Tuberculin Test in the Diagnosis of Childhood Tuberculosis: Analysis of Quantitative and Qualitative Features

ORIGINAL ARTICLE

Tuberculin test in the diagnosis of childhood tuberculosis: analysis of quantitative and qualitative features

Henriqueta M.V. Oliveira,1 Clemax C. Sant’Anna2

Abstract Abstract: tuberculin test is an indirect method for diagnosing tuberculosis. It is important in the diagnosis of suspected childhood tuberculosis because children are often nonbacillary, or cannot produce sputum. Objective: to assess the accuracy of quantitative and qualitative features of tuberculin test in the diagnosis of childhood tuberculosis. Methods: evaluation of diagnostic tests; tuberculin tests with purified protein derivative RT-23 at two tuberculin units were carried out in 158 patients distributed in two groups: 101 Tb-negative and BCG- vaccinated children, and 57 tuberculous children diagnosed clinically, radiologically, and epidemiologically. Test results were analyzed quantitatively (Mantoux s test) and qualitatively (Koch- and Listeria-type reactions). Results: using cutoff for positivity at 10 mm in Mantoux s test, quantitative results indicated 85.9% of sensitivity, and 86.1% of specificity. Qualitative results (Koch-type reaction) indicated 77.2% of sensitivity and 98% of specificity. Conclusion: qualitative analysis of tuberculin test proved useful in the diagnosis of childhood tuberculosis in combination with Mantoux s test.

J Pediatr (Rio J) 2000; 76(2): 115-8: tuberculosis - diagnosis, tuberculin test, child.

Introduction Even though tuberculin was first discovered in 1890, it Taking this into consideration, tuberculin test with only started being used systematically and using different purified protein derivative (PPD) according to Mantoux’s techniques in 1907 and 1908.1,2 As a general rule, it is technique is used in the diagnosis of tuberculosis. In some known that individuals react to tuberculin injection when cases, it is also used in deciding medical procedures such as infected with Mycobacterium tuberculosis, with other in the chemoprophylaxis of children exposed to tuberculous mycobacteria, or, still, when previously vaccinated with patients. BCG. The importance of the tuberculin test for the diagnosis itself lies more in its negative predictive value, which is very high due to the epidemiological situation in Brazil, than in 1 1. Pediatrician and Neonatologist. Chief, Mental Health Service, Hospital its positive predictive value. Municipal Alexander Fleming (HMAF), Rio de Janeiro, RJ. 2. Associate Professor, School of Medicine, Department of Pediatrics, Studies of the prevalence of reaction to tuberculin tests Universidade Federal do Rio de Janeiro (UFRJ). allow assessing risk for infection in different communities,

