Mycobacterial lymphadenitis in Western Australia

Tuber&

and Lung Disease (1992) 73.362-361

0 1992 Longman

GroupUK

Ltd

Tubercleand LungDisease

Mycobacterial lymphadenitis in Western Australia S. C. Pang Chest and Tuberculosis

Services, Health Department

of Western Australia, Perth, Australia

S UMMA R Y. The records of 172 patients with culture-positive mycobacterial lymphadenitis in Western Australia between January 1972 and December 1989 inclusive have been reviewed. Of the 118 children under 7 years of age, the disease was caused by kf. tuberculosis in 4%) the M. avium complex in 74 % and M. scrofulaceum in 20%, whereas in the 54 adults aged 15 years and over, the same organisms were responsible for 89%) 2% and 4% respectively of their diseases. Tuberculous (TBC) lymphadenitis affected mainly adult Asian migrants (71%), while non-tuherculous mycobacterial (NTM) lymphadenitis predominantly affected non-Aboriginal Australian children (92%). The two conditions differed significantly in their distribution of disease in the lymph nodes (PC 0.001). Patients with the TBC disease had a longer (P c 0.001) duration of symptoms before diagnosis but less common (P < 0.02) local complications at presentation than those with the NTM disease. The response of TBC lymphadenitis to medical treatment was excellent with no failure or relapse in the 43 patients followed up to 12 months. Total excision was curative for NTM lymphadenitis although in 10% a second excision was needed because of relapse or residual disease. In a selected group of children, the double Mantoux test was shown to have a 79% sensitivity and a 69% specificity in the diagnosis of the NTM disease. Over the last decade, the prevalence of NTM lymphadenitis in Western Australia decreased while that of TBC lymphadenitis remained steady. R &S U Mk

. Les dossiers de 172 malades atteints d’une lymphadenite mycobactkienne avec culture positive et repertories en Australie de 1’Ouest entre janvier 1972 et decembre 1989 inclus ont ete revus. Chez les 118 enfants QgCsde moins de 7 ans, la maladie Ctaitdue a Mycobactetium tuberculosis dans 4% des cas, au complexe M. avium (MAC) dans 74% des cas et a M. scrofuluceum dans 20% des cas, tandis que chez les 54 adultes Pges de 15 ans et plus, les m&mesorganismes ont ete responsables respectivement de 89%, 2% et 4% des cas pathologiques. Une lymphad&tite tuberculeuse (TBC) affectait principalement des migrants asiatiques adultes (71%), tandis que la lymphadenite mycobactkienne non-tuberculeuse (NTM) affectait principalement des enfants australiens nonaborigenes (92%). Les 2 conditions avaient une difference significative dans la distribution de la maladie au niveau des ganglions lymphatiques (P < 0,001). Les malades atteints de TBC avaient une anciennete superieure (P < 0,001) des symptGmes avant diagnostic, mais ils avaient aussi moms communement de complications locales (P < 0,02) a la presentation par comparaison avec ceux atteints de NTM. La reponse de la 1ymphadeniteTBC au traitement medical etait excellente: aucun echec ni rechute parmi les 43 malades suivis jusqu’a 12 mois. L’excision totale a CtCcurative pour la lymphad&tite NTM, bien que dans 10% des cas une deuxieme excision ait ete nkessaire en raison d’une rechute ou d’une maladie residuelle. Dans un groupe dlectionne d’enfants, le double test de Mantoux a montre une sensibilite de 79% et une specificite de 69% pour le diagnostic de la maladie NTM. Au tours de la derniere decemde la prevalence de la 1ymphadCnit.eNTM en Australie de 1’Ouest a dhninue, tandis que celle de la lymphadenite TBC est demeuree co&ante. R ES UM E N . Se revisaron las fichas clfnicas de 172 pacientes que presentaron una linfadenitis micobacteriana

