Secondary Logo

Journal Logo

INFECTIOUS DISEASE: Original Article

Risk Factors for Positive Tuberculin Skin Test in Guinea-Bissau

Gustafson, Per*†‡; Lisse, Ida†‡; Gomes, Victor†‡§; Vieira, Cesaltina S.§; Lienhardt, Christian; Nauclér, Anders*†‡; Jensen, Henrik†‡; Aaby, Peter†‡

Author Information
doi: 10.1097/01.ede.0000259987.46912.2b

Abstract

ArticlePlus

Click on the links below to access all the ArticlePlus for this article.

Please note that ArticlePlus files may launch a viewer application outside of your web browser.

The tuberculin skin test (TST) was introduced in the beginning of the 20th century and is one of the oldest and most widely used immunologic tests.1 It is being used for diagnosing active tuberculosis, for estimating the prevalence of tuberculosis infection in populations, for identifying individuals in need of prophylactic treatment, and for tracing of tuberculosis transmission to contacts of cases with active and potentially infectious tuberculosis.

The present study was performed in Bissau, the capital of Guinea-Bissau, a country on the African west coast with approximately 1.3 million inhabitants. The country has a high incidence of smear-positive tuberculosis (over 200/100,000 among adults).2 The Bacille Calmette Guerin (BCG) vaccination coverage in the area is high; more than 98% of children are vaccinated in infancy.3 We evaluated potential risk factors for positive TST in children and adults who were in close contact with an infectious case. We also considered the factors associated with a positive skin test in community controls, (ie, in persons without known recent contact with smear-positive tuberculosis cases).

METHODS

The population of the study area has been followed regarding tuberculosis since 1996, using active and passive case finding.2 The present study was performed within a larger case-control study investigating the risk factors for active tuberculosis in adults. The general design of the study has been described in detail.4 Adults aged 15 years and older living in Bissau, Guinea-Bissau, and with newly diagnosed pulmonary tuberculosis were recruited and investigated at Hospital Raoul Follereau, the national referral hospital for tuberculosis, between May 1999 and November 2000. Direct microscopy was performed on sputum smears from 3 consecutive morning sputum samples. Two or more smears positive for acid fast bacilli were required for inclusion in the study. Sputum smears were graded according to the guidelines of the International Union Against Tuberculosis and Lung Disease5: “scanty” for 1–9 acid fast bacilli per 100 oil immersion fields, “1+” for 10–99 per 100 fields, “2+” for 1–10 per 1 field and “3+” for more than 10 per 1 field. These tuberculosis cases are denoted “index cases” in the present paper.

A field assistant visited the household of each index case, a census of the members was tabulated, and socioeconomic and demographic information was recorded. A nurse interviewed each family member regarding disease history and exposure to the index case, checked for presence of BCG scar, and performed a tuberculin skin test (TST) using the Mantoux technique (2TU of RT23, Statens Serum Institut, Copenhagen, Denmark).1 The width and length of the induration were measured 48–72 hours later. The mean of these 2 diameters was used for analysis. An induration of 10 mm or more was considered a positive skin reaction and possibly reflecting tuberculosis infection.

For each index case, we recruited a community control matched within 10-year age bands and living in a randomly selected household in the neighborhood of the index case. Family members of control households were investigated the same way as family members of index case households. The family members are denoted “case-contacts” and “control-contacts,” respectively.

For 244 cases we recruited 231 age-matched controls. A total of 24 cases and 8 controls lived alone and therefore were not included in the analysis.

Houses in the study area are 1-storey, unattached, rectangular buildings, usually with 6–8 rooms and inhabited by 2–4 families. The house is usually owned by 1 of these families. The majority of houses do not have an internal ceiling; this leaves a gap between the internal walls and the roof allowing air to circulate freely among all the rooms.

Statistical Analyses

Potential risk factors for positive TST reaction among case- and control-contacts were evaluated separately for children (<15 years) and adults in multivariable logistic regression models using the method of Generalized Estimating Equations.6 This analysis approach was adopted to take into account the clustering of contacts within households. A compound symmetry structure was used as the working correlation structure, and results are expressed as odds ratios (ORs) with 95% confidence intervals (95% CIs) using empirical standard-error estimates. In a combined analysis of case-contacts and control-contacts, we assessed statistical interaction between type of contacts and risk factors. This was done one-by-one and simultaneously controlling for all other risk factors. The cases and controls themselves were excluded from all analyses. Statistical analyses were performed using SAS for Windows, version 8.0 (SAS Institute, Cary, NC).

