Tubercle (1989) 70,27-36 0 Longman Group UK Ltd 1989

0041-3879/89/0070-0027/$10.00

SKIN TEST RESPONSIVENESS TO A SERIES OF NEW TUBERCULINS OF CHILDREN LIVING IN THREE VIETNAMESE ClTlES

H. M. Ly, D. D. Trach, H. T. Long, N. K. Thuy, N. A. Tuan

National lnsritute of Hygiene and Epidemiology (N/HE), Hanoi, Vietnam

Tran Thi Ninh

National Institute of Dermatology, Hanoi, Vietnam

J. L. Stanford

School of Pathology, University College and Middlesex Hospital Medical School, London

J. T. Hendriks and E. P. Wright

Department of Medical Microbiology, University of Amsterdam, The Netherlands

Summary A skin test survey was conducted among 1035 children aged 7-19 years living in three cities in Vietnam. Fifteen new tuberculins, including leprosin-A, were applied; an induration of 2 mm diameter or more was considered positive. Compared to some other tropical countries, low levels of sensitisation were recorded and remarkable regional differences were found. Positivity to any tuberculins (pooled data) among non-BCG-vaccinated children was significantly lower in Hanoi (13.1 %I and HoChiMinh-City [HCMCI (15.5 %) than in Nha Trang (25.7 %I [p=O.OOl and p=O.O12, respectively]. The proportion of non-vaccinated children responding to Tuberculin ranged from 18.4 % in Hanoi to 54.5 % in Nha Trang. Leprosin-A elicited a response in 14.9 % of the children in Nha Trang, but in very few of those in Hanoi (4.3 %) or HCMC (3.0 %I. Thus, of the three cities studied, significant sensitisation to both M. tuberculosis and M. leprae was demonstrable only in Nha Trang. In Hanoi most of the response was to fast-growing species whilst in HCMC and Nha Trang it was mainly to slow-growing species. These results may account in part for the observed differences in the prevalence of tuberculosis and leprosy between the north and the south of Vietnam.

R&urn6 Une enquete a ete menee parmi une population de 1035 enfants ages de 7-19 ans, habitant dans trois villes du Vietnam. Quinze nouvelles tuberculines, ainsi que la leprosine A, ont et6 appliquees; une induration de 2 mm de diametre ou plus a ete consideree comme positive. Par rapport a certains autres pays tropicaux, des niveaux faibles de sensibilisation ont en fait 6te enregistres; des differences regionales tres notables ont et6 egalement constatees. Le taux de positivite vis-a-vis

Correspondence and address for reprints: J. T. Hendriks, Department of Medical Microbiology, University of Amsterdam, Meibergdreef 39 1105 AZ Amsterdam, The Netherlands.

28 Ly and others

de quelque tuberculine que ce soit (tous resultats regroup&) parmi les enfants non vaccines par le BCG dtait significativement plus faible a Hanoi (13,l %) et g Ho Chi Minh Ville (15,5 %) par rapport a celui de Nha Trang (25,7 %) (p=O,OOl et p=O,O12 respectivement). La proportion d’enfants non vaccines qui reagissaient 6 la tuberculine allait de 18,4 % a Hanoi 6 54,5 % $I Nha Trang. La leprosine A a provoque une reponse chez 14,9 % des enfants de Nha Trang, mais seulement parmi une petite proportion des enfants de Hanoi (4,3 %) ou de Ho Chi Minh Ville (3,0 %). Done, sur les trois villes etudiees, une sensibilisation significative vis-a-vis des deux microorganismes, M. tuberculosis et M. leprae a 6te demontree seulement a Nha Trang. A Hanoi, la plupart des reponses concernaient les especes a croissance rapide, tandis qu’a Ho Chi Minh Ville et a Nha Trang, il s’agissait principalement d’especes a croissance lente. Ces resultats peuvent expliquer en partie les differences observees en ce qui concerne la prevalence de la tuberculose et de la lepre entre le Nord et le Sud du Vietnam.

