in partial fulfilment of the requirements for the degree of

Transmission of tuberculosis caused by Mycobacterium bovis

between possums and possums and cattle

A thesis presented

in partial fulfilment of the requirements for the degree of

Doctor of Philososphy at Massey University

R. Jackson

1995

ii

Abstract

Tuberculosis caused by Mycobacterium bovis was diagnosed in 59 of 632 possums (Trichosurus vulpecula) individually identified over a 52 month period, during a longitudinal study of the naturally occurring disease in possums at a 21 hectare bush pasture location on a farm at Castlepoint in the Wairarapa. The disease exhibited marked spatial and temporal clustering and was continuously present in the population for the whole period.

The disease had a relatively long duration of up to 22 months and four distinct stages were demonstrated in cross-sectional studies. Among tuberculous possums, prevalences of up to 0.15 (?0.11) were recorded in the first stage prior to development of gross lesions. After dissemination started, the disease showed rapid generalisation to multiple sites by haematogenous and/or lymphatic spread to the next stage when gross lesions were evident, particularly in lung, axillary and inguinal lymphocentres. In the third stage, lesions were disseminated through almost all lung lobes, discharging fistulae were common and kidney, intestine and mammary gland were commonly affected by both gross and microscopic lesions. Behaviour and outward signs of health were unaffected prior to the terminally-ill stage, lasting for up to 2 months.

In common with other marsupials studied to date and in contrast with most eutherians, there are no popliteal lymph nodes and efferent drainage from the inguinal lymphocentre passes directly to the deep axillary group of lymph nodes via an inguinoaxillary trunk. All subcutaneous lymph drainage passes through either the superficial cervical or the axillary lymphocentres before entering the venous system.

Studies of survival of Mycobacterium bovis organisms in different natural habitats showed a relatively short period of survival of M. bovis outside hosts and support a conclusion that environmental contamination of pasture, particularly in summer months, may be relatively unimportant in the epidemiology of tuberculosis in cattle, deer and possums.

The weight of evidence favours transmission of infection by the respiratory route and it would seem that transmission of tuberculosis between possums occurs through two major and one minor pathway.

The first major pathway is pseudo-vertical transmission from mother to joey during the rearing process. The second major transmission mechanism is direct horizontal transmission among adult possums with available evidence suggesting that this takes place around the locality where a possum dens, probably during competition and threat/agonistic behaviour and during courting and mating

iii

activity. The third and probably least important pathway is indirect transmission among mature possums.

None of three immunosorbent assays reliably detected possums infected with tuberculosis and poor test performance was exacerbated by inconsistency between results from serially collected samples from known tuberculous possums.

iv

Acknowledgements

I undertook this postgraduate training mainly to better equip me for solving problems in a logical and systematic manner, and to help me make sense of complex issues. I was very fortunate to be given an opportunity to work on a high profile project by my chief supervisor, Professor Roger Morris, who was willing to take on a "bush vet" who scarcely knew the difference between a mean and a median. I am very grateful to him, not just for that opportunity, but also for his generous assistance and responsiveness to needs throughout the study period and for his counsel through my transition from practice to research.

My other supervisors, Dr. Geoff de Lisle and Associate Professor Roger Marshall have constantly encouraged me and given of their time and expertise willingly and I am grateful to them both.

Life within the epidemiology group at Massey University under Professor Morris has been exciting and challenging and has given me a lot of fun and pleasure. All of the postgraduate students and staff in this group have been enthusiastic and loyal and supportive of one another. I thank them all very sincerely for their assistance at various times.

Within the group, Dr. Dirk Pfeiffer has guided me over and around many of the "brick walls" thrown up by analytical techniques and I have particularly enjoyed our many discussions about epidemiological issues.

The long term nature of longitudinal studies makes them more prone to problems than short term studies. We had our share of problems at Castlepoint, but none became serious, and a large part of the smooth running of the study was due to the good sense and friendly cooperation of my good friends Ron Goile, manager of Waio and his partner, Donna Lewis, both of whom have made an outstanding contribution to the longitudinal study. Thanks are also due to Bill Maunsell, the owner of Waio, who has made a quiet but considerable contribution by making his property available for the work.

The use of the library at Massey University has given me much pleasure over the past five years, but the greatest pleasure came from the human resource within the Faculty of Veterinary Science, which shared its talents and knowledge willingly. Fiona Dickinson, secretary to Professor Morris, is a great exponent of all things pertaining to word processing and document layout, and she in particular has answered numerous requests for help and guidance in the preparation of this thesis and other reports.

The nature of my studies carried with it a high ethical cost in terms of use of animals. I trust that I

v

used them wisely and was not wasteful in the extraction and use of information from them.

I am unable to adequately express my feelings for the most outstanding person of all. June and I have now been married for 37 years and her loyalty and support for me have been remarkable. It can't have been much fun helping me go round the trap lines on occasions in foul weather, or scribing for me in a cold shelter and then facing a waist deep evening return through Flagstaff Creek in mid-winter in Westland. Such is the mettle and loyalty of June, which she constantly has demonstrated throughout our life together. Our children too have been highly supportive of us in this venture and I am grateful to them for that. We miss our friends from Alexandra very much and thank them for continuing their friendship with us over time and distance.

Ron Jackson,

Department of Veterinary Clinical Sciences, Massey University,

New Zealand.

18 August 1995

vi

Table of Contents

Abstract

Acknowledgements

Table of Contents

List of Figures

List of Tables

CHAPTER 1 Introduction

CHAPTER 2 Tuberculosis caused by Mycobacterium bovis in New Zealand

PROGRESS AND CURRENT STATUS

Testing regimes

Animal Health Board strategic plan Control costs and funding

THE CASE FOR TUBERCULOUS POSSUMS BEING A SOURCE OF TUBERCULOSIS FOR CATTLE

Evidence from tuberculin testing

The incidence of cattle tuberculosis following possum poisoning operations Restriction endonuclease typing of M. bovis

Persistence of tuberculosis in possum populations Analogy with badgers in the U.K. and Eire Why possums?

