LETTER TO THE EDITOR
Clinical Pathology
Testing
in
Toxicity Studies
GEOFF BROWN Chairman, Association for Comparative Haematology Department of Clinical Pathology, Huntingdon Research Centre, Huntingdon, Cambs PE18 6ES
I refer to the recent communication &dquo;Clinical Pa-
thology Testing Recommendations for Nonclinical Toxicity and Safety Studies&dquo; (8), and congratulate the Joint Task Force of the AACC-DACC/ASVCP for the work they have done in producing it. The need for a review of the existing recommendations by practising Comparative Clinical Pathologists was well overdue. After careful consideration of the report, I offer the following comments and suggestions with regard to the Haematology content. There are many welcome suggestions in the recommendation, in particular, the observations on overnight fasting prior to sampling, the random collection and analysis of samples, the comment regarding the analysis of blood smears at the end of a long-term carcinogenicity study, the list of core tests, and the insensitivity of the whole blood clotting time. The importance of analysing individual samples (as opposed to pooled samples) cannot be stressed too much, since individual reaction to treatment may be masked by pooled samples, especially where such reactions are of an immunological nature, with severity depending on each animal’s im-
munological
status.
The use of the differential white cell count on its own is particularly meaningless. If marked differences are noted between treated animals and controls, even a speculative explanation becomes impossible unless the actual numbers of each cell type are known. A total white cell count is therefore obligatory in these cases: this being so, it is only a very small step to perform a total red cell count at the same time, thus identifying the possibility of erythroproliferative disorders. The use of pattern recognition methods for statistical analysis rather than the frequently adopted Student’s t-test is also to be commended, even if it only reduces the need to explain why statistical differences observed in the latter are not considered biologically significant.
Whilst
agreeing with the comments on the value
of pre-dose samples in rodent studies, there may be a stronger case for such evaluations and comparison with larger animals (dogs, primates, etc). At the other end of the scale, analysis of samples between week 52 and 104 in rodent studies may reveal an unexpected side-effect: that of enhancing age-related
changes, or perhaps more importantly, of delaying such changes. I recall a study performed some years
(unpublished) where the incidence of tumours in the later stages of the study was significantly, and dose relatedly, less than in control animals. The inclusion of an early sample, within 7 days of the commencement of certain studies, is also to be recommended. This will demonstrate early reactions that could otherwise be masked by subsequent adaption or enzyme induction. However, some thought should be given to taking a suitable sample at or close to the time of the maximum blood level of the test compound, in order to exclude shortacting acute reactions (3). The use of the orbital plexus for blood sampling can be of great value in toxicity studies, particularly if the operation is performed by highly experienced staff. In our experience, working in a large toxicity unit where several hundred samples are taken every day, the appointment of one or two staff members solely to this procedure has resulted in extremely ago
high quality samples, including satisfactory samples for platelet function analysis. However, with several countries now apparently banning this procedure on aesthetic grounds, these comments may be largely superfluous. I must however, take issue with the recommendations on Bone Marrow Evaluation, to some extent reflected in the Reticulocyte Count. Just as an enlarged liver, showing work-load hypertrophy, can give relatively normal enzyme levels during the course of a study, so can a similar situation exist 672
Downloaded from tpx.sagepub.com at NEW YORK UNIV LIBRARY on May 30, 2015
673
with the bone marrow, where an erythroid hyperplasia may counterbalance a mild or slowly progressive haemolytic anaemia in the peripheral blood (1). It is therefore no more valid to exclude bone marrow assays than to exclude liver weight/pathology on the grounds of normal liver enzyme values in the peripheral blood during the course of the
views of many Comparative Haematologists in the UK, and taken together with the AACC-
resent the
DACC/ASVCP report, show that there is sufficient common ground, given some further discussion and negotiation, to make the goal of an agreed international approach attainable.