Resumen Se realizo una encuesta con tests cut6neos en una poblacion de 1035, nitios de 7 a 19 arias, habitantes de tres ciudades de Vietnam. Se aplicaron 15 nuevas tuberculinas y la leprosina A; se consider6 coma positiva una induration de 2 mm o m&s. Se registraron niveles bajos de sensiblizacion, comparados con 10s de otros paises tropicales; se constataron igualmente diferencias regionales notables. Las tasas de positividad con respect0 a cualquiera de las tuberculinas (resultados agrupados) entre 10s nines no vacunados con BCG eran significativamente mas bajos en Hanoi (13,l %) y ciudad de HoChiMinh (CHCM) (15,5 %) comparades con la de Nha Trang (25,7 %) [p=O,OOl y p=O,O12, respecivamentej. La proportion de nines no vacunados que reaccionaban a la tuberculina iba de 18,4 % en Hanoi a 54,5 % en Nha trang. La leprosina A provoco una respuesta en 14,9 % de 10s nirios de Nha Trang, pero solo en un escaso numero en 10s de Hanoi (4,3 %) o de CHCM (3,O %). Asi, de las tres ciudades estudiadas, se demostrb una sensibilizacion significativa a ambos microorganismos, M. tuberculosis y M. leprae solamente en Nha Trang.

En Hanoi la mayoria de las respuestas eran contra las especies de crecimiento rapido, mientras que en CHCM y Nha Trang, se trataba principalmente de especies de crecimiento lento. Estos resultados pueden explicar en parte, las diferencias observadas con respect0 a la prevalencia de tuberculosis y lepra entre el notte y el sur de Vietnam. .

Introduction The development of immunity to M. tuberculosis and M. leprae and the protective effect of BCG immunisation against the diseases caused by these pathogenic mycobacteria remain poorly understood phenomena. BCG vaccination trials against tuberculosis and leprosy have yielded widely diverging results in different countries [I, 21. For example, BCG vaccination gave 80 % protection from leprosy in Uganda but only 17 % in Burma [I].

To explain these apparent discrepancies, Stanford [I] and Rook [2] postulated two types of cell mediated immune reactions (CMI) in mycobacterial infections, both affected by specific T-lymphocytes but with different outcomes. One type, the “Listeria like” or non- necrotising response, would contribute to immune protection, whereas the other type, the “Koch-like” or necrotising response, would be associated with the development of disease

Skin tests in Vietnam 29

[I I. Recently, a similar model of two functionally antagonistic (i.e. resistance-promoting and disease-promoting) T-helper cell populations has been proposed for murine leishmaniasis [3, 4, 51. Both cell populations can mediate delayed type hypersensitivity (DTH). It was proposed [I] that previous contact with environmental mycobacteria greatly influences the type of CMI reaction that predominates in individuals and influences their susceptibility to subsequent infection with M. tuberculosis or M. leprae. Varying exposure to slow-growing species could enhance either type of immune reaction, whereas exposure to fast-growing species generally favours a protective immune response [I]. Subsequent vaccination with BCG would then merely enhance the already predominating response and would thus be successful in individuals primed for a protective response, but ineffective in persons with predominantly necrotising responses [2]. If this model is correct, skin test surveys with new tuberculins provide valuable information in regions where mycobacterioses occur. Docu- mentation of the prevailing type of response in a population at risk could contribute to the selection of optimal BCG vaccination strategies in an endemic area.

To measure the level of exposure of human populations to environmental mycobacteria, skin test surveys have been carried out in various countries, using new tuberculins [6, 7, 8, 9, IO].

In Vietnam, both tuberculosis and leprosy are still important public health problems, especially in the south, but there is no information available on the mycobacteria present in the environment nor on the sensitisation to mycobacterial antigens in the population. This study reports the results of an investigation of skin test responses to new tuberculins in schoolchildren from three geographically distinct locations in Vietnam (Fig. 1): Hanoi in the north, and HoChiMinh-City (HCMC) and Nha Trang in the south.