Summary

THE ROLE OF POSSUMS IN THE EPIDEMIOLOGY OF TUBERCULOSIS IN DEER

vii

THE EPIDEMIOLOGY OF TUBERCULOSIS IN POSSUMS

Introduction

Disease investigations

Experimental infection studies Cross-sectional studies

Limitations of cross-sectional study design Disease transmission

A SUMMARY OF ECOLOGY OF THE POSSUM IN NEW ZEALAND AND ITS INFLUENCE ON TUBERCULOSIS CONTROL POLICIES

Classification

Introduction into New Zealand

Reasons for success

Basic Physiology of the possum

Ecology

Population dynamics

General population structure

Sex ratio Body size

Population density Reproduction

Home range, dispersal, emigration and immigration Juvenile mortalities

Adult mortalities

Territorial defence, spacing, and social behaviour Behaviour and territorial defence

viii

TUBERCULOSIS IN OTHER ANIMALS AND THEIR IMPORTANCE AS SOURCES OF INFECTION

Introduction

Herbivores

Deer

Introduction Pathology Epidemiology Ecological factors Sheep

Goats

Rabbits and hares Horses

Flesh eating animals

Cats

Ferrets, Stoats and Weasels Hedgehogs and Rats

THE CURRENT SITUATION AND IMPLICATIONS FOR LONG TERM CONTROL

Introduction

Implications for long term control

CHAPTER 3 A study of the topography of the lymphatic system of the Australian brushtail possum (Trichosurus vulpecula)

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS RESULTS

ix

HEAD AND NECK

Parotid lymphocentre

Parotid lymph nodes

Mandibular lymphocentre

Mandibular lymph node

Deep cervical lymphocentre

Deep cervical lymph node

Palatine Tonsils

Palatine tonsil

Superficial cervical lymphocentres

PECTORAL LIMB, SUPERFICIAL THORAX AND ABDOMEN

Axillary lymphocentre

Superficial axillary lymph node (N.A.V. accessory axillary) Deep axillary lymph nodes (N.A.V. proper axillary)

PELVIC LIMB, TAIL AND MAMMARY GLAND

Inguinal lymphocentre

Inguinal lymph nodes Mammary lymph nodes

ABDOMEN AND ABDOMINAL VISCERA

Iliac lymph node

Renal lymph node Colic lymph node

Cranial mesenteric lymphocentre

Gastric lymphocentre

Gastric lymph nodes

Hepatic lymphocentre

Hepatic lymph nodes

Cysterna chyli

x

Cranial mediastinal lymph nodes

DISCUSSION

ACKNOWLEDGMENTS

CHAPTER 4 A study of environmental survival of Mycobacterium bovis in selected locations in New Zealand

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

Substrate material preparation

Timing of studies, test material collection and subsequent bacteriological examinations

RESULTS

Culture results

Weather data recording

Survival probabilities of

seasons.

Association between seasons and survival of

and den locations

Survival on pasture Survival on forest floor Survival in dens

Location and weather effects

DISCUSSION

ACKNOWLEDGMENTS

xi

CHAPTER 5 Naturally occurring tuberculosis caused by Mycobacterium bovis in brushtail possums (Trichosurus vulpecula): I An epidemiological analysis of lesion distribution

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

Necropsy and data recording procedures Selection of specimens for bacteriology Bacteriology and histopathology

Waio studies; March 1992, July and September 1993 Statistical Analysis

RESULTS

Point prevalence studies

Distribution of gross and microscopic lesions Distribution of gross and microscopic lesions

Test for symmetry of lesion distribution between both sides of the body

Association between number of lesioned sites per individual and individual characteristics Associations between occurrence of lesions of tuberculosis among specific body regions Terminally ill possums

Tuberculous possums with no gross lesions at necropsy

Comparisons between tuberculous and non-tuberculous possums

Comparison of distribution of lesions in lymph nodes and lymphocentres by sex

Comparisons between distributions of gross lesions in possums in this series of studies and previous studies

DISCUSSION

ACKNOWLEDGMENTS

xii

CHAPTER 6 Naturally occurring tuberculosis caused by Mycobacterium bovis in brushtail possums (Trichosurus vulpecula): III Routes of infection and excretion

ABSTRACT

INTRODUCTION

MATERIAL AND METHODS

Bacteriology

RESULTS

Recovery of M. bovis from tracheal washings, urine, faeces and pouch young of tuberculous possums

Occurrence of fistulae draining tuberculous lymph nodes to the exterior

Disease characteristics in tuberculous possums with lesion distributions consistent with early stage disease

Lesion distributions in tonsils, deep cervical lymph nodes and gastric and mesenteric lymphocentres (non-terminally ill possums only)

DISCUSSION

ACKNOWLEDGMENTS

CHAPTER 7 Serological tests for the diagnosis of tuberculosis in possums: Evaluation of three immunosorbent assays

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

Collection of blood samples

Evaluation

xiii

RESULTS

Test evaluation using cutoffs derived from mean plus 2.57 x standard deviation values Evaluation of agreement between tests using Kappa

Evaluation of agreement between tests using Receiver Operating Characteristic curves Correlations between test absorbance indexes

Lesion frequencies and assay test results

Application of the BLOCK assay to sera from the Castlepoint longitudinal study Application of the MPB70 assay to sera from the Castlepoint longitudinal study

DISCUSSION

ACKNOWLEDGMENTS

CHAPTER 8 A longitudinal study of tuberculosis in possums and cattle

INTRODUCTION

MATERIALS AND METHODS

Data analysis

POSSUM ECOLOGY

Trapping statistics Reproduction

Population dynamics General body condition Denning

Immigration and a comparison of known locally recruited possums and immigrants Dispersal

TUBERCULOSIS EPIDEMIOLOGY

Descriptive epidemiology

Pathology observations

Survival of possums

xiv

Epidemiological analysis based on restriction endonuclease patterns of

TUBERCULOSIS IN OTHER ANIMALS AT CASTLEPOINT

Cattle

Goats Sheep Ferrets Pigs

DISCUSSION

TUBERCULOSIS EPIDEMIOLOGY Prevalence and incidence

Disease occurrence in mothers and their offspring Age and sex distribution of disease

Time of death or disease for different categories of possums Temporal dynamics of the disease

Spatial dynamics of the disease Tuberculosis in cattle at the study site

CHAPTER 9 General Discussion

GENERAL DISCUSSION

Stages of tuberculosis in possums

Modes of transmission of tuberculosis among possums

Opportunities for transmission of tuberculosis among possum other than by the pseudovertical mode

A summary of hypotheses about transmission of tuberculosis between possums Comparisons between Australia and New Zealand

Bibliography

Appendix

xv

TECHNIQUE FOR POST-MORTEM EXAMINATION OF POSSUMS FOR THE DETECTION OF TUBERCULOSIS

External examination Internal examination

Macroscopic appearance of tuberculous lesions

Collections of specimens for subsequent culture for M. bovis Collection of specimens for subsequent histopathology Facilities for autopsies

Equipment

Disposal of carcasses and disposable equipment

Protection of operators engaged in handling tuberculous possums and tissues

xvi

List of Tables

Table 2.1. Number of cattle reactors , lesion-non-tested cattle, and movement control (MC herds with cattle reactor incidence rates (%) for STCAs and Surveillance Areas of New Zealand for the testing seasons 1985/6 to 1992/93.