con cultivo positivo, diagnosticada en el oeste de Australia entre enero de 1972 y diciembre de 1989 inclusive. En 118 niiios de menos de 7 aiios de edad, la enfermedad se debia a M. tuberculosis en e14 % de 10s cases, a M. avium (MAC) en el74% y a M. scrofuluceum en el20%, mientras que en 54 adultos de mais de 15 aiios de edad 10s mismos microorganismos eran responsables de 89%) 2 % y 4 % de 10s cases, respectivamente. La linfadenitis tuberculosa (TBC) afectaba principalmente a 10s migrantes asiaticos adultos (71%), mientras que la linfadenitis micobacteriana no tuberculosa (NTM) afectaba principahnente a 10s niiios australianos no aborigenes (92%). Las dos condiciones tenian una diferencia signiflcativa en lo referente a la distribution en 10s ganglios linfaticos

Correspondence to: S. C. Pang, Chest and Tuberculosis Services, Health Department of Western Australia, 17 Murray Street, Perth, Western Australia 6000. 362

Mycobactetial

lymphadenitis

in Western Australia

363

0,001). Los enfermos con enfermedad TBC habian tenido un period0 m8s largo (PcO,OOl) con sintomas antes de1 diagmktico, pero presentaron complicaciones locales con menor frecuencia (P < 0,02) que aqubllos con enfermedad NTM. La respuesta de la linfadenitis TBC al tratamiento con medicamentos era excelente: sin fracasos ni recaidas en 10s 43 pacientes sequidos hasta 12 meses. La eseisidn total fue curativa para la linfadenitis NTM, aunque en el 10% de 10s cases se necesit6 una segunda escisibn, debido a una recaida o a una enfermedad residual. En un grupo seleccionado de niiios, el doble test de Mantoux mostr6 una sensibilidad de 79% y una especificidad de 69% para el diagmktico de la enfermedad NTM. La prevalencia de linfadenitis NTM ha disminuido en el curso de la liltima dCcada en el oeste de Australia, mientras que la prevalencia de linfadenitis TBC ha permanecido constante. (PC

INTRODUCTION Tuberculous (TBC) lymphadenitis has been the commonest extrapulmonary tuberculosis in Australia.‘.* It was reported to make up 9% of tuberculosis notifications in children under 17 years of age in Victoria for the period 1970-l 986” and 11.7% of all tuberculosis notifications in New South Wales in 1986.’ Yet between 1976 and 1985, Joshi et al’ could only find 5 children presenting to the Royal Children’s Hospital in Melbourne with tuberculous as against 86 with non-tuberculous mycobacterial (NTM) lymphadenitis, commenting that lymphadenitis due to M. tuberculosis was rare in the community. Similar findings were reported for Western Australia.6,7 All included culture-negative diseases, and no study on communitywide mycobacterial lymphadenitis in Australia has ever been published. Western Australia with its relative geographical isolation, a single Mycobacteria Reference Laboratory Service and a central Chest and Tuberculosis Control Branch, both under the State Health Department, provides unique conditions for such studies. This paper reports on the I8-year findings between 1972 and 1989 inclusive.

MATERIALS AND METHODS General data All records of patients notified to the Chest and Tuberculosis Services at the Perth Chest Clinic (PCC) for mycobacterial lymphadenitis from January 1972 to December 1989 were reviewed. NTM disease is not notifiable but in Western Australia the State Mycobacteria Reference Laboratory (SMRL), which is solely responsible for all mycobacteriological studies, reports every mycobacterial identification by culture or smear from clinical specimens to the PCC where volunteer notification is encouraged and solicited for surveillance and statistical purposes. Those whose diagnoses were based on histology or clinical grounds, would only be noti tied through the clinicians’ volition. Undernotification of this group was inevitable but the number is estimated to be small because of the clustering of patients at the Princess Margaret Hospital for Children where special interest in the disease was well known.6-s Information about each patient was recorded on

predesigned data collection forms for demographic and clinical details which were then analysed using simple descriptive statistical methods including x’ test for probability values. Children are defined in this study as persons under the age of 15 years. The period of followup assessment was set at 12 months after completing medical or surgical treatment.

