Influence of isoniazid on naturally acquired tuberculin allergy and on induction of allergy by BCG vaccination*

Bull. Org. mond. Sant} 1970, 43, 53-64Bull. Wid HUth Org. Influence of Isoniazid on Naturally Acquired Tuberculin Allergy and on Induction of Allergy by BCG Vaccination* RAJ NARAIN,1 A. S. BAGGA,'2 K. NAGANNA8 & S. MAYURNATH' Previous studies on the influence of isoniazid on the size of the tuberculin reaction have given conflicting results. A controlled study in an area with high prevalence of low-grade allergy has been carried out by the administration of isoniazid or placebo tablets. For those not vaccinated with BCG, isoniazid in a single daily dose of 5 mg/kg body-weight tended to reduce somewhat the size of the tuberculin reaction among those with reactions of12 mm or more at the initial tuberculin test. In people who were vaccinated with BCG, isoniazid given simultaneously resulted in significantly less increase in the size ofpost-vaccination tuberculin reactions as compared with controls; the difference was still significant, in tests conducted 4112 months after the discontinuation of isoniazid. However, in spite of isoniazid, the post-vaccination allergy induced by BCG was quite considerable. This considerable increase in post-vaccination allergy suggests that the vaccination was successful in spite ofthe administration ofisoniazid and makes it clear thatprimary chemoprophylaxis could be combined with BCG vaccination. Administration of isoniazid for 2 months is estimated to have killed about 90 % of the bacilli in the BCG vaccine injected intracutaneously. Available reports regarding the influence of isoniazid on allergy elicited by a Mantoux test show conflicting results. For instance, Marsova (1965) reported stronger reactions after 8 months' treatment with isoniazid among recently infected children with strong reactions. On the other hand, Daniel & Bowerfind (1965) reported that among 15 persons with documented " tuberculin conversion ", chemo- prophylaxis with isoniazid resulted in 9 negative reactors. Rodkey (1965) reported that children * From the Tuberculosis Prevention Trial, Bangalore -20, South India. This, trial is a joint operation sponsored by the Indian Council of Medical Research, Ministry of Health, Government of India; the National Communicable Disease Center, United States Public Health Service; and the World Health Organization. The project is financed by United States PL-480 Funds and by a grant-in-aid from WHO. This paper is also being published in the Indian Journal ofMedical Research. Project Officer. 'Senior Medical Officer. 'Junior Statistical Officer. 'Medical Officer. suffering from tuberculosis became negative to tuberculin after treatment. In these studies, controls were not used, but studies with controls also show a similar difference in results. Among children given isoniazid or placebo for 12 months, a United States Public Health Service Tuberculosis Prophylaxis Trial (1957) reported no significant difference in mean induration or distribu- tion of reactions. Zorini (1963) also reported that prolonged isoniazid treatment did not appreciably modify the pre-existing allergic state. Chapman et al. (1966) reported that reversion to negative tuber- culin reaction occurred as often in the treated as in the untreated children. But Mysakowska et al. (1966) reported " a clear diminution of tuberculin response" after chemoprophylaxis for 9 months. During pilot studies to decide the feasibility of a large-scale chemoprophylaxis trial in India, the in- fluence of isoniazid on naturally acquired tuberculin allergy and on induction of allergy by BCG vaccina- tion has been studied. To avoid bias, randomly allocated controls were given placebo tablets. 