Table 4.1. Numbers of positive, negative and contaminated culture test results from ribbons replicated at each of 3 sites on pasture, a forest floor and in dens and number of samples not tested

Table 4.2. Group medians, means and Log Rank statistics from comparison of survival probabilities of M.bovis organisms on pasture, forest floor and in dens over 4 seasons calculated using 7-day test results as first measurements

Table 4.3. Group medians, means and Log Rank statistics from comparison of survival probabilities of M.bovis organisms on pasture, forest floor and in dens during spring summer and winter using 4-day test results as first measurements

Table 4.4. Log Rank statistics from comparisons of between season survival probabilities of M.bovis organisms in dens using 7-day test results as first measurements

Table 4.5. Log Rank statistics from comparisons of between season survival probabilities of M.bovis organisms in dens calculated using 4-day test results as first measurements

Table 4.6. Cox's proportional hazard regression model for survival of M.bovis Table 5.1. Summary of prevalences from field surveys

Table 5.2. Prevalences of gross lesions and gross and microscopic lesions at body sites in 73 tuberculous possums with gross or microscopic lesions of tuberculosis

Table 5.3. Prevalences of gross lesions and gross plus microscopic lesions at grouped anatomical sites in 73 possums with lesions of tuberculosis

Table 5.4. McNemar's Chi-squared test values for symmetry of lesion distributions

Table 5.5. Summary results from initial simple Poisson regression screening analyses for number of lesions per individual

Table 5.6. Unweighted Poisson regression of number of lesions per individual for 73 cases Table 5.7. Analysis of deviance for goodness of fit in predicting lesion numbers

Table 5.8. Relative risk values for associations between response and design variables in predicting number of lesions per individual

Table 5.10. Results of histopathology and culture tests for tuberculosis carried out on necropsy negative (NN) possums

Table 5.11. Summary results from initial logistic regression screening analyses for presence of tuberculosis Table 5.12. Comparison of frequency of gross lesion occurrence in studies reported here with values from

previously reported studies

xvii

Table 6.1. Results of culture tests for M.bovis from tracheal washings, urine and faeces of tuberculous possums

Table 6.2. Summary statistics for frequency of gross plus microscopic lesions per individual in 71 possums with and without discharging fistulae

Table 6.3. Summary statistics for number of lung lobes containing lesions in individual possums with and without discharging fistulae

Table 6.4. Distributions of gross plus microscopic lesion sites in possums in which four or fewer lesion sites were detected

Table 7.1. Cross-sectional study sera tested. Table showing numbers of sera from possums with positive diagnoses of tuberculosis and the origins of 251 sera classified by study location and diagnostic criteria used for postmortem examination

Table 7.2. Absorbance index means, standard deviations and cut-off points derived from tests on sera from a non-diseased possum population in Northland with 95% confidence intervals shown in parentheses Table 7.3. Summary test results from possums for which the diagnostic criterion was detailed necropsy using

cutoff points calculated from tests on sera from a non-diseased possum population in Northland Table 7.4. Test agreement between CF and MPB70 tests for the sample of 119 possums using cut-off points

derived from a non-diseased possum population in Northland

Table 7.5. Comparison of areas under the curves for CF, MPB70 and Block assays

Table 7.6. Summary of test results from possums for which the diagnostic criterion was detailed necropsy using cut-off points derived from ROC curves at the point of the lowest index value with a corresponding specificity equal to 1.0

Table 7.7. Correlations between test absorbance indexes for tuberculous and non-tuberculous possums diagnosed by detailed necropsy

Table 7.8. Summary statistics for frequency of gross and gross plus microscopic lesion sites per individual possum tested by MPB70 and CF ELISAs

Table 7.9. Summary statistics for frequency of gross and gross plus microscopic lesion sites per individual possum tested by the BLOCK assay

Table 7.10. Results from Wilcoxon rank-sum tests of equality of medians of numbers of gross and numbers of gross plus microscopic lesion sites in tuberculous possums categorised by positive and negative test results

Table 7.11. Testing histories of nine tuberculous possums which were identified as positive by the BLOCK assay

Table 7.12. Testing histories of eight confirmed tuberculous possums which were identified as positive by the MPB70 ELISA

Table 8.1. Number of possums caught in yearly time periods between February 1990 and January 1993 classified by sex and maturity

Table 8.2. Number of rearing episodes per possum for 157 individual female possums over a 52 month period

xviii

Table 8.3. Summary Jolly-Seber statistics for survival probability between successive visits, population size and immigration plus births for each month from and including visits 5 to 50.

Table 8.4. Survival functions S (t) for non-tuberculous possums which remained in the study ( N = 76 failed, 205 censored), non-tuberculous possums which disappeared (N = 243 failed) and tuberculous possums ( N = 43 failed, one censored)

xix

List of Figures

Figure 2.1. Map of New Zealand showing areas endemic for Tb and Special Tuberculosis Control Areas (September 1991) courtesy P. Livingstone

Figure 2.2. Areas of Endemic Tb (shaded black) in New Zealand 1995

Figure 3.1.a-b. (a) Superficial body regions from which lymph drains directly to 1, parotid lymph nodes; 2, deep axillary lymph nodes; 3, inguinal lymphocentres; 4, superficial axillary lymph nodes; 5, mandibular lymph nodes. (b) I, Superficial body regions from which lymph drains directly or indirectly to the superficial cervical lymph nodes; II, Superficial body regions from which lymph drains directly or indirectly to the deep axillary lymph nodes.

Figure 3.2. Diagrammatic representation of the superficial lymph nodes and the deep cervical lymphocentre and their efferent pathways in the brushtail possum.

Figure 4.1. Survival probabilities of M. bovis organisms on pasture, forest floor and in dens, calculated using aggregated data from spring, summer autumn and winter, with 7-day test results as first measurements.

Figure 4.2. Survival probabilities of M. bovis organisms on pasture, forest floor and in dens during spring, summer and winter, calculated using 4-day test results as first measurements.

Figure 4.3. Survival probabilities of M. bovis organisms on forest floor during winter, spring and summer calculated using 4-day test results as first measurements.

Figure 4.4. Survival probabilities of M. bovis organisms in dens during autumn, winter, spring and summer calculated using 7-day test results as first measurement data.

Figure 4.5. Survival probabilities of M. bovis organisms in dens during winter, spring and summer calculated using 4-day test results as first measurement data.

Figure 4.6. Daily minimum temperatures on pasture, forest floor and in dens during spring (4.6a), summer (4.6c) and winter (4.6b) and daily mean temperatures at the study site at those times (4.6d).

Figure 5.1. Frequency of number of sites containing gross lesions per individual in 73 tuberculous possums.

Figure 5.2. Frequency of number of sites containing gross and/or microscopic lesions per individual in 73 tuberculous possums.

Figure 5.3. Box plot of distributions of gross lesions of cross-sectional and terminally ill groups of possums.

Figure 5.4. Box plot of distributions of gross plus microscopic lesions of cross-sectional and terminally-ill groups of possums.

Figure 6.1. Box plot of number of gross plus microscopic lesion sites in tuberculous possums with and without discharging fistulae

Figure 6.2. Box plot of number of lung lobes with lesions in tuberculous possums with and without discharging fistulae

Figure 7.1. Receiver-operating characteristic curves for Culture Filtrate, BLOCK and MPB70 assays

xx

Figure 7.2a Scatter plot of log10MPB70 indexes and log10 CF indexes of tuberculous and non-tuberculous possums

Figure 7.2b Scatterplot of log10MPB70 and log10CF indexes in non-tuberculous possums. r = +0.34 Figure 7.2c. Scatterplot of log10MPB70 and log10CF indexes in tuberculous possums.

r = +0.71

Figure 7.3. Histograms showing frequencies of positive and negative CF ELISA sera categorised by number of gross lesion sites per individual

Figure 7.4. Histograms showing frequencies of negative and positive CF ELISA sera categorised by the number of gross plus microscopic (total) lesion sites per individual

Figure 7.5. Histograms showing frequencies of positive and negative MPB70 ELISA sera categorised by the number of gross lesions sites per individual

Figure 7.6. Histograms of frequencies of negative and positive MPB70 ELISA sera categorised by the number of total lesion sites per individual

Figure 7.7. Histograms showing frequencies of positive and negative Block assay sera categorised by the number of gross lesions per individual.