115 116 Jornal de Pediatria - Vol. 76, Nº2, 2000 Tuberculin test in the diagnosis of childhood tuberculosis:... - Oliveira HMV et alii thus indicating the epidemiological importance of this test. We studied children from low-income families, of both Mass BCG vaccination campaigns have, to a certain extent, sexes, aged from 2 months to 12 years. Patients were made it difficult to assess prevalence of tuberculosis randomly selected and then divided into two study groups: infection. That is because reaction to tests may be caused group 1 - 57 inpatients who fulfilled criteria for diagnosis of both by M. tuberculosis infection and by previous BCG tuberculosis; group 2 - 101 children who were healthy, vaccination. There are, however, new methods for eutrophic, asymptomatic from a respiratory perspective, distinguishing between allergy produced by M. tuberculosis not presenting clinical or radiological evidence of and by other mycobacteria, including BCG.3 In this sense, tuberculosis nor history of tuberculosis infection, BCG- it was possible to assess risk for tuberculous infection in Rio vaccinated, and who were randomly selected at different de Janeiro4 and in São Paulo,5 respectively. outpatient clinics. Tuberculin test was carried out only after Moreover, studies have indicated that Listeria-type the obtainment of verbal consent from the parents or person reactions to tuberculin test are associated with protective responsible for the child. immunity of BCG; whereas Koch-type reactions are Patients were considered as a case of tuberculosis if the associated with ineffectiveness of BCG and tissue damaging following three diagnostic criteria were fulfilled: 6,7 hypersensitivity. a) suggestive clinical status, with evolution of fever for These findings have lead to the hypothesis that tuberculin over 15 days, with or without respiratory or extrapulmonary test with PPD RT-23 at two tuberculin units would allow symptoms, such as the indication of central nervous system quantitative assessment considering the size of induration, infection or peripheral adenomegaly with continuous and qualitative assessment considering the form of evolution for over 3 weeks; induration. Criteria used to classify reactions are presented b) chest X-ray presenting hilar or paratracheal in Table 1. adenomegaly, miliary hypopatency, or alveolar hypopatency The objective of our study is to determine the accuracy persisting for over 15 days with no response to antibiotic of quantitative and qualitative features of tuberculin test in therapy using drugs with no proven effect over M. the diagnosis of childhood tuberculosis. tuberculosis; c) reported contact with tuberculous adult patient. Our tuberculin tests were carried out by one of the authors (HMVO), who received previous training at the Table 1 - Criteria for qualitative classification of tuberculin Tuberculosis and Pneumology Institute, Universidade test Federal do Rio de Janeiro. The training consisted of the Characteristics Reaction type assessment of 32 tuberculin tests by two different examiners whose intra-class correlation coefficient was equal to 99.2%. Koch Listeria Tuberculin tests were administered to volar forearm, Induration Well-delineated Not well-delineated using PPD RT-23 tuberculin, tween 80, at 2 tuberculin units Consistency Hard Soft per ml, obtained from the Tuberculin Unit at Tuberculosis Painful Yes No and Pneumology Institute, Universidade Federal do Rio de Janeiro. Tests were assessed 72 hours after the administration of tuberculin according to Mantoux’s technique.1 We also carried out qualitative assessment related to the occurrence of Koch- and Listeria-type reactions according to Stanford et al.6,7 The smallest induration size considered for assessment was 4 mm. Patients and Methods Data were collected using specific forms and analyzed The study consisted of the evaluation of diagnostic tests using the Epi-info 6 statistical software. carried out prospectively from July 1986 until December 1989 at city, state, and federal public hospitals and healthcare units in the city of Rio de Janeiro, Brazil. This study included patients from the Unidade Auxiliar Results de Cuidados Primários de Saúde Vargem Grande, Group 1 included 57 patients aged 2 months to 12 years Jacarépaguá, and, in cases of hospital admission, from the (average = 35 months), and group 2 included 101 children following institutions: Hospital Geral de Nova Iguaçu, aged 3 months to 12 years (average = 35 months). Table 2 Hospital Municipal Jesus, Hospital Estadual São Sebastião, presents age distribution for both groups. We observed a Hospital Raphael de Paula Souza, and Instituto de predominance of children with less than 2 years of age (less Puericultura e Pediatria Martagão Gesteira (Universidade than 24 months) in our population: 29/57 (50.87%) in group Federal do Rio de Janeiro). 1 and 55/101 (54.45%) in group 2. Tuberculin test in the diagnosis of childhood tuberculosis:... - Oliveira HMV et alii Jornal de Pediatria - Vol. 76, Nº2, 2000 117

Table 2 - Distribution of children with and without tuberculosis (groups 1 and 2, respectively) according to age group

Age groups Group 1 Group 2 (months) n % accumulated % n % accumulated %

02 – 11 16 28.08 28.08 38 37.63 37.63 12 – 23 13 22.81 50.89 17 16.83 54.46 24 – 35 8 14.03 64.92 10 9.90 64.36 36 – 47 3 5.26 70.81 6 5.94 70.30 48 – 59 7 12.28 82.46 5 4.95 75.25 60 – 71 2 3.51 85.97 6 5.94 81.19 72 - 83 2 3.51 89.48 7 6.93 88.12 84 – 95 4 7.02 96.50 7 6.93 95.05 96 – 107 2 1.98 97.03 108 – 119 1 1.75 98.25 2 1.98 99.01 120 – 131 132 - 143 144 - 155 1 1.75 100.00 1 0.99 100.00

BCG had been administered during the 1st month of age Table 4 - Distribution of children according to the presence of to 33 (57.89%) patients in group 1 and to all patients in tuberculosis and Koch-type reactions group 2. Koch phenomenon Tuberculosis Total Table 3 presents tuberculin test results for the 158 Yes No children assessed according to Mantoux’s technique. Results yielded sensitivity of 85.9% (49/57) and specificity of Positive 44 02 46 86.1% (87/101). Negative 13 99 112 Total 57 101 158

Table 3 - Distribution of children according to tuberculin test results and presence of tuberculosis Discussion Tuberculin test Tuberculosis Total After the assessment of tuberculin tests (induration size Yes No in millimeters), our findings yielded sensitivity and specificity lower than those found in the literature (around Reaction (> 10 mm) 49 14 63 90%).10 We believe that these results are coincidental, No reaction 8 87 95 since the indicated technical procedures were carefully Total 57 101 158 followed during testing.