2543 -53- RAJ NARA1N AND OTHERS MATERIALS AND METHODS The study was carried out in 7 villages, situated near each other, in Magadi Taluk of Bangalore District, South India. The populations of these villages had not been previously tuberculin-tested or BCG-vaccinated, nor had they had any previous isoniazid prophylaxis. A census of the population was taken by house-to-house visits. All persons aged 2 years and above were tested (test I) with 5 Interna- tional Units (IU) of PPD-S (tuberculin). Reactions were usually read after 72 hours. All persons 5 years of age or more were X-rayed and sputum was ex- amined from those with any X-ray abnormality. At the time of testing (before the tuberculin reaction could be known and therefore regardless of the size of reaction) half the population, by random selection in several strata by age and sex, was vaccinated with freshliquidvaccinefromtheBCGlaboratory, Madras; the other halfwas " vaccinated " with saline. Separate new syringes were used for the tuberculin, the BCG vaccine and for the saline. The site of vaccination, with BCG or with saline, was examined at 1, 2, 3, 4, 6 and 8 weeks during chemoprophylaxis and again 41/2 months after discontinuation ofthe drug. For the first 4 weeks only the size of induration was recorded. For later readings, the size of the skin breakdown (ulceration, crust or scar) was measured in mm. Families (not individuals) were allocated randomly to receive isoniazid or placebo tablets, which were administered under direct supervision each day to all persons 7 years of age or more. Techniques have been described in another paper (Raj Narain et al., 1970). The dose of isoniazid was 100 mg (1 tablet) for those 7-11 years of age and 200 mg for those 12 years of age or more. The average dose per kg of body- weight for the former was 5.2 mg (range 2.9 mg- 10.0 mg) and for the latter 4.9 mg (range 2.8 mg- 1 1.1 mg). Placebo was also administered in a dosage of 1 or 2 tablets. Administration of tablets was started at the time of tuberculin testing and BCG vaccination. Consumption of isoniazid was cross-checked by urine examination using the Belles-Littleman filter- paper spot test (Feasibility Study for Tuberculosis Prevention Trials in India, 1967). Half the families in each village were randomly selected for urine examin- ation once every week. Thus, all persons, on placebo or on isoniazid, were eligible for urine examination once a fortnight. Urine examination in a village was carried out on a randomly allocated day each week. After tablets had been administered for 60-62 days, the entire population was tested (test II) with 5 IU of PPD-S and was again tested (test III) about 41/2 months after the discontinuation of pills. A different site was chosen for testing each time. To avoid bias in subsequent readings, a new set of cards was used for each reading of the tuberculin reaction and for each examination of the site of vaccination. The new cards had identifying informa- tion, but not the results ofprevious tests, or the details of vaccination or of drug status. The present analysis is confined to those aged 7 years and over, for whom the results of all 3 tuberculin tests are available. Study population The registered dejure population of 7 years of age and over was 2127 in the 7 villages and comprised 493 families. The following persons were excluded: (1) 157 persons who were away from the village during the testing and vaccination; (2) 39 who had a BCG scar, as a result of vaccina- tion elsewhere; (3) 21, because of errors committed during testing, vaccination etc., and duly recorded; and (4) 25 persons from 1 village because