"CLINICAL PATHOLOGY

Removal of blood and urine samples

Food and water were withdrawn overnight prior to collection of samples. Overnight urine samples were collected and blood was withdrawn under light ether anaesthesia from the orbital sinus of the five male and five female rats selected from Groups 1 and 4 (see TREATMENT PROCEDURE) and all rats of Groups 2 and 3. These samples were collected shortly after 4 weeks of treatment (Day 30).

Investigations were also carried for rats of the recovery sub-groups shortly after their two-week recovery period (Day 44). The blood and urine sampling procedures in this instance were similar to those carried out on Day 30.

The blood samples collected were divided into tubes as follows:

EDT A anticoagulant . . . . . . . . . . . . . . for haematological investigations

Citrate anticoagulant . . . . . . . . . . . . . . for coagulation tests

Heparin anticoagulant ............. for biochemical tests

The estimations performed on blood and urine samples have been listed below together with an abbreviated title (used in Tables and Appendices), the methods and the units of measurement applicable:

Haematology

The following parameters were analysed with an Ortho ELT-1500 Analyser, using standard Ortho methodology:

Packed cell volume (PCV) (%)

Haemoglobin (Hb) (g/dl)

Red blood cell count (RBC) (x10^6/mm^3)

Platelet count (Plts) (x10^3/mm^3)

Absolute indices:

Mean corpuscular haemoglobin concentration (MCHC) Calculated: HB (g/dl) x 100 + PCV ( % )

Mean corpuscular volume (MCV) (fl)

Calculated: PCV (%) x 10 / RBC (x10^6/mm3)

Total white blood cell count (WBC) (x10^3/mm3)

The following were analysed using the appropriate methodology as described below:

Differential white blood cell count (Diff) - by standard microscopy of a blood smear stained with modified Wright's stain counting 100 cells (x10^3/mm3):

Neutrophils (N)

Lymphocytes (L)

Eosinophils (E)

Basophils (B)

Monocytes (M)

The percentage distribution of each cell type was determined by microscopy. These values were then converted to absolute values by computer. This inevitably involved a "rounding off" in a proportion of the results and for this reason the measured total WBC may differ slightly from the total of the different cell types.

Additionally, blood film slides were examined for morphological abnormalities. No abnormalities that were considered to be related to treatment were noted, but their incidence is recorded in the haematology appendix.

The following parameter was also analysed:

Thrombotest (TT) - Method of Owren, P.A. (Lancet, 1959, ii, 754)"

"The following sequence of statistical tests was used for bodyweight, food consumption, organ weight and clinical pathology data: if the data consisted predominantly of one particular value (relative frequency of the mode exceeded 75%), the proportion of values different from the mode was analysed by Fisher's exact test (1) followed by Mantel's test for a trend in proportions (2). Otherwise:

Bartlett's test (3) was applied to test for heterogeneity of variance between treatments. Where significant (at the 1 % level) heterogeneity was found, a logarithmic transformation was tried to see if a more stable variance structure could be obtained.

If no significant heterogeneity was detected (or if a satisfactory transformation was found), a one-way analysis of variance was carried out. If significant heterogeneity of variance was present, and could not be removed by a transformation, the Kruskal-Wallis analysis of ranks (4) was used.

Analyses of variance were followed by Student's 't' test and Williams' test (6) for a dose-related response, although only the one thought more appropriate for the response pattern observed was reported. The Kruskal-Wallis analyses were followed by the non-parametric equivalents of the 't' test and Williams' test (Shirley's test, (5)).

Organ weight analyses were initially carried out using analysis of variance as outlined, on the previous page, on absolute organ weight. Covariate analysis of organ weight data (with final bodyweight as covariate) was also be performed using adjusted weights for organs where a correlation between organ weight and bodyweight was established at the IO% level of significance (7).

Additional or alternative statistical methods were used when considered appropriate.

Significant differences between control animals and those treated with the test substance are expressed at the 5% (*P<0.05) or 1 % (**P<0.01) level."

Female CD (Sprague Dawley) Rat