Materials and methods Persons tested

1035 schoolchildren from three cities were tested: 278 from Hanoi, 507 from Nha Trang and 250 from HCMC. Their ages ranged from 7 to 19 years (Table I).

Reagents used for skin-testing

15 new tuberculins (NT) prepared from M. leprae and 14 other mycobacteria were used (Table II) [IO]. With the exception of our Leprosy reagent, all of the mycobacterial reagents were prepared in London as previously described Ill].

Preparation of Leprosin-A (Vietnam)

Lyophilized purified M. leprae, kindly provided by Dr. P. Klatser (Royal Tropical Institute, Amsterdam, The Netherlands), was resuspended in sterile phosphate buffer, pH 7.0, and ultrasonicated on ice for 20 minutes. The crude sonicate was then centrifuged at 35.000x g for 30 minutes at 4°C to obtain the supernatant soluble antigen. This supernatant was passed through a series of membrane filters with pore sizes of 0.45 pm and 0.2 ym. The protein concentration of an aliquot was estimated by ultraviolet absorption at 280 nm. The antigen preparation was stored at 4 “C in sterile borate buffer of pH 8.0. For use, this stock was diluted with the same buffer to obtain a final concentration of 10 pg/ml for skin testing. The diluted reagent was again filtered through sterile 0.2 pm pore membranes and finally distributed into sterile ampoules.

The skin-test reagents were used at a concentration of 2 pg/ml, except Nonchromogenicin and Vaccin, which are used at 20 pg/ml, and Leprosin-A (VI, which was used at a concentration of 10 kg/ml.

30 Ly and others

+2-vc&9 CHINA

I 200miles I

Figure. Map of Vietnam indicating the three cities studied.

Skin tests in Vietnam 31

Table I. Characteristics of schoolchildren studied.

city Male/female ratio

Age

Mean SD Range

KG-scar Total

Yes No Unknown

Hanoi 0.78 12.3 3.0 7-17 1 277 0 278 HCMC 1.07 13.2 2.7 8-19 152 88 10 250 Nha Trang 1.13 11.6 3.7 6-19 66 404 37 507

Table II. Reagents used for skin testing.

Tuberculin Abbreviation Species of origin

Aviumin-A Diernhoferin Flavescin Gilvin Gordonin Kansasin leprosin-A, Vietnam Marinin Neoaurumin Nonchromogenicin Ranin 2-1387 Scrofulin Tuberculin Vaccin Xenopin

AA Dh F Gi Go K LA PJ) M Ne No R2 S T V X

M. avium M. diemhoferi M. flavescens M. gilvum M. gordonae M. kansasii M. leprae M. marinum M. neoaurum M. nonchromogenicum M. fortuitum M. scrofulaceum M. tuberculosis M. vaccae M. xenopi

In HCMC and Nha Trang, a large proportion of the population studied had previously been vaccinated with BCG (strain T888, originally obtained from the Pasteur Institute, Paris).

Method of skin-testing

Each child received four intradermal injections: 0.1 ml of two reagents on each forearm. The children in each city were allotted randomly to one of five groups according to the following scheme (see also Table II):

Group 1: T, LA (V), No ,Dh; Group 2: T, V LA (VI, x ; Group 3: T, M :F ,AA; Group 4: T, Gi ,Ne ,S ; Group 5: T, R2 , K , Go.

Tuberculin was always in the left upper position and the other reagents were changed between the groups of children. The responses were read after 72 hours; a diameter of induration of 2 mm or greater was considered positive.

Comparisons of the results from the three groups of children and for the different antigens were evaluated statistically using the Wilcoxon Signed Rank Test (WSRT) or Fisher’s Exact Test (FET) or, in a few cases, the Chi-squared test (x2).