Ethical Considerations

Informed consent was acquired from all participants before enrollment. Counseling was offered before and after HIV testing. The study was approved by the Ministry of Health in Guinea-Bissau and by the Central Ethical Committee of Denmark.

RESULTS

After exclusion of single households, 220 index case households and 223 community control households were enrolled in the study. Fifty-eight percent (127/220) of the cases and 46% (103/223) of the controls were men. The mean (±SD) age of the cases was 36.7 years (±13.5) and 35.2 years (±12.0) for the controls. There were a total of 1503 members in the case households and 1331 in the control households; tuberculin skin testing was performed in 1059 (70%) case-contacts and 921 (69%) control-contacts. Forty-two percent (443/1059) of the tested case-contacts and 38% (353/921) of control-contacts were men, for a prevalence ratio (PR) of 1.07 (95% CI = 0.98–1.16). The mean age was 19.8 (±16.1) for the case contacts and 15.8 (±14.8) for the control contacts (P < 0.001). BCG scar prevalence was similar among case and control contacts (Table 1).

T1A-9
TABLE 1:
Prevalence of Positive Tuberculin Skin Test Reactions (≥10 mm) for Various Risk Factors in Case- and Control-Contacts, for Children (Age <15 Years) and Adults
T1B-9
TABLE 1:
(Continued)

We assessed the risk factors for positive skin reaction in case households (in which exposure to pulmonary smear-positive tuberculosis was documented), and among control-contacts as a reflection of the general population. The prevalence of positive reactions for various risk factors is presented in Table 1 and the multivariable estimates are shown in Table 2.

T2-9
TABLE 2:
Multivariable Analysis of Risk Factors for Tuberculin Skin Test Reaction ≥10 mm in Case and Control Households, Separated for Adults and Children (Age <15 Years) and Adults

The prevalence of positive skin reaction was higher in case contacts compared with control contacts, 41% (437/1059) versus 22% (201/921) (PR = 1.48; 95% CI = 1.37–1.60). Of the case contacts, 45% (480/1059) had no reaction to tuberculin at all, compared with 68% (623/921) of the control contacts (PR = 0.66; 95% CI = 0.61–0.72). The distribution of skin test results in contacts in both case and control households showed typical bimodal patterns (Figs. 1–4 available with the online version of this article).

Both children and adults living in contact with a tuberculosis case had higher risks of positive skin reactions than those living in contact with a community control (adjusted OR = 2.47 [95% CI = 1.78–3.42] for children and 1.89 [1.31–2.72] for adults). The prevalence of positive skin reaction among case-contacts was similar in males and females among both children and adults (Table 2, Fig. 5 available with the online version of this article). Among the control-contacts, males had a lower risk of positive skin tests than females until the age of 14, after which they had higher risks (Table 2, Fig. 6 available with the online version of this article). The risk of a positive skin test was higher among older children than among younger children, in both case and control households (Table 2).

Compared with the control-contacts, female case-contacts had a higher risk of being skin-test-positive than male case-contacts (adjusted OR = 2.35 [1.62–3.39] for women and 0.96 [0.52–1.79] for men). For children there was, however, no difference between males and females in this regard, (3.54 [2.01–6.25] for girls and 4.20 [2.03–8.68] for boys). It thus appears as if women were more likely than men to express a positive skin reaction after exposure. As shown in Table 3, the prevalence of positive skin reaction was higher for women who lived either with a tuberculosis case or with a skin-positive control, as compared with living with a skin test-negative control. This was not true for men.

T3-9
TABLE 3:
Effect of Exposure for Tuberculin Skin Test Positivity in Women and Men (Adults Only)

Contacts in case and control households who were tested during the early rainy season (from June to August) tended to have a lower risk of being skin test positive compared with those tested during the rest of the year (OR = 0.79 [CI = 0.57–1.10]). Consistent with this trend, the risk of having no skin test reaction during June to August was higher than for those tested during the rest of the year (1.45 [1.03–2.29]).