study. The absence of the Thrombotest (5) from the clotting profile leaves a gap which could be easily filled. This test measures essentially the same Factors as Quick’s Prothrombin Time, but has been shown to be more sensitive in dogs (7) and in rats (4,6). In our own experience, the recent screening of approx. 200 dogs for Factor VII Deficiency showed that normal dogs gave a Prothrombin Time of 5.8-7.2 seconds (Simplastin: Organon-Teknika), with deficient dogs giving times of 7.9 seconds upwards. The same samples analysed by Thrombotest gave a range for normal dogs of 13-18 seconds, with deficient dogs in excess of 42 seconds-clearly an advantage when error into account. Similarly, Prothrombin values for a control group of rats was 12.2 seconds, compared to 13.4 seconds for a group treated with Coumarin (unpublished data). The same samples analysed by Thrombotest gave mean values of 22.3 and 51.6 seconds, respectively. Since the test can be performed on whole blood, sample size is minimal, and the test can be performed &dquo;at the bedside,&dquo; using only 0.05 ml blood pipetted direct from a tail-vein transection. Whilst the views expressed in this letter are essentially personal, and have been recently published as an Editorial Comment (2), they nevertheless rep-
taking experimental mean
REFERENCES 1. Brown G (1991). The left shift index: A useful guide to the interpretation of marrow data. Comp. Haematol. Int. 1: 106-111. 2. Brown G (1992). Editorial Comment: Haematology tests in toxicology—Time for a re-think? Comp. Haematol. Int. 2: 231-235. 3. Ganney BA, Brown G (1991). Disseminated intravascular coagulation in the rat: A case history from a toxicological study. Comp. Haematol. Int. 1: 172-177. 4. Godsafe PA, Singleton BK (1992). The use of the whole blood thrombotest time (1/51) as a routine monitor of vitamin-K dependent blood coagulation factor levels in the rat. Comp. Haematol. Int. 2: 51-55. 5. Owren PA (1959). Thrombotest: A new method for controlling anticoagulant therapy. Lancet ii: 754. 6. Schofield MA, Hall DE (1991). Detection of hepato-
toxicity by means of blood coagulation changes. Comp. Haematol. Int. 1: 200-204. 7.
Spurling NW, Burton LK, Pilling T (1974). Canine Factor VII deficiency: Experience with a modified thrombotest method in distinguishing between geno-
8.
Weingand K, Bloom J, Carakostas M, Hall R, Helfrich M, Latimer K, Levine B, Neptun D, Rebar A, Stitzel K, Troup C (1992). Clinical pathology testing recommendations for nonclinical toxicity and safety stud-
types. Res. Vet. Sci. 16: 228-239.
ies. Tox. Path. 20: 539-543.
Downloaded from tpx.sagepub.com at NEW YORK UNIV LIBRARY on May 30, 2015
Clinical Pathology
Testing
in
Toxicity Studies
GEOFF BROWN Chairman, Association for Comparative Haematology Department of Clinical Pathology, Huntingdon Research Centre, Huntingdon, Cambs PE18 6ES
I refer to the recent communication &dquo;Clinical Pa-
thology Testing Recommendations for Nonclinical Toxicity and Safety Studies&dquo; (8), and congratulate the Joint Task Force of the AACC-DACC/ASVCP for the work they have done in producing it. The need for a review of the existing recommendations by practising Comparative Clinical Pathologists was well overdue. After careful consideration of the report, I offer the following comments and suggestions with regard to the Haematology content. There are many welcome suggestions in the recommendation, in particular, the observations on overnight fasting prior to sampling, the random collection and analysis of samples, the comment regarding the analysis of blood smears at the end of a long-term carcinogenicity study, the list of core tests, and the insensitivity of the whole blood clotting time. The importance of analysing individual samples (as opposed to pooled samples) cannot be stressed too much, since individual reaction to treatment may be masked by pooled samples, especially where such reactions are of an immunological nature, with severity depending on each animal’s im-
munological
status.
The use of the differential white cell count on its own is particularly meaningless. If marked differences are noted between treated animals and controls, even a speculative explanation becomes impossible unless the actual numbers of each cell type are known. A total white cell count is therefore obligatory in these cases: this being so, it is only a very small step to perform a total red cell count at the same time, thus identifying the possibility of erythroproliferative disorders. The use of pattern recognition methods for statistical analysis rather than the frequently adopted Student’s t-test is also to be commended, even if it only reduces the need to explain why statistical differences observed in the latter are not considered biologically significant.