32 Ly and others

Results

Geographical differences in sensitisation

The percentage of positive responders to Tuberculin in children not vaccinated with BCG ranged from 18 % in Hanoi to 27 % in Nha Trang and HCMC. The result from HCMC was very similar to the 24.3 % positivity obtained by others in a recent tuberculin survey with PPD RT23 amongst schoolchildren in that city [J. F. Broekmans & K. E. Planting, unpublished data]. Leprosin-A was recognised by 14.9 % of the children in Nha Trang, but evoked very few positive responses in HCMC (3.0 %) and Hanoi (4.3 %).

The results, expressed as percentages of children with positive reactions, in the three different cities for all 15 mycobacterial antigens are shown in Table ll,l. The overall percentage of positive responses from the pooled data for 15 reagents was significantly lower in Hanoi (mean+SD: 13.1 %f7.5) and HCMC (15.5 %f13.7) than in Nha Trang (25.7 %?I 1.51 (WRST, p=O.Ol and p=O.O12, respectively).

However, when the data from Table III were analysed according to whether the reagents were prepared from slow- or fast-growing species (omitting Tuberculin and Leprosin-A), Hanoi and HCMC are seen to be quite different. Mean responses to the two groups of reagents were similar in Hanoi (12.0 and 14.6 %, respectively), whereas in HCMC responses

Table III. Results of skin tests with new tuberculins for non-BCG-vaccinated children in Hanoi, HCMC and Nha Trang.

Hanoi (HAI

HCMC 1HCI

Nha Trang WI

p value*

T

LA (V)

% 511277 (18.4)

5/l 15 (4.3)

Slow growers AA S K M Go X

7156 (12.5) 1 l/54 (20.4)

6/53 (11.3) 5/56 (8.9) 6!53 (11.3) 4/54 (7.4)

mean SGs 39/326 (12.0)

Fast growers V F R2 Dh Gi No Ne

mean FGs

17/54 (31.5) 5/56 (8.9)

lo/53 (18.9) 4/61 (6.6) 4/54 (7.4)

13/60 (21.71 4/54 (7.4)

57/392 (14.5)

% 24i188 (27.3)

1133 (3.0)

6/19 (42.1) 5/15 (33.3) 3/21 (14.3) 709 (36.8) 2/21 (9.5) 2/22 (9.1)

27/117 (23.1)

l/22 (4.6) o/19 (0.0) l/21 (4.8) 2/11 (18.2) 0115 (0.0) l/l1 (9.1) 3/15 (20.0)

8/l 14 (7.0)

% 1 lo/404 (27.2)

24/161 (14.9)

41/78 (52.6) 20/86 (23.3) 19/79 (24.0) 16/78 (20.5) 29/79 (36.7) 10/91 (11.0)

145/491 (29.5)

36/91 (39.6) 23i78 (29.5) 25i79 (31.6) 15/70 (21.4)

W86 (7.0) lW70 (25.7) 17/86 (19.8)

140/560 (25.0)

HNHC: HA/NT: HUNT: ns 0.0048 “S

“S 0.0032 “S

0.0092 "S ns 0.0085 ns ns

0.0019

0.009 "S "S ns “s ns ns

0.021

<0.0001 ns ns “S

0.0008 ns

<o.ooo

ns 0.0029 ns ns ns ns “S

<o.ooo

ns ns llS

ns “S

ns

“S

0.0008 0.0031 0.0082 ns “S

ns flS

<o.ooo

Total mean: 13.1 15.5 25.7 (sd): (7.5) (13.7) (11.5)

Remarks: For each tuberculin the number of positive reactions and the number of children tested are given. The percentage with positive responses is given in parentheses. *The significance level was calculated by Fisher’s Exact Test: p<O.Ol; ns=non significant.