Figure 7.8. Histograms of frequencies of positive and negative Block assay sera categorised by the number of total lesions per individual

Plate 8.1 The middle region of the northern side of the study site where possum density and tuberculosis prevalence was high

Plate 8.2 The manuka clad souther side of the study site Figure 8.1. Trapcatch statistics

Figure 8.2. Individuals captured and individuals clinically examined at monthly visits

Figure 8.3. Relative frequency of ages of female and male possums for which death was recorded.

Figure 8.4. Temporal distributions of births and periods of rearing pouch young in possums

Figure 8.5. Relative frequency of periods between successive births measured in 30 day interval periods Figure 8.6. Changes in mean bodyweight and mean testicle size following independence in male possums

identified as pouch young

Figure 8.7. Distributions of the proportion of immature possums to mature possums in the catch of new possums each month aggregated over four years

Figure 8.8. Temporal dynamics of Jolly-Seber population parameters for the possum population from visit number 5 to visit number 50

Figure 8.9. Temporal dynamics of Jolly-Seber population parameters for the possum population from visit 5 to visit 50 showing the relationship between immigration/births and disappearance

Figure 8.10. Temporal pattern of average body weights of mature male and female and immature possums over 52 months.

xxi

Figure 8.11. Temporal patterns of body weight of adult male and female possums (note restricted range of values shown on the Y axis)

Figure 8.12. Scatterplot of den site tracking effort (N = 818 occasions) and the number of different dens (N = 595 den sites) used by individual possums

Figure 8.13. Captures of new possums stratified by age groups

Figure 8.14. Aggregated 36 month data showing the number of months for which possums were captured after initial capture depending on month of capture

Figure 8.15. New cases of tuberculosis over the first 45 months of the study, stratified by age and sex (no new cases were recorded between visits 46 and 52)

Figure 8.16a. Temporal distribution of incident cases of tuberculosis in mature females without pouch young present

Figure 8.16b. Temporal distribution of incident cases of tuberculosis in mature females with pouch young present

Figure 8.17. Estimated survivor functions for possums which died from misadventure and possums which were not lost to follow-up

Figure 8.18. Kaplan-Meier survivor functions for infected and non-infected possums

Figure 8.19. Estimated survivor function curves of groups of possums stratified by sex and tuberculosis infection status

Figure 8.20. Incidence and prevalence of tuberculosis in possums at Castlepoint from April 1989 to July 1993 Figure 8.21a. Average monthly point prevalences calculated from data from all 52 months of the study period

under consideration

Figure 8.21b. Average monthly point prevalences for males and females calculated from data for all 52 months of the study, using the numbers of male and female possums clinically examined at each visit for calculation of the denominators

Figure 8.22a. Average monthly cumulative incidences calculated from the 52 months of the period under consideration

Figure 8.22b. Average monthly cumulative incidences for males and females calculated from data for all 52 months of the study period

Figure 8.22c. Number of mature male and mature female incident cases of tuberculosis at each month Figure 8.22d. Number of total male and total female incident cases of tuberculosis at each month Figure 8.23. Temporal distribution of restriction endonuclease types of Mycobacterium bovis isolates Figure 8.24a. Trap sites at which tuberculous possums were caught during the study period

Figure 8.24b. Trap sites at which tuberculous possums were never caught during the study period

xxii

Figure 8.25. Spatial distribution of capture sites plus den sites used by tuberculous possums infected with particular restriction endonuclease types of Mycobacterium bovis isolates, based on capture site data taken from the period of four months prior to time of diagnosis of tuberculosis to time of death

Figure 8.26. Spatial distribution of restriction endonuclease types of Mycobacterium bovis isolates based solely on capture site data from four months prior to time of diagnosis of tuberculosis to time of death Figure 8.27. Spatial and temporal distribution of restriction endonuclease Type 4 over 52 months of the study

based on locations of den sites used by possums infected with that type

Figure 8.28. Spatial and temporal distribution of restriction endonuclease Type 4a over 52 months of the study based on locations of den sites used by possums infected with that type

Figure 8.29. Spatial and temporal distribution of restriction endonuclease Type 4b over 52 months of the study based on locations of den sites used by possums infected with that type

Figure 8.30 Spatial and temporal distribution of restriction endonuclease Type 10 over 52 months of the study based on locations of den sites used by possums infected with that type

CHAPTER 1

Introduction

2

INTRODUCTION

In 1989, as part of a national integrated effort, a series of epidemiological studies under the direction of Professor R.S. Morris at Massey University, Palmerston North, was undertaken to define and clarify the role of the brushtail possum in the problem of tuberculosis in animals in New Zealand to enable that information to be used to design control methods which would reliably reduce the prevalence of tuberculosis in cattle and deer. At that time it was not known whether the possum was a true reservoir host, and its relationship with tuberculous feral pigs, ferrets and cats and their uncertain role in the maintenance of disease in domestic stock was regularly questioned by farmers, pest managers and scientists. At that time it was clear that possum pest control programmes were inadequate to eradicate tuberculosis in possums and there was a recognition that new and improved control programmes were required.

It was not known how cattle acquired the disease from possums. There were many theories but it was most commonly thought that cattle became infected through grazing pastures contaminated by tuberculous possums, particularly at the forest pasture margins, where prevalence of tuberculosis in possums was thought to be highest. No longitudinal studies had been made and there was consequently no knowledge of the temporal patterns of the disease in possums. The course of the disease was thought to be rapid and the host response poor and inadequate. Experimental studies had been carried out but knowledge of the naturally occurring disease relied on case studies which has used either non-random case selection and/or non-standardised necropsy procedures. The interpretation of pathological findings suffered because the gross anatomy of the lymphocentres and lymphatic pathways had not been described.

Within the general scientific community there was a mediocre understanding of tuberculosis, many features of which had been established by researchers during the first half of the century. As was the case in human medicine, measures which had given acceptable initial progress in tuberculosis control had led to a sense of complacency, although in the absence of problems, there had been no real requirement to have a high level of understanding.

The Massey University studies were part of an integrated national and international scientific effort to formulate short, medium and long term control programmes. The series of studies described here were a natural progression from the longitudinal study at Castlepoint designed by Pfeiffer (1994), and a part of this thesis describes the continuation of that longitudinal study.