during analysis ofthe data it was found that 6 persons givenBCG had no scar at all, while 17 given saline had a typical BCG scar. In addition to these 23 persons, 7 other persons who were given BCG had a normal scar. All 30 per- sons were examined consecutively in 1 village within an hour or so, during which it is likely that once the BCG syringe and twice the saline syringe were filled with the wrong solution. Of these 30 persons 25 were 7 years of age or more and have been excluded. Of the remaining 1885, a total of 1676 (89%) accepted vaccination with BCG or with saline (con- trols). The number allocated to isoniazid or placebo tablets among the BCG-vaccinated and the controls, and numbers tested and read at all the 3 tests are shown in Fig. 1. RESULTS Effect of isoniazid on naturally acquired tuberculin allergy Only the controls vaccinated with saline are con- sidered in this section. Actual pill consumption can be estimated by the results of urine examination among those given isoniazid. From the 320 such persons, 1118 samples of urine were examined during the 2 months; 899 54 INFLUENCE OF ISONIAZID ON TUBERCULIN ALLERGY FIG. I ALLOCATION OF STUDY POPULATION STUDY POPULATION 1676 CONTROLS (VACCINATED WITH SALINE) 838 ALLOCATED TO ISONIAZID OR PLACEBO GROUP TESTED AND READ ) AT ALL TESTS J ISONIAZID 463 I 320 (69%) PLACEBO 375 269 (72%) BCG-VACCINATED 838 ISONIAZID PLACEBO 421 417 299 303 (71 %) (73%) (80%) were positive. The frequency of errors during the administration of isoniazid can be judged from the 269 persons on placebo tablets; ofthe 929 samples examined from them, 34 (3.7%.) were positive. Of the 320 persons tested in the isoniazid group and the 269 tested in the placebo group, 2 and 1 respectively had typical BCG lesions; these 3 findings indicate errors during vaccination with saline. Histograms of tuberculin reactions by age at the initial test, both for the placebo and the isoniazid groups, can be seen in Fig. 2. The distribution is bimodal and possibly reactions of 12 mm or more at the initial test may be taken as positive reactions; among the remaining reactions it was difficult to decide which should be classified as negative or as intermediate reactions. Rather arbitrarily we have classified 0-mm-5-mm reactions as negative and those of 6 mm-li mm as intermediate reactions. The difference in reaction size for each person (test-II reaction minus test-I reaction and test-III reaction minus test-I reaction) was calculated. Means of such differences for the isoniazid and the placebo groups, separately for males and females, are shown in Table 1. The distribution of these differences, separ- ately for each sex and for those with reactions of 0 mm-1l mm and 12 mm or more, at test I, is shown in Fig. 3. The influence of isoniazid, if any, could be seen by a comparison of the mean differences between the isoniazid and the placebo groups at test II and test III. The largest difference (placebo minus isoniazid) both at test II and at test III is seen in Table 1 among males with reactions of 12 mm or more at the first test. Isoniazid could be expected to reduce the size of the tuberculin reaction only among those infected with Mycobacterium tuberculosis, and that may be the reason for this trend among the males with reactions of 12 mm or more. The possible small influence of isoniazid on those with reactions of 12 mm or more could be expected to be greater among those who took isoniazid more regularly. Those who consumed pills for at least 48 days of the total of 60-62 days, and who had at least 3 of the 4 urine samples positive, were compared with those who took isoniazid less regularly and had a smaller number of urine samples