Skin tests in Vietnam 33

were much more to slow growers (24.2 %) than to fast growers (8.1 %). In Nha Trang, responses to slow growers (28 %) were similar to HCMC, but responses to fast growers were much greater (25.0 %). As well as these differences, the pattern of sensitisation to different individual species in both groups varied considerably between the three cities. Only M. xenopi and M. gilvum appeared to be absent as sensitising species.

Influence of age

The profile and age distribution of the three populations studied did not differ significantly (Table I), and the age distribution over the five groups (see Materials and methods) was similar (data not shown). The observed geographical differences in sensitisation could not, therefore, be ascribed to age differences. The observation of Stanford et al. [IO], that positivity increases with age, was confirmed for most tuberculins, most notably AviuminA, Tuberculin, Nonchromogenicin, Scrofulin and Vaccin (data not shown).

Effect of prior BCG vaccination

The effects of BCG vaccination can only be assessed in HCMC and Nha Trang since only one of the children tested in Hanoi had been vaccinated. Table IV shows the data from the two cities for BCG-vaccinated children which can be compared with the data for non-vaccinated children in Table III.

Significant increases due to BCG vaccination in responsiveness to Tuberculin, Aviumin, Scrofulin and Marinin were found in both cities. Differences in responses to individual reagents after BCG vaccination in the two cities reflected differences seen in the responses of the non-vaccinated children.

Distribution of indurations in tuberculin responsive children

The distribution of indurations to Tuberculin in non-BCG-vaccinated children was also

Table IV. Skin test results in BCG-vaccinated children in HCMC and Nha Trang.

HCMC Nha Trang

% % T 112/152 (73.7) 36/66 (54.55) LA(V) IO/62 (16.1) 2/14 (14.3)

s/0wgr0!Ners AA 21/32 (65.6) 15/15 (100.0) S 15/25 (60.0) 14/34 (41.2) K 10/33 (30.3) 113 (33.3) M 25/32 (78.1) .?/15 (13.3) Go 9/33 (27.3) 2/3 (66.7) X 10141 (24.4) 119 (11.1)

Fasf growers V F I72 Dh Gi No Ne

9/41 (21.9) 719 (77.8) 7/32 (21.9) 8/15 (53.3) 9/33 (27.3) 113 (33.3) 201 (18.2) 3/5 (60.0) 6/25 (24.0) 2/34 (5.9) 7/21 (33.3) 3/5 (60.0) 4/25 (16.0) 10/34 (29.4)

34 Ly and others

Table V. Categories of responders.

Category Hanoi (HA) HCMC (HC) Nha Jrang (NJ) p value(*)

% % % I. A// four reagents BCG-scar - 12/l 51 (7.95) 4766 (6.1) “S

positive <0.02 <0.08 no KG-scar 3/277 (1.1) 1188 (1.1) 13/404 13.22) ns ns ns

2. A// four reagents BCG-scar - 22/151 (14.6) 15/66 (22.7) ns negative <o.ooo <o.ooo

no KG-scar 1837277 (66.1) 60/88 (68.2) 187/404 (46.3) <O.OOl <O.OOl 3. Some reagents BCG-scar - 117/151 (77.5) 47/66 (71.2) ” - ns

positive <o.ooo <0.002 no BCG-scar 91/277 (32.8) 27/88 (30.7) 204f404 (50.5) ns <o.ooo <o.ooo

Remarks: (“) Significance levels calculated by Fisher’s Exact test: p<O.Ol; ns=non significant.

determined (data not shown). Due to the low number of unvaccinated children studied in HCMC, extensive statistical analysis was not possible, but when the Tuberculin non- responsive children were included in the comparison, Chi-square testing indicated a different distribution in Hanoi from that in HCMC and Nha Trang (p=O.O21 and p=O.O19, respectively).

BCG-vaccinated children tended to have larger responses to Tuberculin than did non- BCG-vaccinated children (6.121fi3.10 compared with 5.58f3.01 in HCMC and 7.05rt4.30 compared with 5.34f5.07 in Nha Trang).