In neither case- nor control-contacts could we establish a correlation between positive TST reaction and markers of crowding, such as the number of persons living in the households or the size of the dwelling. Potential markers of economic situation, such as ownership of the house or presence of animals in the house had no impact on skin test reactivity (Table 2).

Sleeping in the same bed or in the same room as the index case increased the risk of positive skin test response in case-contacts. However, risk of positive skin test reaction was not increased by the proximity to the case during daytime, by the bacterial load of the case's sputum, or by the sex or age of the index case, in either children or adults (Table 2).

Potential markers of exposure to tuberculosis were associated with skin test reaction also among the control contacts. A history of tuberculosis in the family was found to increase the risk of a positive skin test in both children and adults, and a positive skin test reaction in the control increased the risk among the adult contacts (Table 2). Furthermore, if the control was HIV-positive then the adult contact had a considerably reduced risk of being tuberculin-skin-test-positive. Contacts of HIV-positive tuberculosis cases were also somewhat less likely to have a positive skin-test response. Presence of a BCG scar was found to be a risk factor for a positive skin test in adults in the case households, but this effect was not seen in any of the other groups.

Among adult control contacts the risk of positive skin reaction appeared to be higher if the control person was a man (Table 2), but this association did not persist after adjusting for HIV status (OR = 0.99; CI = 0.57–1.72).

DISCUSSION

We estimated the prevalence of positive tuberculin skin-test reactions among family members in close and recent contact with a smear-positive tuberculosis case, and in family members without such recent contact, and we investigated potential risk factors for positive skin-test responses. We have presented data using a TST reaction of 10 mm or more as positive in both case and control contacts. Performing the analyses with 5 mm as the cut-off point for case contacts, as recommended by some tuberculosis programs, had no impact on the results. Several studies have shown that contact with tuberculosis cases increases the risk of a positive skin test reaction.7–10 In our data, both children and adults with recent and close exposure to smear-positive tuberculosis cases were more likely to have positive skin test than those who did not have such contact. The prevalence was greater for case-contacts than for control-contacts of skin reactions (regardless of size).

In persons without recent tuberculosis contact, the prevalence of positive tuberculin skin-test reaction after the age of 15 was higher among men than women. Overall, men in the community control households had twice the risk of positive skin tests compared with women. The finding that men in the general population had a higher prevalence of positive skin-test reactions than women has also been reported in several studies from both Africa10,11 and Europe.12 However, in our study, among persons in close contact with a tuberculosis case, there was no sex difference in the prevalence of positive skin test reactions, suggesting that women are more prone to develop a positive reaction once exposed. In contrast to men, women were more likely to be skin-test-positive when living in contact with a tuberculosis case or a positive control, compared with living with a skin-test-negative control. However, given that the tuberculin test is mainly an indicator of past exposure, there ought to be more tuberculin positive women in the general population as well, if women are truly more likely to react to exposure as indicated above. To reconcile these findings, we suggest that women might have a better primary or booster response to tuberculosis than men; women may therefore become less severely infected and may be less likely to maintain the same level of tuberculin positivity over time, compared with men. Whatever the mechanism, these trends suggest that there are major differences in the way men and women respond to tuberculosis exposure.

Crowding or economic status had no apparent association with skin test responses, either in adults or children, in either case or control households. The intensity of exposure13–15 and bacterial load14–16 have previously been reported to be important risk factors for skin test positivity. We could not establish an effect of the infectivity of the tuberculosis case as assessed through the bacterial load in sputum, similar to a study performed in The Gambia.10 However, according to previous studies,15–17 the major difference in infectivity lies between those who are smear-positive and those who are positive in culture only. It should be emphasized that all tuberculosis cases in our study were positive in direct microscopy in at least 2 sputum samples. Our study shows that close contact to the case during the night increases the risk in both children and adults. Among controls, a reported history of tuberculosis in the family increased the risk of positive skin test reaction in both children and adults. Furthermore, adult contacts had an increased risk of skin test positivity if the control person was positive as well, presumably due to similar exposure to stimuli that might render positive reactions.