Double Mantoux test The tuberculins used for the Mantoux test over this period varied from purified protein derivative (PPD)-B (Battey), PPD-A (avium) to PPD-I (intracellulare) for the non-tuberculous mycobacteria, and from PPD-S (Seibert) to PPD-H (human) for M. tuberculosis. For clarity they are collectively referred to as B and H respectively in the results. The dose was 5 iu for children and 10 iu for adults. To assess the value of the double Mantoux test, ’ the results from the non-Aboriginal Australian children with diseases caused by the M. avium complex or M. scrojiilaceum were analysed. All would not have had BCG vaccinations and were most unlikely to have been simultaneously infected by M. tuberculosis. The sensitivity and specificity were calculated according to the following definitions: I. when the induration from PPD-B is greater than that from PPD-H by 3 mm or more, the double Mantoux test in the patient is categorized as both TP (truepositive, for NTM infection) and TN (true-negative, for TBC infection); 2. when the induration from PPD-B is less than that from PPD-H by 3 mm or more, the test is categorized as both FN (false-negative, for NTM infection) and FP (false-positive, for TBC infection); and 3. when the two indurations are within 3 mm of each other, the test is (i) TN if the PPD-H induration is under 5 mm, FP if over 4 mm, and (ii) FN if the PPD-B induration is under 5 mm, TP if over 4 mm. The cut-off point for significant difference between the indurations in the double test is difficult to determine as there is no uniform or generally accepted standard in the literature.7-‘0 A minimum of 3 mm is adopted here as this probably represents the limit of accuracy from technical and observer errors.“,‘2 In the small number of patients who had the test repeated after an interval, the

364

Tubercle and Lung Disease

results from the second test were considered separately if the categorization had changed and discarded if not. Similar calculations were not applied to the other groups of patients due to the small number of data available.

Table 1. Source of bacteriological Source

RESULTS

Exclusions A total of 274 patients were found to have mycobacterial lymphadenitis during the period and, of these, 102 had culture-negative disease. 95 of the latter had histological evidence of caseating or necrotizing granulomata with or without positive microscopy for acid-fast bacilli. They were excluded from all analyses but included in the histograph (Figure) to indicate the 3-yearly prevalence of the condition in Western Australia. This left a total of 172 patients with culture-positive disease in the study, consisting of 53 with TBC and 119 with NTM lymphadenitis.

diagnosis

No. of patients Tuberculous

Total Non-tuberculous mycobacterial

Excisional biopsy Needle aspiration Sterile swab Sputum or gastric washing

39 7 3 4

112 0 7 0

151 7 10 4

Total

53

119

172

lymphadenitis. The bacteriological diagnosis is shown in Table 1, while the distribution of the different mycobacterial species in the patients is shown in Table 2.

Tuberculosis and non-tuberculous mycobacterial lymphadenitis Certain demographic features and the anatomical sites of the two diseases are presented separately for adults and children in Table 3 to highlight their differences. The results from the two age groups in each disease, however, have been combined for statistical analyses due to the small number of patients in the childhood TBC and adult NTM subgroups. There was an overall male:female ratio of about 2:3 in both conditions. TBC lymphadenitis was mainly an adult infection of the Asian migrants (71%) and NTM lymphadenitis predominantly a childhood condition of the non-Aboriginal Australians (92%). The anatomical distribution of the two diseases in the various lymph nodes was significantly different (P
Diagnosis Of the total of 172 patients, the diagnosis in 168 was confirmed bacteriologically from the infected lymph nodes obtained by excisional biopsies in 151, by needle aspirations in 7 and by sterile swabs obtained at incisional drainage in 10. The remaining 4 patients had positive cultures from sputa or gastric washings for M. tuberculosis, 3 of whom were children with primary pulmonary tuberculosis. The sole adult was a Vietnamese with bacteriologically confirmed tuberculosis in the right upper lobe of the lung and bilateral supraclavicular

50 !I

:

Number 40-

: : / L

30-

: /

:

x

of

Patients

X X X X X X X X

lo-

i 1972-74

Figure - Mycobacterial lymphadenitis in Western Australia culture-negative. NTM lymphadenitis, culture-p0sitive.m

1975-77

X

1978-80

n

X X X

LX

/ A--iJEL X

q

: /

//

i

X X X X X X

1981-83

X X X X X

X

1987-89

1984-86

1972-1989. TBC lymphadenitis, culture-positive. NTM lymphadenitis, culture-negative.