Comparison of categories of responders

Category 1 responders have been defined as individuals that recognise either group i antigen or the specific antigens of all of the four species tested [IO]. In Vietnam, this category was poorly represented, the percentages ranging from 1.1 % to 3.22 % in non- BCG-vaccinated children, but were increased by BCG vaccination (Table V); significantly so in the case of HCMC.

Category 2 (containing those reacting to none of the four reagents applied) was correspondingly well represented in all three cities, although it was most common in Hanoi and in HCMC. This category was markedly reduced by BCG vaccination. Amongst the non- vaccinated, category 3 responders (those responding to the specific antigens of only those species they have met) were most numerous in Nha Trang, and the category was significantly increased by BCG vaccination in both cities. Although not designed to recognise the recently described category 4 responders to slow-grower associated (group ii) antigen [I 21, low levels of positivity to Xenopin in all situations do not suggest its significant presence, even in Nha Trang.

Discussion The composition of the environmental mycobacterial flora, as well as the level of sensitization to these mycobacteria in the human host, are postulated critical factors in determining the t-ode of action and the efficacy of BCG vaccination against tuberculosis and leprosy in a given location. A high effectiveness of BCG vaccination against tuberculosis may correlate with a low prevalence of environmental mycobacteria [I31 and low efficacy of vaccination may correlate with a high prevalence of these organisms [141.

The present survey has revealed remarkable differences in levels of sensitisation to

Skin tests in Vietnam 35

environmental mycobacteria in the three cities studied. In general, sensitisation to variou:, new tuberculins occurred with a low frequency in Hanoi and in HCMC (Table III and V!. Some reagents evoked such a low percentage of positive responses that the data resemble those collected in North Lebanon, a country known for its low prevalence of environmentai mycobacteria [7]. The positive reactions with some of these reagents, such as Gilvin anil Xenopin (Table Ill), probably reflect cross-reactivity with group i common antigens [IO], and the species from which they are made may be absent from the Vietnamese environment The highest proportions of positive reactions to both slow-growing and fast-growin species were recorded in Nha Trang (Table Ill). Although data on the local occurrence oi mycobacterial species are lacking, it can be assumed that the composition of the environmental mycobacterial population differs markedly in Nha Trang from those in Hanoi and HCMC. Geographical differences probably account for this observation: Hanor and HCMC are both large cities located in or near river deltas, whereas Nha Trang is situated near the coast, is less humid and has a climate described as Mediterranean [13]. When the data from Hanoi and HCMC were compared, it appeared that although the frequencies of response to the 15 new tuberculins were similar (Table Ill), the predominating environmentai species recognised in Hanoi are fast growers, while in HCMC slow-growing species occur in higher frequencies. Such a difference could, if the hypothesis of Stanford [I] is correct, partly explain the higher incidence of tuberculosis and leprosy in south Vietnam comparer? to the north.

Mycobacterioses continue to be an important public health problem in Vietnam Tuberculosis prevalence ranges from 2.3 % in the north to 10.3 % in south Vietnam [unpublished data]. The annual risk of infection in HCMC is estimated to be 2.8 % (J. F Broekmans and K. E. Planting, unpublished data]. The prevalence of leprosy for the whole of Vietnam is 0.5 per thousand, but is unevenly distributed, with a tendency to higher prevalence rates in the south, in some isolated mountain areas 5.5 % (P. Feenstra, personal communication).

Prior BCG vaccination was associated, in this study, with a significant increase in positivity to slow-growing species, most notably Tuberculin (Table IV], as reported b\/ others [6]. In similar surveys, this has been interpreted as a good prognostic factor for the success of BCG vaccination campaigns in India 191. A recent tuberculin survey performed in HCMC failed to demonstrate a clear-cut difference in size of induration between BCi; vaccinated and non-vaccinated children, raising doubts on the efficacy of the BCG vaccination applied [J. F. Broekmans and K. E. Planting, unpublished data]. However, the response sizes to Tuberculin we obtained suggest protective immunity rather than pre existing tuberculous infection in almost all Tuberculin positive individuals (data not shown;. Further study, including a follow up of BCG vaccinated children, is required to clarify this point.