Studies of the many complex issues of tuberculosis have seldom produced clear-cut conclusions and have commonly led to controversy. Wildlife disease has been described by Morris (1995) as the

3

and calculating the usual indices used to characterise the epidemiology of a disease. Thus the study of tuberculosis in a wild species presents a particularly daunting challenge, not only from the inherent difficulties associated with tuberculosis and dealing with uncontrolled wild animal populations, but also from acceptance from the general and scientific community that many conclusions will necessarily be derived by inference from multiple sources and from consideration of established general principles of host response to infection with Mycobacterium bovis. Where possible within this thesis, conclusions and inferences have been drawn following assessment using epidemiological techniques which tested the goodness of fit of particular explanations, and there has been a strong underlying endeavour throughout to make deductions which have application to control programmes.

4

CHAPTER 2

Tuberculosis caused by

Mycobacterium bovis in New Zealand

5

PROGRESS AND CURRENT STATUS

public health concern about the sale of milk from tuberculous cattle in town supply herds and was accompanied by legislation which directed that test positive cattle be slaughtered. The scheme was voluntary at its sta rt but progressed to compulsory participation with all town supply herds under test by the

end of the 1950s. In 1961, the scheme was extended to the manufacturing dairy industry which supplies milk for the export trade in dairy products. Beef cattle test ing on a national basis started in 1967, and by

1977, all breeding cattle in New Zealand were under test and a surveillance system was in place at slaughter houses for all beef animals. The dairy factory supply and beef cattle control schemes were under taken primarily with the aim of

ensuring continued access to competitive international markets but it is now recognised that changes in consumer perceptions about the wholesomeness and safety of New Zealand meat have the potential to pose even greater thre ats to profitable marketing. The total trade at risk is:

Dairy $3.42 billion

Beef $1.10 billion

Venison $0.12 billion

Velvet $0.06 billion

Total $4.70 billion source Livingstone (1995)

Initial progress in controlling the disease was rapid, altho ugh in some areas prevalence was very high, with

up to 60% of herds infected. By 1979

- 80,

the percentage reactors had reduced from an initial 8.6% to

0.05% in dairy cattle and from 1.5% to 0.1% in beef cattle. Since that time, there has been little posit ive

progress in the reduction of incidence of reactors. The continued effort has only held reactor incidence steady and has had an undetermined effect of reducing the rate of dissemination and expansion of endemic areas. The reason for the continued lack of progress is the widespread established Australian brushtail

tuberculous possum ( Trichosurus vulpecula ) populations and the reservoir status of possums for the disease.

6

The New Zealand population size was estimated to be approximately 70,000,000 by Bat cheler and Cowan

(1988).For reasons of control, Ministry of Agriculture and Fisheries (MAF) authorities have categorised New Zealand into either Special Tuberculosis Control Areas, (STCAs), in which tuberculosis is endemic, or Surveillance Areas, where t he disease is not endemic. An STCA has a central endemic zone, surrounded

bya "fringe"zone,whichisinturnsurroundedbya non

- endemic

zone. Eachzoneisdefined

geographicallyandtakessecurityaffordedbylocalterrainintoaccount. Thecentral e ndemiczone

comprises a region containing known tuberculous possums, the fringe zone covers an area considered wide enough to contain tuberculous possums which might migrate into it, while the surrounding non

- endemic

zone provides further confidence of con tainment. An area is declared endemic when tuberculosis is found

in wild

- life

within that area. The non

- endemic

areas contain minor endemic areas which are well defined

geographically and these areas are termed Special Tuberculosis Investigation Areas (S TIAs) (Livingstone,

1992). It is conceded that the disease cannot be eliminated from STCAs at this stage but eradication of disease from STIAs is considered technically feasible and is currently underway in some of these areas.The number of endemic area s increased from 9 in 1980 to 23 at present, (6 STCAs and 17 STIAs), but

more importantly, the area of New Zealand which is classified as endemic increased from 10% in 1980 to about 24% at present and contains approximately 12% of all herds. The endemic a reas as at 1992 are

shown in Figure 2.1 (reproduced from Livingstone, 1992) and the STCAs as at 1995 in Figure 2.2 (reproduced from Animal Health Board information sheet).For the most part, cattle and farmed deer contract tuberculosis after direct contac t with tuberculous farmed

animals or wildlife reservoirs. Of the non

- farmed

species, the possum is currently considered to be by far

the most important. The slow progress in reducing the incidence of disease is evident in Table 2.1 (compiled from Livingst one, 1992, and Chief Veterinary Officer, 1992; 1994).

7

Table 2.1. Number of cattle reactors , lesion-non-tested cattle, and movement

control (MC herds with cattle reactor incidence rates (%) for STCAs and

Surveillance Areas of New Zealand for the testing seasons 1985/6 to 1992/93.

8

Summary data 85/8 6

86/8 7

87/8 8

88/8 9

89/9 0

90/9 1

91/9 2

92/93

No of reactors 5768 5922 5940 6183

Lesion-non- tested

data not reported in same format

712 871 689 831

No of M C Herds

as in later years 1231 1292 1314 1383

Reactor incidence %

STCAs 0.43

2

0.38 6

0.41 8

0.42 4

0.35 7

n.a. n.a. n.a.

Surveillance 0.04 8

0.04 4

0.07 4

0.03 1

0.06 9

n.a. n.a. n.a.

New Zealand 0.23 2

0.20 8

0.23 9

0.22 2

0.23 9

0.21 0.20 0.18

n.a. = not available

Lesion-non-tested refer to animals detected at slaughter. These animals are not

included in the calculation of reactor incidence.

test under cattle of No.

ed slaughter reactors

tuberculin of

= No.

incidence Reactor

9

Reactorincidenceisthebestavailablemeasureoftherateatwhichcattleare becominginfected.

Calculation of a true rate is difficult because not all cattle are tested each year and some herds are tested more frequently than once annually.Testing regimesThe basic testing regimes differ between STCAs and Surveillance areas. In STCAs, all stock older than 6

to 12 months are tested annually with the caudal fold intradermal tuberculin test and for test interpretation any visible or palpable injection site swelling is classed as a reactor. In Surveillance areas, breeding stock from 12 to 24 months of age are tested every 2 to 3 years and only injection site swellings which are greater than 4 mm are considered reactors. In large herds in Surveillance areas, a sample of up to 250 animals is considered sufficient to determine herd status. Accredited deer herds in Surveillance areas are

tested every two years. Cattle in endemic areas or from movement control herds are tested prior to and after movement to new areas.When reactors are detected, the frequency of testing is increase d. The time periods for the steps in moving

from infected herd status to free status are subject to change, but generally follow the following pattern: Initially an infected herd is classed as MCI (Movement Control Infected). If the following test is cl ear the

herd is classed as MCC (Movement Control Clear). One more clear test gives F (Free) status. Two F tests are required before A (Accreditation) status.The present effort is maintaining the status quo as far as reactor incidence rates are concerne d, is almost

certainlyhavingsomesuppressiveeffectontherate ofextensionstoexistingendemicareasandis inhibiting creation of new endemic areas. However, the real concern is that the total area of endemic areas is increasing and has increased st eadily over the past 20 years, although part of the increase in size may

have been due to better definition of previously poorly defined areas.Both herds and individual cattle are at a higher risk of exposure to M. bovis in endemic areas than in non -

ende mic surveillance areas. The probability of a herd becoming infected in endemic, fringe and STIAs is

about 16, six and eight times respectively greater than the risk for becoming infected in surveillance areas. Eight - eight

percent of slaughtered reactor c attle and 93% of MC herds came from STCAs in 1989/90

(Livingstone, 1992), indicating that endemic areas were well defined geographically.