0

TBC lymphadenitis,

Mycobacterial

Table 2. Australia

Bacteriology 1972- I989

Mycobacterial

species

of mycobacterial

lymphadenitis

No. of patients (%) Adults

M. tuberculmia M. avium complex M. scrr$rlucYJrrm M. uvium complex and M. .scr~fulawum M. chdmar M fortuitum M.
48 (88.9%) 1(1.9%) 2 (3.7%)

TOtal

54 (100.2%)

TBC lymphadenitis Adults Children

1 (0.8%) 0 0 I (0.8%) 1 (0.8%)

I18 (99.8%)

Antituberculous therapy was the main form of treatment for TBC lymphadenitis, and 43 of the 53 patients (Table

Duration of symptoms

Complications presentation

at

5 3 2

6 2 3

Ethnic origin Australian (nonAboriginal) Aboriginal Asian migrant+ Non-Asian migrant

0.75-3 2.3

6 0 34 (7 1%) 8

2 2

1.3

15-47 31.0 IS.7

2

1

I I

0

2

Site of disease’ Cervical upper lower unspecified total Axillary Mediastinal Inguinal

11 I6 17 44 (80%) 7 2 2

0 1 1 2 2 4

Multiple/bilateral

10

> 0.70

l-6 2.8 I.1

IO4 (92%) 3

I 5 < 0.00

3 , 0 5

7x 0 71 99 (87.6%)

1

I I 2

I 0 13

3

2

4

I

“As calculated by x2 test between the TBC and NTM groups, iMigrants from South-east Asia and the Indian subcontinent, *Multiple-site disease in some patients.

6) so treated and available for assessment on follow-up had favourable results at 12 months. The majority (24 patients) were treated with 6- or 9-month regimens as generally recommended.“,‘4 A substantial number (18 patients) were given treatment for 12 months and over, mainly because they were diagnosed before short-course chemotherapy for TBC lymphadenitis was shown to be effective.‘> Two patients who stopped treatment at 3 (RH plus pyrazinamide) and 6 months (RH) because of drug intolerance and default respectively remained well when assessed 12 months later. Both had their lymph

and complications

at presentation

Duration of symptoms in months O-3 >3 No data TBC NTM TBC NTM TBC NTM

Total TBC

NTM

Skin inflammation Abscess formation Cutaneous sinus

0 5 2

I4 22 2

0 0

0

0

I

I

I

3 I

I 9 2

0 8 4

15 32 6

Present Absent

7 6

38 37

1 16

2 3

4 19

12 27

12* 41’

52’ 67’

13’

751

17+

5’

23

39

53

Total

II3 46 67 1.1.5

18-83 35.7 16

P value*

NTM lymphadenitis Adults Children

1.1.7

Age in years Range Mean SD

and outcome

Table 4.

48 17 31

No. of patients Male Female Overall M:F ratio

and in particular its predilection for the upper cervical nodes and the absence of primary infection in the lung. No abdominal lymphadenitis from either cause was diagnosed. All children with the NTM disease had normal chest X-rays. Of the 119 patients with NTM lymphadenitis, 113 (95%) were children with a mean age of 2.8 (range l-6) years. Almost 88% of them had the disease in the cervical nodes and invariably the upper (pre-auricular, submental, submandibular and jugulo-digastric). 13 involved the inguinal nodes compared with 1 the axillary. These further support the view of MacKellar’ that, in children at least the portal of entry is either oropharyngeal or through a skin injury in the limb. The duration of symptoms before diagnosis was generally longer but the complications of skin inflammation, abscess formation and cutaneous sinus were less common in the patients with the TBC than in those with the NTM disease (Table 4). Both were statistically significant. The results of the double Mantoux test and its sensitivity and specificity in differentiating the NTM from the TBC disease are given in Table 5.