positive. No significant difference was seen. Effect ofisoniazid on induction ofallergy by BCG Only those persons vaccinated with BCG are con- sidered in this section (see Fig. 1). From the 299 persons on isoniazid, the number of urine samples examined was 1057; 815 (77%) were positive and are indicative of actual pill consumption. Of the 1067 samples examined from the 303 persons on placebo, 31 (3%.) were positive, and indicate the frequency of errors in the administration of isoniazid. All the vaccinated persons had a BCG scar. The mean ofincreases in size ofthe reaction at test II and test III in comparison with the reaction at test I, both for the placebo and the isoniazid groups is shown separately for the two sexes in Table 2. Those cases in which the mean of the increases in reaction size were significantly less in the isoniazid group than in the placebo group are indicated in the table. The mean post-vaccination increase in diameter ofindura- tion among those with negative or intermediate reactions at the first test was significantly greater in the placebo group than in the isoniazid group, both at test II and at test III. For those with initial reactions 55 - RAJ NARAIN AND OTHERS FIG. 2 NON-VACCINATED GROUP: DISTRIBUTION OF SIZE OF REACTION AT TEST I SEPARATELY BY AGE AND BY MEDICATION 1 7 TO 14 YEARS '5 TO 29 YEARS 30 YEARS OR MORE | 7 YEARS OR MORE I w z hi U) 56 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 40 DIAMETER OF INDURATION (mm) WHO 00274 Placebo group Isonlazld group TABLE I CONTROL GROUP (VACCINATED WITH SALINE): MEAN DIFFERENCE IN REACTION SIZE AT TEST II AND TEST III FROM REACTIONS AT TEST I FOR THE ISONIAZID (I) AND PLACEBO (P) GROUPS FOR 3 LEVELS OF INDURATION AT TEST I Males Females SizeonNof mbertested Mean difference In reaction size Numbertested Mean difference in reaction sizelnuattest AttestsedNumerteteat est and read Attest 11 a Attest IlI b and read At testlI 1a At test lIlb I _ I I P I I I P 0-5 60 33 2.2 2.1 3.6 3.7 69 57 1.7 2.3 4.0 3.9 6-11 37 28 -0.1 -1.0 1.7 1.4 58 52 -0.3 0.6 2.3 2.5 >12 42 56 -1.3 0.4 1.3 2.2 54 43 -1.1 -1.4 1.1 0.8- a Test done Immediately after 2 months of Isonlazld administration. b Test done 4Y months after discontinuatIon of Isonlazld administration. FIG. 3 DISTRIBUTION OF DIFFERENCES IN REACTION SIZE FOR THE ISONIAZID AND PLACEBO GROUPS-FOR 2SIZES OF REACTION AT TEST I: NON-VACCINATED GROUP, BOTH SEXES a FROM TEST II TO TEST I 0-1lmm AT TEST I 2 12mm AT TEST I 21 04ow. X 21 1' 20 is 10 s -S 0 10 15 20 -10 FROM TEST III TO TEST I 20 is 10 5 -5 0 s 10 1s 20 -10 -s DIFFERENCE IN SIZE OF REACTION (mm-) a Curves smoothed by taking 3-point moving averages. INUMBER 96 ISO NIAZID --- MEAN! -t 2STANDARD' 36 DEVIATION INUMER -99 \P L AC E BO - .3EAN-0 'PLACEBO - VLSTANDARD3 4PVIATION S 0 s 10 1s 0 S 10 is WHO 00275 I~~~~~~~~~~~~~~~~~~~~~ NUMSER 224 ISONIAZID- EAN 10 .s STANDARD:3 9jI % ~~~~~~DEVIATtON UMuBsER 170 PLACEBO - 5<JEAN1?lSTANDARD 3-7IDVIATION /''\\~Ii / ssS~~~~~~~~~~~~~~ MOSER: 224 ISO N I AZID -- sMEAN, 31ISONIAZID STANDARD:40 I IP~~~~~~~~~~~~JEVIATION -. 5:~~~~~~0170 PLACEBO- ANI2D9 1 IOND:W^TA 44 S / /- % / I\ rNIEQ'96 ISONIAZID._,_ EAN:121STANORD:36 ,EVIATION I K~~~~4USR:99 , PLACEBO ,ANEA 1 6I / 'ARD: 3 STANO :36II / %E VIATION II's\ ns . Iz RAJ NARAIN AND OTHERS a of 12mm or more the differences were not significant, _4 C Q although the placebo group consistently showed a __________ greater increase in post-vaccination allergy. oi O @ % t b Distribution by- size of reaction, for 3 age-groups <: cm w OD t "and for all ages, for the isoniazid and the placebo z ~ _ o .