Data regarding the influence of BCG vaccination on TSTs are conflicting. A meta-analysis showed that immunization with BCG increased the risk of a positive skin test,18 although several studies have shown that the skin test reaction wanes with time after BCG vaccination.7,9,11,12,18,19 In our study, we could not establish a correlation between the presence of a BCG scar and a positive skin test reaction among the family members in the control households or in children living with tuberculosis cases. However, in adults living in case households, an association between the presence of a BCG scar and a positive skin-test reaction was observed (adjusted OR = 1.45; 95% CI = 1.01–2.07). This finding may be due to a boosting phenomenon, as adults with a BCG scar had larger skin reactions compared with adults who did not have a scar (mean ± SD = 9.4 mm ± 7.4 and 7.9 ± 7.2, respectively; P = 0.02). We were not able to confirm vaccination status with vaccination cards and had to use the presence of a BCG scar instead. Because of potential problems with scar reading, results should be interpreted with caution. Traumatic scars and scars from smallpox vaccination may be misclassified as BCG scars.20 We tried to minimize this problem by using assistants who were trained in scar reading and had performed scar reading surveys in the area before. It is also well known that subjects with a recorded BCG vaccination do not all develop and retain a recognizable scar.21–23

Even though the TST is the most commonly used test to assess potential tuberculosis infection, several papers have pointed out that there are limitations to using this skin test for that purpose.1,19,24 False-positive reactions may occur due to cross-reaction with other mycobacteria sharing the same antigens,25 such as environmental nontuberculous mycobacteria,26 or as a result of BCG vaccination.27–29 Similarly, repeated tuberculin testing may result in boosting phenomena.30 Conversely, a negative skin test result cannot rule out tuberculosis infection in situations of immunosuppression, which may lead to a hampered reaction to tuberculin. HIV-infection,31 as well as other viral32 or bacterial33 infections, or even a recent vaccination with live virus,34–36 as well as malnutrition37,38 may reduce sensitivity to tuberculin. Seasonality in active tuberculosis has previously been reported39,40 and we found that season might be 1 of the immunosuppressive factors that affects the interpretation of skin test responses. The risk of negative skin test response was increased during the beginning of the rainy season, from June to August, compared with the rest of the year. Such a pattern has been demonstrated in studies previously performed in Bissau, where children who were tuberculin-tested during the early rainy season had lower frequencies of positive reactions.3,41 This may be due to a reduction in the cell-mediated immunity during the rainy season,42 or to the fact that other common infections during the rainy season, such as malaria43 and respiratory infections,44 may have a negative effects on the immune response. Shaheen et al41 and Garly et al3 have reported a seasonal effect when adjusting for such infections. TST performed during the early rainy season may cause false negative reactions; and extra care should be taken when interpreting the result. Follow-up testing may be required in negative case contacts.

Newer serological tests, such as whole-blood interferon γ (IFN-γ) assays, may in the future play an important role in diagnosing tuberculosis infection. Studies from high-endemic populations have shown high agreement between TSTs and IFN-γ-assay, both in children and adults, an advantage for the newer tests being that they do not give false-positive reactions to prior BCG vaccination.45,46 The newer tests are, however, still hampered by the higher cost and the need for laboratory infrastructure. In our study we found that a positive TST is closely related to tuberculosis exposure, and that having a BCG scar had no impact on the skin-test result for unexposed persons. We believe that the inexpensive and easily administered TST remains a useful tool for diagnosing tuberculosis infection in high-endemic populations.