Treatment

365

Tuberculous (TBC) and non-tuberculous mycobacterial (NTM) lymphadenitis in Western Australia 1972-1989

5 (4.2%) 87 (73.7%) 23 (19.5%)

! (I .9%) I (I .9%) I(l.9%) 0

in Western Australia

Table 3.

in Western

Children

0

lymphadenitis

*The difference between the two groups is significant

(P
‘The difference

(P
between the two groups is significant

119

366

Tubercle and Lung Disease

Table 5.

The double Mantoux test

Table 6. Drug treatment of tuberculous Australia 1972-1989.

Total number of tests where PPD-B induration is greater by >2 mm: 50 Total number of tests where PPD-H induration is greater by >2 mm: 10 Total number of tests where PPD-B and PPD-H indurations differ by <3mm:26 Number of TPs: 68 Number of FPs: 27 Number of TNs: 59 Number of FNs: 18 Sensitivity = TP/(TP+FN) x 100% = 79% Specificity = TN/(TN+FP) x 100% = 69%

lymphadenitis

in Western

Drug regimen

Duration in months 3 6 9

RH(E)-RH RHZ(E)-RH Others

0 1 0

3 0

11 9 0

13 1 4

25 14 4

Total

1

4

20

18

43

R = rifampicin,

I

H = isoniazid, E = ethambutol,

Total 12&over

Z = pyrazinamide.

DISCUSSION nodes excised and this may have contributed to the favourable results. All 119 patients with NTM lymphadenitis had surgical intervention (Table 7). Five had incisional drainage only, and of these, 3 were known to remain well at 12 months. The remaining 114 had total excisions, 15 of which were carried out for recurrence after a previous incisional drainage; second excisions were required in 8 patients because of relapse or residual disease in the same site. Of the total 86 patients known to have a favourable outcome of 12 months, 12 of 15 (80%) who had incisional drainage procedures failed against 7 of 71 (10%) who had total excisions. These findings are similar to those of other reports5-7.‘6X’7and emphasize once again the importance of adequate local excision. Additional antituberculous therapy was given to 6 patients with the NTM disease for periods varying from 3 to 16 months. They all had strongly positive reactions to PPD-H. Three were children, including 1 Aboriginal and 1 Vietnamese. They were treated with rifampicin (R) and isoniazid (H) for 3,4 and 6 months respectively. One adult Australian had disseminated M. chelonae infection with involvement of several groups of lymph nodes, possibly related to her previous corticosteroid therapy for sarcoidosis; she had 16 months of multiple drug treatment including R, H, ethambutol (E) and prothionamide. The other two adults were Asian and European migrants with M. gordonae and the M, avium complex respectively isolated from their affected nodes. The Asian was lost to follow-up but all the other 5 patients had a favourable outcome at 12 months.

Table 7. Surgical treatment and outcome of non-tuberculous mycobacterial lymphadenitis in Western Australia 1972-1989 Surgery

Outcome at 12 months Favourable Unknown

Incisional drainage alone Incisional drainage plus total excision Total excision Second excision required for relapse or residual disease

3 12

2 3

5 15

64

27

91

7

1

8

Total (no. of patients)