Whilst
agreeing with the comments on the value
of pre-dose samples in rodent studies, there may be a stronger case for such evaluations and comparison with larger animals (dogs, primates, etc). At the other end of the scale, analysis of samples between week 52 and 104 in rodent studies may reveal an unexpected side-effect: that of enhancing age-related
changes, or perhaps more importantly, of delaying such changes. I recall a study performed some years
(unpublished) where the incidence of tumours in the later stages of the study was significantly, and dose relatedly, less than in control animals. The inclusion of an early sample, within 7 days of the commencement of certain studies, is also to be recommended. This will demonstrate early reactions that could otherwise be masked by subsequent adaption or enzyme induction. However, some thought should be given to taking a suitable sample at or close to the time of the maximum blood level of the test compound, in order to exclude shortacting acute reactions (3). The use of the orbital plexus for blood sampling can be of great value in toxicity studies, particularly if the operation is performed by highly experienced staff. In our experience, working in a large toxicity unit where several hundred samples are taken every day, the appointment of one or two staff members solely to this procedure has resulted in extremely ago
high quality samples, including satisfactory samples for platelet function analysis. However, with several countries now apparently banning this procedure on aesthetic grounds, these comments may be largely superfluous. I must however, take issue with the recommendations on Bone Marrow Evaluation, to some extent reflected in the Reticulocyte Count. Just as an enlarged liver, showing work-load hypertrophy, can give relatively normal enzyme levels during the course of a study, so can a similar situation exist 672
Downloaded from tpx.sagepub.com at NEW YORK UNIV LIBRARY on May 30, 2015
673
with the bone marrow, where an erythroid hyperplasia may counterbalance a mild or slowly progressive haemolytic anaemia in the peripheral blood (1). It is therefore no more valid to exclude bone marrow assays than to exclude liver weight/pathology on the grounds of normal liver enzyme values in the peripheral blood during the course of the
views of many Comparative Haematologists in the UK, and taken together with the AACC-
resent the
DACC/ASVCP report, show that there is sufficient common ground, given some further discussion and negotiation, to make the goal of an agreed international approach attainable.
study. The absence of the Thrombotest (5) from the clotting profile leaves a gap which could be easily filled. This test measures essentially the same Factors as Quick’s Prothrombin Time, but has been shown to be more sensitive in dogs (7) and in rats (4,6). In our own experience, the recent screening of approx. 200 dogs for Factor VII Deficiency showed that normal dogs gave a Prothrombin Time of 5.8-7.2 seconds (Simplastin: Organon-Teknika), with deficient dogs giving times of 7.9 seconds upwards. The same samples analysed by Thrombotest gave a range for normal dogs of 13-18 seconds, with deficient dogs in excess of 42 seconds-clearly an advantage when error into account. Similarly, Prothrombin values for a control group of rats was 12.2 seconds, compared to 13.4 seconds for a group treated with Coumarin (unpublished data). The same samples analysed by Thrombotest gave mean values of 22.3 and 51.6 seconds, respectively. Since the test can be performed on whole blood, sample size is minimal, and the test can be performed &dquo;at the bedside,&dquo; using only 0.05 ml blood pipetted direct from a tail-vein transection. Whilst the views expressed in this letter are essentially personal, and have been recently published as an Editorial Comment (2), they nevertheless rep-
taking experimental mean
REFERENCES 1. Brown G (1991). The left shift index: A useful guide to the interpretation of marrow data. Comp. Haematol. Int. 1: 106-111. 2. Brown G (1992). Editorial Comment: Haematology tests in toxicology—Time for a re-think? Comp. Haematol. Int. 2: 231-235. 3. Ganney BA, Brown G (1991). Disseminated intravascular coagulation in the rat: A case history from a toxicological study. Comp. Haematol. Int. 1: 172-177. 4. Godsafe PA, Singleton BK (1992). The use of the whole blood thrombotest time (1/51) as a routine monitor of vitamin-K dependent blood coagulation factor levels in the rat. Comp. Haematol. Int. 2: 51-55. 5. Owren PA (1959). Thrombotest: A new method for controlling anticoagulant therapy. Lancet ii: 754. 6. Schofield MA, Hall DE (1991). Detection of hepato-
toxicity by means of blood coagulation changes. Comp. Haematol. Int. 1: 200-204. 7.
Spurling NW, Burton LK, Pilling T (1974). Canine Factor VII deficiency: Experience with a modified thrombotest method in distinguishing between geno-
8.
Weingand K, Bloom J, Carakostas M, Hall R, Helfrich M, Latimer K, Levine B, Neptun D, Rebar A, Stitzel K, Troup C (1992). Clinical pathology testing recommendations for nonclinical toxicity and safety stud-
types. Res. Vet. Sci. 16: 228-239.
ies. Tox. Path. 20: 539-543.
Downloaded from tpx.sagepub.com at NEW YORK UNIV LIBRARY on May 30, 2015