In conclusion, this survey has documented differences in sensitisation levels tu environmental mycobacteria in three geographically distinct regions in Vietnam. These different sensitisation patterns may be related to observed differences in the prevalence 01 tuberculosis and leprosy between north and south Vietnam. Further studies, including patients with tuberculosis and leprosy, are required to determine whether skin testing with new tuberculins can help to determine optimal BCG vaccination strategies in a geographi- cally diverse country such as Vietnam.

References

1 Stanford JL, Shield MJ, Rook GAW. How environmental mycobacteria may predetermine the protective efficxv of BCG. Tubercle 1981; 62: 56-62.

36 Ly and others

2 Rook GAW, Bahr GM, Stanford JL. The effect of two distinct forms of ceil-mediated responses to mycobacteria on the protective efficacy of BCG. T&em/e 1981; 62: 63-68.

3 Liew FY, Hale C, Howard JG. Immunologic regulation of experimental cutaneous leishmaniasis. V. Characteriza- tion of effector and specific suppressor cells. J lmmunoll982; 128: 1917-1922.

4 Mitchell GF. Host-protective immunity and its suppression in a parasitic disease: murine cutaneous leishmaniasis. lmmunol Today 1984; 5: 224-227.

5 Titus RG, Ceredig E, Cerottini JC, Louis JA. Therapeutic effect of anti-L3T4 monoclonal antibody GK1.5 on cutaneous leishmaniasis in genetically susceptible BALB/c mice. J lmmunol 1985; 135: 2108-2144.

6 Lema E, Stanford JL. Skin-test sensitization by tubercle bacilli and by other mycobacteria in Ethiopian schoolchildren. Tubercle 1984; 65: 285-293.

7 Bahr GM, Stanford JL, Rook GAW, Rees RJW, Grayha GJ, Abdelnoor AM. Skin sensitization to mycobacteria amongst school children prior to a study of BCG vaccination in North Lebanon. Tubercle 1986; 67: 197-203.

8 Stanford JL, Cunningham F, Pilkington A, Sargeant I, Bhatti N, Bennett E. A prospective study of BCG given to young children in Agra, India-A region of high contact with environmental mycobacteria. Tubercle 1987; 68: 39-49.

9 Stanford JL, Sheikh N, Bogle G, Baker C, Series H, Mayo P. Protective effect of BCG in Ahmednagar, India. Tubercle 1987; 68: 169-176.

10 Stanford JL. Immunologically important constituents of mycobacteria: antigens. In The Biology of the Mycobacteria Vol 2, C. Ratledge and J.L. Standford, eds., Academic Press, New York 1983; 85-127.

11 Shield MJ, Stanford JL, Paul RC, Carswell JW. Multiple skin testing of tuberculosis patients with a range of tuberculins, and a comparison with leprosy and Mycobacferium ulcerans infection. J Hyg 1977; 78: 331.

12 McManus IC, Lockwood DNJ, Stanford JL, Shaaban MA, Abdul Ati M and Bahr GM. Recognition of a categon/ of responders to group ii, slow-grower associated antigens amongst Kuwaiti senior school children using a statistical model. Tubercle 1988; 69: 275-81.

13 Chaparas SD. Immunity in tuberculosis. Bull WHO 1982; 60: 447-462. 14 Shield MJ. The importance of immunologically effective contact with environmental mycobacteria. In The

Biology of the Mycobacteria Vol 2, C. Ratledge and J.L. Stanford, eds., Academic Press, New York 1983; 343-415.

15 Vu Tu Lap. Vietnam. Geographical Data. Foreign Languages Publishing House, Hanoi. 1979; 248 pages.