Table 5 - Distribution of children according to the presence of tuberculosis and Listeria-type reactions

Listeria phenomenon Tuberculosis Total Table 4 presents distribution of children according to Yes No Koch-type reaction. Results yielded sensitivity of 77.2% (44/57) and specificity of 98.0% (99/101). Table 5 presents Positive 6 55 61 distribution according to Listeria-type reaction. Results Negative 51 46 97 yielded sensitivity of 12.3% (6/57) and specificity of 54.4% Total 57 101 158 (46/101). 118 Jornal de Pediatria - Vol. 76, Nº2, 2000 Tuberculin test in the diagnosis of childhood tuberculosis:... - Oliveira HMV et alii

Considering that over half of our population included procedures, tuberculin tests could be useful in corroborating children with less than 2 years of age, that is, 51% of group he diagnosis of tuberculosis - as long as it is used in 1 and 54% of group 2, and that at 2 years of age there still combination with clinical, radiological, and epidemiological may be influence of BCG over tuberculin tests,3,8,10 it is results. important to emphasize that BCG did not affect our tests. In We would also like to emphasize that it is very important analyzing group 2, whose 101 patients were all BCG- to observe proper training of healthcare professionals for vaccinated, it is possible to observe in Table 3 that reactions quantitative and qualitative assessment of tests. to tuberculin test of over 10 mm corresponded to just 10% of the total. Reactions in all the other patients were lower than 10 mm. In this sense, however, Mantoux’s reaction of Acknowledgments children with tuberculosis (group 1) indicated that 86% were of over 10 mm. This suggests that stimulus caused by We would like to thank professors José Uéleres Braga M. tuberculosis was more intense than that caused by the and Antonio Ruffino-Netto for their expert contribution, vaccine. and professors Gerson Carakushansky, Aunir José Carneiro, and Roni Marques for their support. In considering the qualitative assessment of tuberculin test, that is, the assessment of the form of induration independently from its size, we observed an important increase in specificity. As we mentioned earlier, Koch-type reactions are related to hypersensitivity, observed in tuberculosis illness or infection; whereas Listeria-type References reactions are related to protective immunity of intracutaneous 1. Ruffino Netto A. Prova tuberculínica. Rev Ass Med Brasil 1979; BCG vaccination. Thus it is possible to assess protective 25:257-9. immunity of BCG by observing the occurrence of Listeria- 2. Carneiro JF. Contribuição ao estudo da alergia tuberculínica. 9 type reactions. Rev Serv Nac Tub 1964; 8:31. Based on the fact that Koch- and Listeria-type reactions 3. ten Dam HG, Hitz KL. Determining the prevalence of tuberculosis were identified in children with induration greater than or infection in population with no-specific tuberculin sensitivity. Bull WHO 1980; 58:475. equal to 4 mm, it is possible to understand the practical 4. Camacho LAB. Estimativa do risco de infecção tuberculosa em applicability of this type of assessment even in children população com alta cobertura vacinal com BCG [dissertation]. 6 considered non-responsive to Mantoux’s test. Escola Nacional de Saúde Pública; 1988.186p. We observed that the qualitative assessment of Koch- 5. Arantes GR, Nardy SMC, Weiler RMG, Belluomini M, Nogueira type reactions in tuberculin tests, due to its high specificity, PA. Estimativa do risco de infecção tuberculosa em populações is important for the diagnosis of childhood tuberculosis. In vacinadas pelo BCG. Rev Saúde Pública (São Paulo) 1992; 26:96-107. this sense, findings of Listeria-type reactions would allow 6. Stanford JL, Lema E. The use of a soniccate preparation of ruling out M. tuberculosis infection. Mycobacterium tuberculosis (new tuberculin) in the assessment In general, pediatricians may face difficulties in assessing of BCG vaccination. Tubercle 1983; 64:275-82. tuberculin tests in children vaccinated with BCG. Even 7. Stanford JL, Shield MJ, Rook GAW. How environmental though there is a tendency of tuberculin reactions towards mycobacteria may predetermine the protective efficacy of BCG. Tubercle 1981; 62:55-61. decreasing its size with time, recent BCG revaccination 8. Sant’Anna CC. Tuberculose na infância. Rotina de diagnóstico campaigns in children at school entry age have reestablished e tratamento. J Bras Med 1998; 74:100-8. the difficulty in assessing tuberculin tests because of the 9. Aroson JD, Aroson CF. The correlation of the tuberculin reaction renewed antigenic stimulus. In assessing tuberculin tests with roentgenographically demonstrable pulmonary lesion in qualitatively, we hope to contribute to the understanding of BCG vaccinated and control persons. Am Rev Resp Dis 1953; the importance of Mantoux reaction despite previous 68:713-8. intracutaneous BCG vaccination in children. This study 10. Starke JR, Correa AG. Management of mycobacterial infection and disease in children. Pediatr Infect Dis J 1995; 14:455-70. does not, however, assess children and adolescents revaccinated with BCG. Responses to tuberculin tests among revaccinated subjects have not been extensively studied, and it is possible that Koch-type reactions may be found in individuals revaccinated with BCG.

Correspondence: Recommendations Dra. Henriqueta M.V. Oliveira We recommend carrying out tuberculin tests with PPD Rua Aristides Caire 172 – Meier RT-23 at two tuberculin units, measuring induration and CEP 20775-090 – Rio de Janeiro, RJ, Brazil checking for Koch-type reactions. In following these Phone/fax: + 55 21 501.0417