REFERENCES

1. Huebner RE, Schein MF, Bass JB Jr. The tuberculin skin test. Clin Infect Dis. 1993;17:968–975.
2. Gustafson P, Gomes VF, Vieira CS, et al. Tuberculosis in Bissau: incidence and risk factors in an urban community in sub-Saharan Africa. Int J Epidemiol. 2004;33:163–172.
3. Garly ML, Bale C, Martins CL, et al. BCG vaccination among West African infants is associated with less anergy to tuberculin and diphtheria-tetanus antigens. Vaccine. 2001;20:468–474.
4. Lienhardt C, Bennett S, Del Prete G, et al. Investigation of environmental and host-related risk factors for tuberculosis in Africa. I. Methodological aspects of a combined design. Am J Epidemiol. 2002;155:1066–1073.
5. International Union Against Tuberculosis and Lung Disease. Technical Guide: Sputum Examination for Tuberculosis by Direct Microscopy in Low Income Countries. 5th ed. Paris: International Union Against Tuberculosis and Lung Disease; 2000.
6. Diggle PJ, Liang K-Y, Zeger SL. Analysis of Longitudinal Data. New York: Oxford Science Publications; 1994.
7. Lockman S, Tappero JW, Kenyon TA, et al. Tuberculin reactivity in a pediatric population with high BCG vaccination coverage. Int J Tuberc Lung Dis. 1999;3:23–30.
8. Narain R, Nair SS, Rao GR, et al. Distribution of tuberculous infection and disease among households in a rural community. Bull World Health Organ. 1966;34:639–654.
9. Saiman L, San Gabriel P, Schulte J, et al. Risk factors for latent tuberculosis infection among children in New York City. Pediatrics. 2001;107:999–1003.
10. Lienhardt C, Fielding K, Sillah J, et al. Risk factors for tuberculosis infection in sub-Saharan Africa: a contact study in The Gambia. Am J Respir Crit Care Med. 2003;168:448–455.
11. Fine PE, Bruce J, Ponnighaus JM, et al. Tuberculin sensitivity: conversions and reversions in a rural African population. Int J Tuberc Lung Dis. 1999;3:962–975.
12. Jentoft HF, Omenaas E, Eide GE, et al. Tuberculin reactivity: prevalence and predictors in BCG-vaccinated young Norwegian adults. Respir Med. 2002;96:1033–1039.
13. Lienhardt C, Sillah J, Fielding K, et al. Risk factors for tuberculosis infection in children in contact with infectious tuberculosis cases in the Gambia, West Africa. Pediatrics. 2003;111(5 Pt 1):e608–e614.
14. Almeida LM, Barbieri MA, Da Paixao AC, et al. Use of purified protein derivative to assess the risk of infection in children in close contact with adults with tuberculosis in a population with high Calmette-Guerin bacillus coverage. Pediatr Infect Dis J. 2001;20:1061–1065.
15. Grzybowski S, Barnett GD, Styblo K. Contacts of cases of active pulmonary tuberculosis. Bull Int Union Tuberc. 1975;50:90–106.
16. Shaw JB, Wynn-Williams N. Infectivity of pulmonary tuberculosis in relation to sputum status. Am Rev Tuberc. 1954;69:724–732.
17. van Geuns HA, Meijer J, Styblo K. Results of contact examination in Rotterdam, 1967–1969. Bull Int Union Tuberc. 1975;50:107–121.
18. Wang L, Turner MO, Elwood RK, et al. A meta-analysis of the effect of Bacille Calmette Guerin vaccination on tuberculin skin test measurements. Thorax. 2002;57:804–809.
19. Snider DE Jr. Bacille Calmette-Guerin vaccinations and tuberculin skin tests. JAMA. 1985;253:3438–3439.
20. Fine PE, Ponnighaus JM, Maine N. The distribution and implications of BCG scars in northern Malawi. Bull World Health Organ. 1989;67:35–42.
21. Grindulis H, Baynham MI, Scott PH, et al. Tuberculin response two years after BCG vaccination at birth. Arch Dis Child. 1984;59:614–619.
22. Elliott A, Bradley AK, Tulloch S, et al. Tuberculin sensitivity in rural Gambian children. Ann Trop Paediatr. 1985;5:185–189.
23. Young TK, Mirdad S. Determinants of tuberculin sensitivity in a child population covered by mass BCG vaccination. Tuber Lung Dis. 1992;73:94–100.