10

For an area to be classed as endemic, tuberculosis must be confirmed in wildlife, but detection of disease in possu ms or other wildlife relies on time

- consuming

and expensive surveys involving killing and necropsy.

The efficacy and reliability of this method of detection has not been determined and there is always concern that negative results may be false. A large n umber of possums need to be examined in surveys to detect

disease if prevalence of gross lesions at autopsy is low. If the prevalence is 1% and diseased animals are randomly distributed, approximately 300 possums need to be examined to be 95% certain of de tecting one

diseased animal from a large population. A larger sample is required for examination if the disease occurs in clusters of infected animals, a situation which applies to tuberculosis in possums (Coleman, 1988; Hickling, 1991; Pfeiffer, 1994).

T hepopulations. The first possibility involves spread from existing MC herds, untested tuberculous herds or tuberculous herds which have not been recent status of Surveillance areas is at risk from two possible methods of spread of infection into possum ly tested and have accredited status Past experience of dealing with

tuberculosis in cattle herds suggests that the risk of spread from the above sources is very low. There is generalagreementthatthetransmissionofinfectionfromcattletopossumsis a rareoccurrence.

Maintenance of the prevalence of disease at a low level through regular testing with slaughter of reactors reduces the probability of such an occurrence even further (Morris et al ., 1994).

The present level of testing appears adequate to detect herd infection and natural control is probably

assisted by husbandry and ecological factors within the surveillance area which act as natural barriers to disease establishment. This combination of a low probability for transmission from cattle to possums, early

detection systems for disease in cattle and existing natural ecological barriers reduces the risk of areas becoming endemic.When an unusually high prevalence of disease is detected in a herd, attention is directed to neighbouring herds and the animal population within that area, for early containment and resolution of the problem.

Deer herds with high prevalence of tuberculosis are known to have established endemic disease in some locations and such herds are given special attention. F ollowing detection of high risk herds, an early

decision is usually made as to whether total depopulation of the affected herd is warranted along with

11

vigorous possum control. A compulsory testing and slaughter policy for deer was introduced about 1989

an d the likelihood of transmission of tuberculosis from farmed deer to possums should now be very low.

The more likely and more serious method of spread involves migration of tuberculous possums from endemic areas. This poses the more serious risk because of difficulty of detection at an early stage and

difficulty for containment. If a disease outbreak in farmed animals comes from introduced diseased farm animals, it is normally detected at a relatively early stage during routine surveillance testing. If on the other

hand, possum dispersal from an endemic area introduces disease into local possum populations, detection may be delayed, and meanwhile the infected possum population may expand further. Infected possums probably need to come into close contact with susceptible farm animals for successful direct transmission

and such contact probably occurs relatively uncommonly (Paterson, 1993).During a lag phase, infected dispersing possums have opportunities to establish new foci of infection at the limits of their dispersal ranges. A further delay in detection in local cattle at new foci means that an

infected possum front may be well in advance of recognised disease in cattle.Containment of infected possums is achieved by creating poison buffer zones (fr inge areas). The widths of

buffer zones are set sufficiently wide to contain dispersing animals. Livingstone (1988) estimates that the average yearly spread over a 7 year period in the vicinity of Mangakino, into the Tirohanga dairying area, and through the pine plantations into the Tutakau Road area, was approximately 5 km per year.

Animal Health Board strategic planA strategic plan for 1993 to 1998 dated October 1992, was drafted by the Animal Health Board (AHB). It proposedintensificationofpossum controlinandaroundendemicareasandsetoutthefollowing

objectives:(a) In endemic areas to reduce the percentage of MC herds (deer and cattle combined) by 30

- 50%

and the number of reactors by 50 - 70%.

(b) To reduce the percentage of MC herd s in the non

- endemic

areas to 0.2% (the internationally

recognised level for official freedom from tuberculosis).

(c) To prevent the establishment of new endemic areas and expansion of existing endemic areas into farmland free of feral/wild vectors.

12

(d) To encourage individual farmers to take responsibility for the control of tuberculosis within their herds.

Only one non

- endemic

region, viz. Bay of Plenty, had achieved the 0.2% status at January 1992. In the

Central North Island endemic area, 21.7% of cattle herds and 27.2% of deer herds were on movement

control at January 1992 compared with 1.95% of cattle herds and 2% of deer herds in the non

- endemic

regions.Control costs and fundingThe possum has an enhanced pest status because of its role as a re servoir and vector of bovine tuberculosis

and deciding the best way to deal with this pest presents a real dilemma to control agencies. Livingstone (1991) estimated that approximately $700 million would be needed to be spent over the next 20 years using

c urrent

control techniques to eradicate tuberculosis from possums. He also conceded that this was not a viable option on the grounds of cost, the uncertainty of being able to achieve a successful outcome and heightening public concern over the use of 1080 poison.

The Animal Health Board budget for 1994

- 95 proposed an expenditure of $36.1 million, estimated to rise

to $43.4 million for 2000/01, allocated as follows (Animal Health Board, 1995): Expenditure 1994/95 2000/01

Reactor compensation 1.8 1.8

Testing/epidemiology 9.3 9.3

Scheme management 4.6 4.9

Research* 1.8 2.0

Information/education 1.1 2.1

Vector control 17.5 23.3

TOTAL (millions) $36.1 $43.4

* AHB funded research only. Total funding for tuberculosis related research is $12

million per annum, of which 75% is funded by the Crown.

13

Funding for 1994/95 will be derived from: Industry levies, grants and sponsorships 21.5

Landowners/regions 3.5

Government 9.8

Industry (direct payments) 11.9

TOTAL (millions) $46.7

Farmer monetary input into possum control through a levy on all cattle slaughtered is substantial and has increased dramatically over the past few years. The Animal Health Board has taken responsibility for co -

ordinating the control effort by co ntrol of spending and policy making, although until recently it lacked full

legal status. The Biosecurity Act 1994 gave management powers to the AHB to enable it to levy funds, develop and implement strategies for the eradication of M. bovis tuberculosis. The Biosecurity Act

requires a consultative approach to pest (includes M. bovis ) management be followed, directs responsibility

for control programmes to those involved, and requires that the benefits of implementing the management strategy must outweigh the costs. Currently the Animal Health Board is conducting a consultative process

to help it decide the best mix of individual and centralised responsibility for tuberculosis control. The Board has conceded that a long term research effort and prolonged control will be necessary to solve the

problem of bovine tuberculosis (Alspach 1991).Some appreciation of the daunting nature of possum control in New Zealand and the limitations of the methods used can be gained on reflection that despite the vigorous control measures which have been in

place for the past 40 years, the possum population has continued to increase and colonisation of new areas has continued.