86

33

119

Total

The incidence of mycobacterial lymphadenitis in Australia is unknown in spite of the many reports in the literature.5-8’16 This again cannot be estimated in the present study because of the large number of culturenegative cases. The great majority of these were undoubtedly mycobacterial but differentiation into TBC or NTM forms cannot be ascertained. It is nevertheless obvious from the Figure that in Western Australia the prevalence of NTM lymphadenitis has been decreasing over the last decade while that of TBC lymphadenitis has remained unchanged, except for minor fluctuations perhaps reflecting different migrant intakes. The three major causes of mycobacterial lymphadenitis in Western Australia (Table 2) are the M. avium complex, M. tuberculosis and M. scrofidaceum. No case due to M. bovis was recorded for the whole l&year period. This is in contrast to the findings from South-east where 65 of 2272 mycobacterial isolations England” from TBC lymphadenitis were bovine strains. The higher prevalence of the M. avium complex over M. scrofulaceum as a cause of NTM lymphadenitis had been noted6.7 and is similar to experience elsewhere.‘s~19 M. kansasii, M. chelonae and M. fortuitum are known human pathogens but the significance of M. gordonae in our 2 patients is uncertain. Clinically, both were not immunocompromized. The source of non-tuberculous mycobacteria is generally accepted as environmental.20~2’ Various mycobacteria have been isolated from animals, soil, water systems (both natural and artificial) and raw milk in Western Australia.22.23 Transmission in children would be through inserting parts of the contaminated environment into their mouths,” drinking unboiled water or unpasteurised milk, and skin injuries from contaminated objects or animals. The last mode of infection explains the great predominance of inguinal node involvement over the axillary since Australian children often go barefoot in outdoor activities. As in tuberculosis, immunity should play a part in the pathogenesis of the NTM disease and BCG vaccination has been implicated by Wickman to have a protective effect in Sweden. In Western Australia BCG was given to school leavers or first-year high school students between 1949 and 1985.25 Although its effect on NTM

Mycobacterial

diseases in the community cannot be assessed, the overwhelming preponderance of NTM lymphadenitis in the local children of l-6 years old tends to support Wickman’s view. Furthermore, the sharp contrast in the relative proportions of the two conditions between the non-Aboriginal Australians and the Asian migrants suggest probable cross-immunity from not only BCG vaccination but also prior TBC infections similar to that observed for pulmonary M. kansasii disease.26,27 The accuracy of the double Mantoux test in differentiating NTM from TBC infections as observed previously”x 1s not supported by the present study in which the test was shown to have a sensitivity of 79% and a specificity of 69% (Table 5). The selection of the subjects would have excluded any complications from BCG vaccinations or simultaneous tuberculous infections. The method of calculation and the definitions may be subject to debate but the proportion of tests with PPD-H reactions greater than (10/86) or within 3 mm of PPD-B reactions (26/86) would make an alternative conclusion unlikely. The reason for the discrepancy is not clear. Difference in the tuberculins used or in the expertise with the technique* may be implicated but the major factor must be the inherent variability of the test technically, ‘)-“J*-~” in host responses3’ and due to the shared antigens in the PPDs.” Further evaluations with the new tuberculins3’ would be desirable and appropriate.

Acknowledgement The author is grateful Hospital for Children, detailed notification of permission of Dr Peter Western Australia.

to the medical staff of the Princess Margaret Perth, in particular Dr P. L. Masters, for their their patients. This paper is published with the Brennan, Commissioner, Health Department of

References I.

2.

3.

4. 5.

6 7

Dwyer DE, Collignon PJ, MacLeod C, Sorrel1 TC. Extrapulmonary tuberculosis - a continuing problem in Australia. Aust N 2 .I Med 1987; 17: 507-5 11. National Health and Medical Research Council. Epidemiology. In: Tuberculosis in Australia and New Zealand into the 1990s. Canberra: Australian Government Publishing Service 1990; 2: 4-10. Mulholland EK, Gilbert GL, Kempster RE. Childhood tuberculosis in Victoria, 1970-86. Aust Paediatr J 1989; 25: 31-34. Plant AJ, Rushworth RL, Wang Q, Thomas M. Tuberculosis in New South Wales. Med J Aust 1991; 154: 86-89. Joshi W, Davidson PM, Jones PG, Campbell PE, Robetton DM. Non-tuberculous mycobacterial lymphadenitis in children. Eur J Pediatr 1989; 148: 751-754. MacKellar A, Hilton HB, Masters PL. Mycobacterial lymphadenitis in childhood. Arch Dis Child 1967; 42: 7&74. MacKellar A. Diagnosis and management of atypical mycobactetial lymphadenitis in children. J Pediatr Surg 1976; 11: 85-89.