=__ _ groups at the 3 tests, is shown in Fig. 4. Both theisoniazid and the placebo groups at the second as well IF m CR q X a?a. as at the third test show a shift of the mode to the =X" _ right, but the shift is greater in the placebo group. 0O Distributions of difference in reaction size, calcu- -~<-_ w. lated for each person (test-Il reaction minus test-I _________ reaction and test-HI reaction minus test-I reaction) F .. for both the isoniazid and placebo groups are shown a. " 3 4 X e -. Xo Oin Fig. 5, separately for those with reactions in test-I CzzO _ _ _ of 0 mm-5 mm and 6 mm-l1 mm. It can be seen z < r @ X clearly that the post-vaccination increase in tuberculinoz o < In C!_ q> allergy was much less in the isoniazid group. The _________ distributions for those with test-I reactions of 12 mm cc U =or moreare not shownin Fig. 5; BCG is usually not *0 0@ eD _ X o e o ". given to such a group and the differences between the a:______ - E -llisoniazid and placebo groups were not significant " L'E . -. (Table 2). :0o 0L.. R _ _ E > I_ _ @ Influence of isoniazid on the size of the BCG lesion cx X ; <; The size of induration or of skin breakdown wasIJJ z measured, for the first 6 readings, by 2 readers * > @l(onefor 4villages the other for 3 villages). For the 0 - 0 co seventh reading, 41/2 months after discontinuation oo < E gIn V f . of isoniazid, the size of scar or skin breakcdown x - oo was measured by the same reader in all the 7 xE ° ° villages. Lu >-.qptN NXa_ o 0 0 The results are shown in Fig. 6 separately for those Z 'H o @_=| E"swithinitial tuberculin reactions of 0 mm-5 mm,2 0 cm la 0 e X O6mm-llmmand 12mmormore. Thedetailscanbe P a r, _-4 co seen in Table 3. Due to rather marked differences <JX__ _O between their readings, the results of the 2 readers a. _ o0 are shown separately. There was no significant , m° X e Xt difference between the sexes in thesizeof the BCG Qz U CD XO________ _X lesion or of skin breakdown at the various readings, zJX_ " " Q E. Q hencethe data for both sexes have been combined.wi 0 q C aQ _Xs < _ o, 0 - Themeansize of lesion for the isoniazid group was a____ *: i @ smaller than for the placebo group except for the z X readingsbyReaderX, taken duringthe first2weeks 0. aoX A V O 4.- for those with initial reactions of 12 mm or more. ..@X s> @ @ ~~~~Reader Y showed greater differences between the - 00 thet @ isoniazid and placebo groups than Reader X. At o z _ @ X X > O Oco0 the sixth week and thereafter, differences between the a______l_ HX0 ¢ placebo and isoniazid groups were small and are -- 0G=F a a showninsetin Fig.6 on an enlarged scale. As shown -Z 0 in Table 3 most of the differences between the .Nuan @ E: z iz\ Xplacebo and isoniazid groups were statistically 0: = X F significant. 58 INFLUENCE OF ISONIAZID ON TUBERCULIN ALLERGY FIG. 4 VACCCINATED GROUP: DISTRIBUTION OF SIZE OF REACTION SEPARATELY BY AGE AND BY MEDICATIO N 7 TO 14 YEARS 15 TO 29 YEARS 30 YEARS OR mORE 7 YEARS OR MORE _ _ _ _ INITIAL TEST _40 NUMBER: 103 MEAN: 5 0 mm NUMBER: 76 MEAN' 8 5mm SA NUMBER:124 MEAN:131 mm t_ NUMBER: 303 MEAN: 92 mrm -30 - 20 - 10 0- -7eZZA 4zzzzA Pi2 M VA 0 NUMBER:99 NUMBER:84 NUMBER:116 NUMBER:299 30 MEAN: 49 MEAN :E Bm2 mm MEAN: 13-Smm MEAN:9 2mm- 30 20 - 20 SECOND TEST: AFTER TWO MONTHS OF MEDICATION MEAN: 180mm MEAN: 179mm MEAN: 1784mm MEAN :18 7mm 30 --30 20 - 20 10- T 10 MEAN:15-7mmn MEAN,:17 1mm MEAN:17-3mm MEAN: 16 7mm 30 -30 20 J -20 10 10 THIRD -TEST: 4' MONTHS AFTER STOPPING MEDICATION 4 MEAN: 195mm MEAN:18 9m MEAN: 20-1mm MEAN: 196mm 30 - 30 20 -20 10- 10 fl.fl~~~~~~~~~~~~~&~~~~~ -~~~0 MEAN: 16-Sm MEAN: 17-5mm A' MEAN: 180mm I Al ~ .-. *1 FI a 10 20 30 0 10 20 30 DIAMETER OF 0 fnli MEAN: 17-4mm 30 20 100 10 20 30 0 10 20 30 40 IIRATIfnu(mm) WHO 00276 MY//X Placebo group Isonlazld group 40 - 30 - 20 - 10 - z w U w 0. 