14

THE CASE FOR TUBERCULOUS POSSUMS BEING A SOURCE OF TUBERCULOSIS FOR CATTLE

Tuberculosis was first recorded in possums caught by a trapper on a farm at the mouth of the Mokihinui River, Buller County, in September 1967 (Ekdahl et al.,1970). Twenty of the 25 possums which he submitted for examination showed gross lesions of tuberculosis. Attention was further focused on the possible role of the possum when, in 1970, there was a serious outbreak of tuberculosis in a dairy herd at Seddonville, just a few miles upstream from the original discovery of tuberculous possums.

The area in which this outbreak occurred supported a large possum population and was bounded by bush on three sides and by the river on the fourth. A post-mortem survey of possums in the vicinity revealed 12% prevalence of the disease. The isolation of these cattle by physical boundaries from other cattle and the presence of a significant level of disease in possums suggested an hypothesis that possums were acting as vectors of the disease. Up to this time, it had been considered that cattle contracted the disease from tuberculous in-contact cattle.

On the basis of these findings it was decided to further investigate the role of the possums as a source of infection. The area in which the outbreak occurred was deliberately destocked for a period of six months. Twenty-nine tuberculin tested negative calves were then introduced to graze the area. Six months later, 26 of these animals reacted to the caudal fold intradermal test (CFT) and of these, 16 exhibited gross lesions of tuberculosis at slaughter. (Davidson, 1976). This was the first real evidence which implicated possums as a source of infection.

Although the disease in possums warranted prolonged and intensive investigation at this early stage, only limited studies were in fact carried out. Consequently, for many years control of the disease was hampered by significant deficiencies in the understanding of tuberculosis in possums. More recently, comprehensive and detailed studies have been undertaken by researchers at Massey University, Wallaceville Animal Research Centre, Landcare and within the Ministry of Agriculture and Fisheries (MAF) to remedy those deficiencies.

15

Ekdahl et al.(1970) and Smith (1972) have reported on the nature of the disease in possums and Smith drew attention in particular to the large numbers of M. bovis organisms which were present in lesions in affected animals. Some experimental infection studies were made by O'Hara et al.(1976) in New Zealand and Corner and Presidente (1981) in Australia, which indicated rapid progression of disease and a presumably limited host response to counteract the infection, leading to the presence of large numbers of organisms in lesions. In summary, these early studies showed the possum to be a susceptible host with a capability of producing large numbers of M. bovis organisms, thus having the potential to be a potent vector of the disease.

There has been a gradual accumulation of evidence during the past 25 years incriminating the possum as the most important source of M. bovis infections in cattle. Indictment of the possum comes not from a single source but from collective evidence from multiple studies.

Evidence from tuberculin testing

Some evidence of the role of possums in maintaining tuberculosis in cattle comes from the results of routine testing and the inability of a test and slaughter policy to eradicate the disease in certain parts of New Zealand. The intra-dermal tuberculin test has been the basis of the tuberculosis eradication scheme which was instituted in New Zealand in the 1950s This test is reliable, it has been used world wide and its use in national control programmes has enabled eradication in most countries where the source of infection has been confined to cattle.

The situation faced on the West Coast of the South Island in the early 1970s had presented a new problem to the Ministry of Agriculture and Fisheries. Close examination of testing records for the region revealed unexpected difficulties being experienced in the testing program which could only be explained by the presence of a source of herd re-infection other than cattle. Conversely, and in line with predictions and expectations, many areas of New Zealand were experiencing no difficulties with eradication.

Unfortunately the problem of herd breakdowns was not long confined to the West Coast region.

Other regions followed. The northern region of the West Coast was the next area to report tuberculosis in possums, while in 1969, the Wairarapa also identified possums as a factor in intractable herd problems. The Western Bays of Lake Taupo, Central North Island followed in 1972 (Batcheler and Cowan, 1988). The situation continued to deteriorate after the primary recognition of the particular problems in these three regions. Other areas in which the disease was initially considered eradicated from cattle have subsequently joined the list of endemic areas. In 1991, there were 19 defined areas wherein tuberculous possums have been found (Livingstone, 1991). Since that

16

report, two additional endemic areas were notified, at Waipawa in Hawkes Bay and Otaki in the Manawatu, while at least two other areas were placed under investigation.

As previously pointed out, cattle in STCAs are more likely to be slaughtered as Tb reactors than cattle in surveillance areas (Livingstone, 1991; Chief Veterinary Officer, 1994). In other words, the risk of cattle becoming infected is much higher in those areas where the presence of tuberculous possums has been confirmed.

Tuberculin testing evidence thus incriminates the possum on two grounds, viz. the inability of a proven test to eradicate the disease and the higher risk of cattle becoming infected in areas containing tuberculous possums.

The incidence of cattle tuberculosis following possum poisoning operations

The hypothesis that tuberculous possums are responsible for cattle tuberculosis has been tested by monitoring the course of events following removal of tuberculous possums by control operations.

The incidence of tuberculosis, i.e. the rate at which new cases appear, reduced after possum poisoning operations. This observation has been made a number of times in different locations (Pannett, 1991;

Hoyle, 1991; Anon, 1986).

Restriction endonuclease typing of M. bovis

Subtle differences in the genetic makeup of M. bovis organisms are uniquely expressed in DNA sequences within genes. Scientists at Wallaceville (Collins and de Lisle, 1984; 1985) developed a technique based on restriction endonuclease analysis of DNA (Rea typing), which enables them to classify M. bovis bacteria into distinct types based on restriction fragment patterns. To date, they have identified in excess of 50 different patterns using this technique (de Lisle, pers. comm.).

Different restriction types were found in the three main endemic areas, viz. the Wairarapa, the Central North Island and the West Coast, indicating that problems in these areas evolved independently.

Specific types have come from both farmed and non-farmed animals in the same localities (Collins et al., 1986; de Lisle et al., 1990). The types of M. bovis found in possums have been found to be the same as those found in cattle and deer in the same locality. This is consistent with a cycle of infection between those species and adds further weight to the possum vector theory.

Persistence of tuberculosis in possum populations

For some time it was questioned whether tuberculosis is self- sustaining in possum populations or

17

that possums living in the ecological circumstances applying in New Zealand, where they occur in great numbers, have attained reservoir status (Morris et al., 1994)

A longitudinal study of naturally occurring tuberculosis in possums in the wild at Castlepoint (Pfeiffer 1994) produced strong evidence to support transmission of infection from mother to offspring and also indicated that transmission between adults was largely confined to local social groupings, covering as little as 2 to 4 ha of den-sites. These findings strongly suggested that spread of disease may be related to behavioural factors associated with breeding activity, and possibly also with tree marking or concurrent or sequential usage of dens, although concurrent den-sharing appeared to be a rare event at Castlepoint. These latter findings point to transmission pathways capable of maintaining the disease in possums independently of cattle.

