54 Phospholipids mmolll 51 y=0'98x+ 0,003 0,98 0,96 Total bilirubin JLmolll 49 See text ,9 m Creatinine JLmolll 49 y=i'ox+ 0' n = Number of data pairs available after exclusion of data resulting from clotting &/or haemolysis. a = Statistical analysis performed on logarithms of data. Antilogs shown. m = Significant differ'ence between means (P< = 0-05). v = Variance of solid heparin data is greater than that for heparin solution (P< 0'05).

4 Aqueous lithium heparin as an anticoagulant 275 for bilirubin, calcium, chloride and sodium, regression equations for these analytes were not calculated as they are considered inappropriate for clustered data. The regression equations for AST and potassium indicate some degree of systematic bias, although this should be interpreted with caution as the range of data is still rather limited. The distribution of data for bilirubin was distorted by adjustment of results that fell below the detection limit of the assay and so precluded comparison of variances. Regression equations were not calculated for LDH and HBDH as a straight line relationship was not apparent (Figs 1 & 2). The variances of data for LDH and HBDH activities in plasma prepared with solid heparin were significantly greater than those for plasma prepared with heparin solution; this is also evident in the scatter diagrams. A number of data points for both HBDH and LDH show a marked positive bias for the solid heparin sample, which appear as deviations from a main cluster of data representing the majority of results. One matched pair show higher LDH, HBDH, and (to a lesser extent) AST activities, for the heparin solution sample. This may have been due to accidental aspiration of cells from the plasma-cell interface during plasma separation. Discussion As platelets contain high activities of LDH and HBDH, it has been recommended that plasma rather than serum be used for measurement of enzyme activity in animal species with high platelet counts, such as the rat (Methfessel & Deml, 1967). It is therefore important that plasma preparation is free from the effects of partial clotting, when significant amounts of LDH with HBDH activity are released from disintegrating platelets. This problem is much less severe for AST as its activity in platelets is much less than that for LDH and HBDH. The use of solid heparin as an anticoagulant for blood collected from the retro-orbital sinus of rats has frequently failed to prevent sample coagulation. No improvement was seen when the amount of heparin was doubled (Moen & c: 250 u.,g 200 o ::l ~ 150 -Q) J: IUfl@30 C Solid heparin Fig. 1. Lactate dehydrogenase activity in rat plasma from blood collected from the retro-orbital sinus using liquid or solid lithium heparin as anticoagulant c: 0 u::: 0- ::l <')0 Ul 40 ~. -Ill :Ja. -Q) 30.. J: IUti (j,) 30 C Solid heparin Fig. 2. a-hydroxybutyrate dehydrogenase activity in rat plasma from blood collected from the retro-orbital sinus using liquid or solid lithium heparin as anticoagulant. Slaughter, unpublished data). The study reported here, however, has shown that using heparin in solution greatly improves its anticoagulant effect and indicates that it is the immediate availability of heparin and not its initial amount that is crucial. A number of factors can be attributed to the apparent increase in coagulability of blood collected from the retro-orbital sinus compared with that from the dorsal aorta. The comparatively large bored, flat-ended capillary tube used in the retro-orbital sinus technique is more likely to cause significant tissue damage than the finer, 70

5 276 Slaughter & Moen bevelled needle used to penetrate the dorsal aorta, thus increasing the chances of activation. Contact activation by the glass capillary itself and the involvement of varying degrees of stress in different methods of blood collection (Landaburu et al., 1971) may be further contributory factors. Since this study we have been using liquid heparin tubes supplied by Teklab containing 32 units per 2 ml tube. The tubes can be stored for 3 months at 4 C with no deterioration in performance; the only significant finding was a O' 5 % increase in sodium (P = O' 02) for blood collected into stored tubes when compared to freshly prepared tubes. Furthermore, since using these tubes routinely we have seen marked reductions in the variability of plasma LDH and HBDH activities, as predicted from the results of this study. This has also been observed for AST even though no significant reduction in variance was seen in this study. The maximum sensitivity of clinical chemistry tests to drug-induced in vivo biochemical alterations depends on minimizing changes caused by any other variables. Artefactual changes in plasma concentration of analytes resulting from sample processing is one important such variable. We have found that aqueous heparin eliminates the problem of clotting and reduces the variability of results for LDH, HBDH and AST when used for collecting blood from the retro-orbital sinus of the rat and we therefore recommend it for this purpose. Acknowledgments The authors thank Ian McPherson for his help and advice on statistical analysis and also the technical staff in the Chemical Pathology Unit for their assistance. References Ciuti R & Rinaldi G (1989) Serum and plasma compared for use in 19common chemical tests performed in the Hitachi 737 Analyser. Clinical Chemistry 35, Cornbleet PJ & Gochmann N (1979) Incorrect least-squares regression coefficients in method-comparison analysis. Clinical Chemistry 25, Kluge A & Haug H (1970) Influence of coagulation processes and thrombocyte disintegration on enzyme activities in human blood. Blut 21, Landaburu RH, Castellanos DE, Giavedoni E & Lo Presti CA (1971) Neurohumoral control of blood coagulation. I. Effects of adrenergic drugs or ACTH administration and stress. Acta Physiologica Latinoamericana 21, Methfessel J & Deml L (1967) Comparative investigations on enzyme activities in blood-plasma and serum of man, rabbits and rats. Zeitschrift fur Klinische Chemie und Klinische Biochemie 5, Neptun DA, Smith en & Irons RD (1985) Effect of sampling site and collection method on variations in baseline clinical pathology parameters in Fischer-344 rats. Fundamental and Applied Toxicology 5, Snedecor GW & Cochran WG (1980) Statistical Methods, 7th edn, pp Iowa: The Iowa State University Press Wechsler SJ (1983) Blood collection techniques and normal values for ferrets, rabbits and rodents-a review. Veterinary Medicine/Small Animal Clinician 78, Westgard JO & Hunt MR (1973) Use and interpretation of common statistical tests in method-comparison studies. Clinical Chemistry 19, 49-57