30 - 20- 10- A.. 59 .vv.BW_"w RAJ NARAIN AND OTHERS FIG, 5 DISTRIBUTION OF DIFFERENCES IN REACTION SIZE FOR THE ISONIAZID AND PLACEBO GROUPS FOR 2 SIZES - OF REACTION AT TEST 1: VACCINATED GROUP, BOTH SEXES a FROM TEST 1I TO TEST I 0-5mm AT TEST I 6-1lmm AT TEST FROM TEST III TO TEST I DWFEtENCE N SIZE OF REACTION (mm) WHO 00277 a Curves smoothed by taking 3-point moving averages. To sum up, isoniazid taken for 2 months, and started simultaneously with BCG vaccination, resulted in significantly smaller increases of post- vaccination allergy and this difference between the isoniazid and the placebo groups persisted for 41/2 months after the discontinuation of the drug. DISCUSSION BCG is sensitive to isoniazid and therefore an influence of isoniazid was to be expected. In guinea- pigs there was considerable reduction both in the production of immunity and of tuberculin sensitivity when administration of isoniazid was started before BCG vaccination (Toyohara, Kudoh & Obayashi 1959). It is not known what the result would have been had administration of isoniazid been started a few days after BCG vaccination. In other studies, when the administration of isoniazid was started 8 weeks after vaccination, it had no influence on the production of immunity m mice (Lambert, 1960). In man, when isoniazid was started47days after vaccina- tion, there was no significant influence on post- vaccination allergy (Aronson, Taylor & Kirk, 1956). The mean increase in siz of the post-vaccination reaction, although less in the isoniazid group, was still considerable (7.5 mm at test II and 8.2 mm at test III; soe Table 2, both sexes). The post-vaccina- z up toi 60A INFLUENCE OF ISONIAZID ON TUBERCULIN ALLERGY tion increase in size of the reaction to tuberculin ndicates successful vaccination. Therefore, in spite of simultaneous isoniazid intake for 2 months, vaccination was successful. This observation may be of considerable practical importance when isoniaizd chemoprophylaxis is advised for tuberculin-negative contacts. The effectiveness ofBCG vaccination given simultaneously would not be reduced. Alternatively, BCG may be given first and isoniazid administration delayed for a few days. Other studies (Lambert, 1960; Aronson, Taylor & Kirk, 1956) have suggested that delayed administration of isoniazid could make the advantages of BCG vaccination more certain. Thus the benefits of both vaccination and chemoprophy- FIG. 6 CHANGE IN MEAN SIZE OF VACCINATION LESION WITH TIME FOR THE ISONIAZID AND PLACEBO GROUPS BY SIZE OF THE INITIAL TUBERCULIN REACTION, SEPARATELY FOR EACH READER READER X C C < 0 0 VI g z z Q it J w VI z 4 hi READER Y 25 20 IS 10 s i I WEEK OF OBSERVATION !m WHO '00278 -I---lsonlazid group Placebo group S 10 9- 07 1I.l(. "'6 ..t1 "'!"-. -. S - II,1w'.. - - - - - - 0-S mm 7 - nv1 -- 61 RAJ NARAIN AND OTHERS TABLE 3 VACCINATED GROUP: SIZE OF INDURATION OR SKIN BREAKDOWN (mm)G DUE TO BCG VACCINATION AT VARIOUS INTERVALS IN THE ISONIAZID AND PLACEBO GROUPS BY SIZE OF INITIAL TUBERCULIN REACTION SEPARATELY FOR EACH READER Size Reader X Reader Y Interval of Initial Charateritic (4 villages) (3 villages)reaction a s ____________ (mm) Isonlazid I Placebo Isoniazid Placebo Number 72 68 69 65Mean size 10.2 11.2 1.1 C 3|3 1week 6-11 ~~~Number 45 57 63 681 week -1 iMean size 12.3 b 14.6 3.5 b 6.3 >12 Number 59 48 67 71Mean size 22.9 22.0 16.4 17.0 __5 Number 66 56 69 64Mean size 11.8 12.3 563 10.0 2 weeks e-11 Number 38 55 67 63Mean size 13.2 c 16.3 8.8 C 13.2 Number 49 40 64 64312 Mean size 23.5 21.2 16.0 b 19.4 | Number 66 64 67 65 Mean size 14.3 15.6 7.4 c 12.6 3weeks 6-11 Number 43 50 69 6334weeks 60,-11 iiMean size 15.4 b 17.9 10.1 C 