Analogy with badgers in the U.K. and Eire

Some regions of England and Ireland experience a problem with intractable tuberculosis in cattle which is very similar to the New Zealand problem (Nolan and Wilesmith, 1994). It was realised that a problem existed towards the end of the 1960s, and in 1971 an infected badger was found on a farm in Gloucestershire which had recently experienced an episode of tuberculosis (Muirhead et al., 1974).

Following these findings, a strong web of circumstantial and research evidence was gathered, similar to that presented here, which implicated the badger as a wildlife reservoir of infection. The most conclusive evidence came after two problem areas were depopulated of badgers. This step proved to be successful in preventing re-infection of cattle (MAFF, 1979; Little et al., 1982). The existence of this proven wildlife reservoir in analogous circumstances to brushtail possums adds plausibility to the possum reservoir theory.

Why possums?

Tuberculosis caused by M. bovis occurs in an exceptionally wide variety of warm blooded animals (Francis, 1958; Lepper and Corner, 1983; Thoen et al., 1984; Pritchard, 1988) but none of these species which occur in New Zealand is as abundant as possums, nor does any show the consistent overall prevalences of tuberculosis which are found in possums. No other species presents a similar threat in terms of numbers of infected animals capable of transmitting the infection.

Summary

In summary, the evidence implicating the possum relies on circumstantial evidence from a variety of sources, viz.

(a) the disease occurs in cattle in contact with tuberculous possums

18

(b) possums are susceptible hosts capable of producing large numbers of organisms, qualities which are consistent with the concept of their being a successful reservoir

(c) cattle in STCAs are at a higher risk than cattle in surveillance areas (d) tuberculosis in possum populations is self-sustaining

(e) analogy with badgers

(f) weight of numbers of infected possums, allowing many transmission opportunities.

It may be argued that this evidence is not fully conclusive despite its compelling nature. The argument would be strengthened if the tuberculosis status of possums in non-endemic areas was known and if total removal of infected possums from a region would prevent the recurrence of the disease in cattle. There would be considerable difficulties in obtaining comprehensive evidence regarding the tuberculosis status of possums in non-endemic areas and such an exercise would be an expensive undertaking. The sequence of events following eradication of tuberculosis in possums in

"island" situations is presently being determined (Hoyle, 1991; Livingstone, 1991).

19

THE ROLE OF POSSUMS IN THE EPIDEMIOLOGY OF TUBERCULOSIS IN DEER

Although the reservoir and vector status of possums is well established, the same cannot be said with the same degree of certainty for deer, although it appears highly likely. The part that possums play in the epidemiology of the disease in farmed and wild deer has not been investigated to the same extent as in cattle.

The first case of tuberculosis in red deer was recorded in 1955 (Davidson, 1976), and interest was heightened when a case in a free living wild deer was reported from Mohikinui on the West Coast in 1970, an area where tuberculous possums were known to exist. Although there was a thriving wild venison export industry in the 1970s, the inspection of carcasses in the formative years was superficial and was not designed for the detection of disease. Consequently, there was no satisfactory determination of prevalence in wild deer made at that time.

In 1978, the first case of tuberculosis in a farmed deer was recorded on a property contiguous to forest containing tuberculous possums. The farmed deer on this property had originated from the adjoining forest (Livingstone, 1980, cited by Beatson 1985).

When improved inspection procedures (which included an inspection of viscera and the carcass of wild deer), were introduced through regulations in 1975, the number of notifications of tuberculosis increased. Between 1970 and 1983, 161 isolates of M. bovis were made from wild deer while a further 340 came from farmed deer (de Lisle, 1985). Most of these cases were recorded after 1978 and it is likely that the number of isolates from wild deer is an under-representation. This is because inspection, until recently, did not include the retro-pharyngeal lymph nodes which are the most commonly affected lymph nodes in deer (Livingstone, 1980; Beatson and Hutton, 1981; Wilcockson, 1986; Leeming, 1991). Despite this deficiency, the results clearly show that tuberculosis existed in free living wild deer in certain regions throughout the country. Infected wild deer have originated from the West Coast of the South Island, Southland, Ekatahuna County, the Wairarapa and the Central North Island. The disease has been found in wild red, fallow and sika deer (de Lisle and Havill, 1985).

Widespread trading and movement of deer accompanied by a voluntary tuberculosis testing programme created difficulties for determination of the source of infection for farmed deer. Carter (1991) calculated a very high relative risk of 21 for the probability of deer herds in STCAs going on to movement control compared with herds in surveillance areas. However, the disease in farmed deer

20

may conceivably originate from infected cattle, infected deer, infected premises or infected possums.

The epidemiology should become clearer as infection in farmed animals on individual properties is better controlled and the confounding influence of movement between herds reduced. Control of tuberculosis in farmed deer has progressed satisfactorily.

The epidemiology of tuberculosis in farmed and wild deer is poorly understood but it is reasonable to assume that tuberculous possums may be responsible for infecting deer in the wild and on occasions, farmed deer. It is worthy of note and a matter of some concern that there is some evidence from investigations of disease outbreaks in Southland and the MacKenzie basin which incriminates tuberculous farmed deer as a source of infection for possums (Carter 1988).

21

THE EPIDEMIOLOGY OF TUBERCULOSIS IN POSSUMS

IntroductionBovine tuberculosis was introduced into Australia and New Zealand by tuberculous cattle at the time of European settlement. Although the disease has persisted in cattle in both countries, there is no evidence to suggest that tuberculosis has ever established in wild possum populations in Australia, a situation which

contrasts with the New Zealand experience. Australian veterinary authorities have recently claimed to have eradicated bovine tuberculosis from cattle herds throug hout Australia. Many of the Australian herds share

habitat with possums, but possums have never been implicated as a source of re

- infection

of cattle in that

country, either during or following their successful eradication programme. It is not known fo r how long the disease was present in possums in New Zealand prior to its initial

recognition in September, 1967. This diagnosis was made at Wallaceville Animal Research Centre from a possum trapped at the mouth of the Mohikinui River. Tuberculosis had b een present for some time on the

farm where the tuberculous possums were first discovered and this same farm had experienced a very large increase in incidence of the disease in cattle on the farm about 4 months prior to the diagnosis of tuberculosis in po ssums.

In the neighbourhood over the next two years, the incidence increased in low incidence herds and new cases occurred in herds previously clear of the disease, suggesting that the disease had become established in possums not long before 1967, and ha d then spread relatively slowly among the local population in the

lower Mohikinui valley.On the other hand, there were farmer reports of a disease resembling tuberculosis having been discovered in the Buller and Inangahua area in the mid 1950s. In addit ion, tuberculous possums were found at widely

scattered locations throughout the lower Buller County from late 1968 onwards, and it was considered unlikely that disease could have spread from Mohikinui with such rapidity (Davidson, pers. comm.).In the No rth Island, tuberculosis was first found in a possum on a station in the Wairarapa in 1969, but the

disease had probably been established in possum populations for some time prior to that event. A persistent cattletuberculosisprobleminthe Wainui

- o - mat a valleynearWellingtoninthe mid

- 1960s

hadbeen

attributed to infected wild pigs, but in hindsight may have originated from tuberculous possums. Infected