24. Rieder HL. Methodological issues in the estimation of the tuberculosis problem from tuberculin surveys. Tuber Lung Dis. 1995;76:114–121.
25. Daniel TM, Janicki BW. Mycobacterial antigens: a review of their isolation, chemistry, and immunological properties. Microbiol Rev. 1978;42:84–113.
26. Edwards LB, Acquaviva FA, Livesay VT, et al. An atlas of sensitivity to tuberculin, PPD-B, and histoplasmin in the United States. Am Rev Respir Dis. 1969;99(Suppl):1–132.
27. Stewart CJ. Skin sensitivity to human, avian and BCG PPDs after BCG vaccination. Tubercle. 1968;49:84–91.
28. Sepulveda RL, Burr C, Ferrer X, et al. Booster effect of tuberculin testing in healthy 6-year-old school children vaccinated with Bacillus Calmette-Guerin at birth in Santiago, Chile. Pediatr Infect Dis J. 1988;7:578–581.
29. Miret-Cuadras P, Pina-Gutierrez JM, Juncosa S. Tuberculin reactivity in Bacillus Calmette-Guerin vaccinated subjects. Tuber Lung Dis. 1996;77:52–58.
30. Thompson NJ, Glassroth JL, Snider DEJr, et al. The booster phenomenon in serial tuberculin testing. Am Rev Respir Dis. 1979;119:587–597.
31. Garcia-Garcia ML, Valdespino-Gomez JL, Garcia-Sancho C, et al. Underestimation of Mycobacterium tuberculosis infection in HIV-infected subjects using reactivity to tuberculin and anergy panel. Int J Epidemiol. 2000;29:369–375.
32. Bloomfield AL, Mateer JG. Changes in skin sensitiveness to tuberculin during epidemic influenza. Am Rev Tuberc. 1919;3:166–168.
33. Mitchell AG, Nelson WE, LeBlanc TJ. Studies in immunity. V. Effect of acute diseases on the reaction of the skin to tuberculin. Am J Dis Child. 1935;49:695–702.
34. Starr S, Berkovich S. Effects of measles, gamma-globulin-modified measles and vaccine measles on the tuberculin test. N Engl J Med. 1964;270:386–391.
35. Brody JA, Overfield T, Hammes LM. Depression of the tuberculin reaction by viral vaccines. N Engl J Med. 1964;271:1294–1296.
36. Brody JA, McAlister R. Depression of tuberculin sensitivity following measles vaccination. Am Rev Respir Dis. 1964;90:607–611.
37. Harland PS. Tuberculin reactions in malnourished children. Lancet. 1965;2:719–721.
38. Sinha DP, Bang FB. Protein and calorie malnutrition, cell-mediated immunity, and B.C.G. vaccination in children from rural West Bengal. Lancet. 1976;2:531–534.
39. Thorpe LE, Frieden TR, Laserson KF, et al. Seasonality of tuberculosis in India: is it real and what does it tell us? Lancet. 2004;364:1613–1614.
40. Leung CC, Yew WW, Chan TY, et al. Seasonal pattern of tuberculosis in Hong Kong. Int J Epidemiol. 2005;34:924–930.
41. Shaheen SO, Aaby P, Hall AJ, et al. Cell mediated immunity after measles in Guinea-Bissau: historical cohort study. Br Med J. 1996;313:969–974.
42. Lisse IM, Aaby P, Whittle H, et al. T-lymphocyte subsets in West African children: impact of age, sex, and season. J Pediatr. 1997;130:77–85.
43. Akinwolere OA, Williams AI, Akinkugbe FM, et al. Immunity in malaria: depression of delayed hypersensitivity reaction in acute Plasmodium falciparum infection. Afr J Med Sci. 1988;17:47–52.
44. Kauffman CA, Linnemann CCJr, Schiff GM, et al. Effect of viral and bacterial pneumonias on cell-mediated immunity in humans. Infect Immun. 1976;13:78–83.
45. Pai M, Gokhale K, Joshi R, et al. Mycobacterium tuberculosis infection in health care workers in rural India: comparison of a whole-blood interferon gamma assay with tuberculin skin testing. JAMA. 2005;293:2746–2755.
46. Dogra S, Narang P, Mendiratta K, et al. Comparison of a whole blood interferon-gamma assay with tuberculin skin testing for the detection of tuberculosis infection in hospitalized children in rural India. J Infect. In press.

Supplemental Digital Content

© 2007 Lippincott Williams & Wilkins, Inc.