14.9 >12 Number 56 43 68 6312 Mean size 22.4 22.5 15.70c 18.9 10-5 Number 65 66 66 63Mean size 11.80 14.6 8.20c 12.9 4weeks 6-11 ~~~Number 44 53 66 684 weeks 6-11 Mean size 11.7 17.2 8.2 c 14.1 12 Number 54 44 65 67Mean size 17.6 18.6 12.2 c 17.2 Number 64 68 71 630e-11 Mean size 7.2 b 7.8 6.0 c 6.8 6 weeks 6-11 Number 42 49 68 64Mean size 7.2 c 8.4 6.0 c 7.0 ; 12 Number 52 43 69 60Mean size 9.3 9.5 6.60 7.0 3~ s Number 68 67 70 65Mean size 7.0 7.3 5.6 c 6.4 8weeks ~Number 41 55 70 678 weekhsd 611 Mean size 7.2 c 8.3 5.8 c 7.0 12 Number 52 46 65 68Mean size 9.1 9.3 6.3 c 7.8 1 ~ 1os Number 131 12105 Mean size 7.2 c 8.2 Oy2months4 6-11 Number 107 112Mean size 7.40 8.9 Number 106 107tiA12 Mean size 9.6 b 10.7 a The first 4 readings recorded size of induration and the last 3 readings recorded size of skin breakdown (crust, ulcer or scar). b Difference in reaction size significantly less than with placebo at the 95 % confidence level. c Difference in reaction size significantly less than with placebo at the 99 % confidence level. d All readings by Reader X. 62 INFLUENCE OF ISONIAZID ON TUBERCULIN ALLERGY Iaxis could be made available almost simultaneously to such contacts. One may attempt to quantify the influence of the present dose schedule. The mean increase in post- vaccination allergy in the isoniazid group was less than that in the placebo group by 1.4 mm at test II and by 2.2 mm at test III. It has been shown that the mean size of induration was reduced by 2 mm- 3 mm when only one-tenth of the standard dose of BCG was given (Edwards, Palmer & Magnus, 1953). A reduction of this order would suggest that the present dose-schedule rendered ineffective almost 90% of the bacilli in the vaccine. The influence of isoniazid persisted for 41/2 months after its discontinuation, as both the post-vaccination allergy and the size of BCG scar continued to be significantly smaller in the isoniazid group. This suggests that the present dose-schedule killed some 90% ofthe bacilli and did not only prevent them from multiplying. In the latter case the influence of isoniazid would not have persisted 41/2 months after discontinuation of the drug. Sex differences Some of the data have been presented by sex. In Table 2, it can be seen that the mean increase in the size of post-vaccination induration among those with initial reactions of 0 mm-5 mm or 6 mm-l mm, in both the isoniazid and the placebo groups, was con- sistently greater in the females than in the males. This greater increase of post-vaccination allergy in the female does not seem to have been reported earlier. But it is known that female guinea-pigs have a higher degree of sensitivity to BCG than male guinea-pigs. Table 2 also seems to suggest that the differences between the placebo and isoniazid groups, as seen by the mean increase ofpost-vaccination allergy, were greater in the male than in the female, particularly at test III. Also among the unvaccinated controls (Table 1) it was among the infected males (> 12 mm) that the difference between the placebo and isoniazid groups was the largest. There were no significant differences between the sexes in the size of BCG scar or skin breakdown at various intervals. ACKNOWLEDGEMENTS The authors are grateful to Dr G. Wijsmuller, Medical Officer, United States Public Health Service, US Depart- ment of Health, Education, and Welfare, for many helpful and constructive suggestions. The authors also wish to thank the field teams and the team leaders for their hard work, the staff of the Statis- tical Section of the Feasibility Study for Tuberculosis Prevention Trial for tabulation and statistical help and Mr B. Rama Rao and Mr K. T. N. Iyengar for secretarial