lymphadenitis

in Western Australia

367

8. Masters PL, Smyth JT. A double Mantoux test applied to screening children for mycobacterial infections: its value in distinguishing infections by anonymous strains. Aust Paediatr J 1965; I: 166-175. 9. Johnston WW, Smith DT. MacVandiviere H. New aspects of mycobacterial skin tests. III. Simultaneous or sequential infection with different mycobacteria. Arch Environ Health 1965: I I : 3749. IO. Houk VN, Kent DC, Baker JH, Sorensen K. Comparison of paired tuberculins. Arch Environ Health 1968; 16: 3645. II. Nissen Meyer S. Hougen A, Edwards P. Experimental error in the determination of tuberculin sensitivity. Public Health Rep 195 I: 66: 56 l-569. 12. Carruthers KJM. Observer and experimental variation in tuberculin testing. Tubercle 1970: 5 I :48-67. 13. American Thoracic Society. Treatment of tuberculosis and tuberculosis infection in adults and children. Am Rev Respir Dis 1986; 134: 355-363. 14. Subcommittee of the Joint Tuberculosis Commtttee. Chemotherapy and management of tuberculosis in ihe United Kingdom: recommendations of the Joint Tuberculosis Committee of the British Thoracic Society. Thorax 1990: 45: 403-408. 15. British Thoracic Soctety Research Committee. Short course chemotherapy for tuberculosis of lymph nodes: a controlled trial. Br Med J 1985; 290: 1106-l 108. 16. Jones PG, Campbell PE. Tuberculous lymphadenitis in childhood: the significance of anonymous mycobactetia. Br J Surg 1962: 50: 302-3 14. 17. Harris BH, Webb HW, Wilkinson Jr AH, Santelices AA. Mycobactetial lymphadenitis. J Pediatr Surg 1982: 17: 589-590. 18. Grange JM, Collins C, Yates M. Bacteriological survey of tuberculous lymphadenitis in Southeast England: 1973-80. J Epidemiol Community Health 1982; 36: 157. 16 I, 19. Lai KK. Stottmeier KD, Sherman IH, McCabe WR. Mycobacterial cervical lymphadenopathy relating etiologic agents to age. J Am Med Assoc 1984; 251: 12861288. 20. Grange JM. Other mycobacterial diseases. In: Mycobacteria and human disease. London: Edward Arnold 1988: I38- 152. 21. O’Brien RJ. The epidemiology of nontuberculous mycobacterial disease. In: Snider Jr DE, ed. Clinics in chest medicine. Philadelphia: Saunders 1989: IO: 407418. 22. Health Department of Western Australia. Scientific report of the mycobacteria reference laboratory services for the years 1979 and i982-1984. 23. Dunn BL, Hodgson DJ. ‘Atypical‘ mycobacterin in milk. J Appl Bacterial 1982; 52: 373-376. 24. Wickman K. Clinical significance of nontuberculous mycobacteria: a bacteriological survey of Swedish strains isolated between 1973 and 198 I Stand J Infect Dis 1986: 18: 337-345. 25. Porter RM, Boag TC. Bacillus Calmette Guerin (BCG). In: The Australian Tuberculosis Campaign 1948-1976. Melbourne: Robert Menzies Memorial Foundation 199 1; I6:7 I. 26. Anh CH, Lowell JR, Onstad GD, Shuford EH, Hurst GA. A demographic study of disease due to Myobactrrium kansasii or M. ir~tracellulare-avium in Texas. Chest 1979; 75: 120-123. 27. Pang SC. Mycobuterium kutzsasii infections m Western Australia (1962-1987). RespirMed 1991; 85: 213-218. 28. Loudon RG, Lawson RA Jr, Brown J. Variation in tuberculin test reading. Am Rev Respir Dis 1963: 87: 852-86 I, 29. Bearman JE, Kleinman H, Glyer VV, La Croix OM. A study of variability in tuberculin test readings. Am Rev Respir Dis 1964; 90: 913-919. 30. Chaparas SD, MacVandiviere H. Melvin 1, Koch G, Becker C. Tuberculin test. Variability with the Mantoux Procedure. Am Rev Respir Dis 1985; 132: 175-177. 31. Editorial. New tuberculins. Lancet 1984; i: 199-200. 32. Stanford JL. Grange JM. The meaning and structure of species as applied to mycobacteria. Tubercle 1974: 55: 14.3-l 52.