R1tSUMIt INFLUENCE DE L'ISONIAZIDE SUR L'ALLERGIE TUBERCULINIQUE NATURELLEMENT ACQUISE ET SUR LE DtVELOPPEMENT DE L'ALLERGIE APRES VACCINATION PAR LE BCG Au cours d'etudes pilotes visant A explorer les possi- bilit6s de recours a la chimioprophylaxie antituberculeuse de masse en Inde, on a recherche l'effet de l'administration d'isoniazide sur l'allergie a.la tuberculine, acquise natu- rellement ou suscit6e par le BCG. Les investigations ont port6 sur la population de sept villages du sud de l'Inde. Apres avoir soumis tous les habitants a une epreuve tuberculinique, on les a r6partis al6atoirement en deux groupes dgaux: le ler a subi la vaccination par le BCG, le 2e (groupe temoin) a requ une injection de solute salin. La moitie des sujets vaccin6s et la moitie des sujets temoins ont ensuite 6t6 trait6s quo- tidiennement par l'isoniazide a la dose moyenne de 5 mg/kg pendant 60-62 jours; les autres ont re4u un placebo. Des examens d'urine de contr6le par la methode de Belles-Littleman ont montre une prise effective du medicament dans environ 80% des cas. Tous les sujets inclus dans l'etude ont subi une 2e epreuve tuberculinique des l'arret du traitement et une 3e apres un delai de 4 mois 1/2. Dans le groupe non vaccine, l'effet de l'isoniazide s'est marqu6 par une l6gere reduction des dimensions de la reaction tuberculinique chez les sujets infectes, c'est- a-dire presentant une induration de 12 mm ou plus lors de l'epreuve tuberculinique initiale. Dans le groupe vaccine par le BCG, le taux moyen d'augmentation des dimensions de la r6action tuberculinique a 6t6 moindre chez les sujets traitds par l'isoniazide que chez les sujets 63 64 RAJ NARAIN AND OTHERS traites par placebo. Cette diff6rence d'allergie postvacci- nale entre les deux categories a 6t6 a nouveau constat6e lors de la 3e 6preuve tuberculinique. Bien que moins prononc6e qu'apres administration d'un placebo, I'augmentation postvaccinale du diametre de la reaction n'en a pas moins ete consid6rable chez les sujets trait6s par l'isoniazide. Le medicament n'a donc pas contrecarr6 l'action du BCG chez les sujets tubercu- culino-n6gatifs et il semble possible d'associer la chimio- th6rapie primaire i la vaccination. On estime que le traitement par l'isoniazide aux dosages indiqu6s pendant 2 mois a eu pour effet d'inactiver pres de 90% des bacilles contenus dans le vaccin. Par ailleurs, certaines constata- tions donnent A penser que le sexe pourrait avoir une influence a la fois sur la r6action au BCG et sur la reponse i la chimioprophylaxie par l'isoniazide. REFERENCES Aronson. J. D., Taylor, H. C. & Kirk, D. L. (1956) Amer. Rev. Tuberc., 74, 7 Chapman, J. S., Black B. G., Dyerly, M. & Kinderberg, S. (1966) Amer. Rev. resp. Dis., 94, 485 Daniel, T. M. & Bowerfind, E. S., Jr (1967) Amer. Rev. resp. Dis., 95, 500 Edwards, L. B., Palmer, C. E. & Magnus, K. (1953) BCG vaccination, Geneva, p. 110 (World Health Organization: Monograph Series, No. 12) Feasibility Study for Tuberculosis Prevention Trial in India (1967) Amer. Rev. resp. Dis., 96, 17 Lambert, H. P. (1960) Amer. Rev. resp. Dis., 82, 619 Marsova, D. (1965) Rozhl. Tuberk, 25, 262; Abstracted in: Amer. Rev. resp. Dis., 1967, 95, 342 Mysakowska, H., Pietron, E., Grodzki, S., Srednicka, D., Cyganiewicz, E., Rozynska, M. & Smajkiewicz, L. (1966) Amer. Rev. resp. Dis., 93, 628 Raj Narain, Mayurnath, S., Subba Rao, M. S., & Murthy, S. S. (1970) Bull. Wid Hlth Org., 43, 41 Rodkey, G. W., (1965) Amer. Rev. resp. Dis., 92, 316 Toyohara, M., Kudoh, S., & Obayashi, Y. (1959) Tubercle (Edinb.), 40, 184 United States Public Health Service Tuberculosis Pro- phylaxis Trial (1957) Amer. Rev. Tuberc., 76, 942 Zorini, A. 0. (1963) Dis. Chest., 43, 1