PEDIATRICS Vol. 108 No. 2 August 2001, pp. 305-310

Risk Factors for Positive Mantoux Tuberculin Skin Tests in Children in San Diego, California: Evidence for Boosting and Possible Foodborne Transmission

Richard E. Besser, MD^*, , Bilge Pakiz, EdD^*, Joann M. Schulte, DO^§, Sonia Alvarado^*, Elizabeth R. Zell, MStat, Thomas A. Kenyon, MD, MPH^§, and Ida M. Onorato, MD^§

From the ^* Department of Pediatrics, University of California, San Diego School of Medicine, San Diego, California;  Division of Bacterial and Mycotic Diseases, National Center for Infectious Diseases, and ^§ Tuberculosis Elimination, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objectives.  Source case finding in San Diego, California, rarely detects the source for children with tuberculosis (TB) infection or disease. One third of all pediatric TB isolates in San Diego are Mycobacterium bovis, a strain associated with raw dairy products. This study was conducted to determine risk factors for TB infection in San Diego.

Design.  Case-control study of children 5 years old screened for TB as part of routine health care visit. Asymptomatic children with a positive (10 mm) Mantoux skin test (TST) were matched by age to 1 to 2 children with negative TST from the same clinic. We assessed risk factors for TB infection through parental interview and chart review.

Results.  A total of 62 cases and 97 controls were enrolled. Eleven cases and 25 controls were excluded from analysis because of previous positive skin tests. Compared with controls, cases were more likely to have received BCG vaccine (73% vs 7%, odds ratio [OR] 44), to be foreign born (35% vs 11%, OR 4.3), and to have eaten raw milk or cheese (21% vs 8%, OR 3.76). The median time between the most recent previous TST and the current test was 12 months for cases and 25 months for controls. Other factors associated with a positive TST included foreign travel, staying in a home while out of the country, and having a relative with a positive TST. There was no association between contact with a known TB case. In a multivariable model, receipt of BCG, contact with a relative with a positive TST, and having a previous TST within the past year were independently associated with TB infection.

Conclusions.  We identified several new or reemerging associations with positive TST including cross border travel, staying in a foreign home, and eating raw dairy products. The strong associations with BCG receipt and more recent previous TST may represent falsely positive reactions, booster phenomena, or may be markers for a population that is truly at greater risk for TB infection. Unlike studies conducted in nonborder areas, we found no association between positive TB skin tests and contact with a TB case or a foreign visitor. Efforts to control pediatric TB in San Diego need to address local risk factors including consumption of unpasteurized dairy products and cross-border travel. The interpretation of a positive TST in a young child in San Diego who has received BCG is problematic.  Key words:  tuberculosis, epidemiology, Bacille Calmette-Guérin, Mycobacterium tuberculosis, Mycobacterium bovis..

To appropriately design tuberculosis (TB) control strategies, one must have an understanding of local risk factors for infection and disease. In 1996, the American Academy of Pediatrics issued recommendations for TB screening of children, which moved from a policy of universal screening to one based on risk.¹ Young children with TB infection represent recent, ongoing transmission in a community. They are also at greatest risk for activation of their infections, and of developing disseminated disease.^2-6 As such, they are important targets for prevention efforts.

San Diego County, California, has been recognized by the Centers for Disease Control and Prevention as one of the 12 highest TB incidence areas in the country.⁷ Between 1985 and 1992, reported cases of TB in children <20 years old increased by 400% from 15 cases in 1985 to 75 cases in 1993.⁸ This rise occurred after a steady decline in pediatric TB during the preceding 3 decades. The southern boundary of San Diego County borders Mexico and the city of Tijuana. The San Diego-Tijuana area is the nation's most active international area with 65 million legal and over 1 million illegal crossings each year.⁹ The bidirectional flow of individuals across the border links the 2 regions closely and makes the control of TB in San Diego increasingly challenging.

Each year 35 to 50 cases of TB disease and 700 to 1000 cases of TB infection are diagnosed in children  5 years of age in San Diego (unpublished data, San Diego Department of Health). In San Diego, between 1980 and 1997, 34% of pediatric TB cases from which isolates were recovered were caused by Mycobacterium bovis.¹⁰ The rarity of M bovis infections in adults in San Diego suggests that these infections are acquired through the ingestion of contaminated, unpasteurized dairy products rather than via the respiratory route from actively infected adults.¹¹ Because of this difference in route of transmission, prevention strategies designed for Mycobacterium tuberculosis are unlikely to be effective for disease caused by M bovis.

The location of San Diego on the border with Mexico and the high proportion of disease caused by M bovis, might suggest that the risk factors for TB transmission in this population are different from those found in other studies.^12-14 We conducted a study of risk factors for TB infection in young children in San Diego to address, in particular, whether border-related activities such as travel, raw dairy consumption, and BCG-receipt are associated with having a positive Mantoux skin test (TST).

	METHODS

Top Abstract Methods Results Discussion Conclusion References

Study Design

Study participants were enrolled at 5 pediatric clinics serving predominately low-income, Hispanic families in San Diego County. All children younger than 6 years of age, being skin tested as part of routine well-child care, were eligible for enrollment in the study. The Child Health and Disability Prevention Program, a program that provides free complete health examinations and immunizations for the majority of children in these clinics, recommends routine annual Mantoux skin testing, although compliance with this recommendation is unknown. Children were ineligible to participate in this study if the skin test was being placed as part of a TB contact investigation or as part of an evaluation of symptoms felt to be compatible with TB, or if they had been evaluated as part of a previous contact investigation.

Tuberculin skin testing was performed by the Mantoux method (0.1 mL or 5-TU strength purified protein derivative).¹⁵ Skin tests were placed by nurses in the clinics and were read 48 to 72 hours later by a trained health care worker. A test was considered to be positive if the area of induration measured at least 10 mm at the largest diameter. In these clinics, as part of the evaluation of children with positive skin tests, chest radiographs were obtained, and reports were made to the health department so that investigations could be undertaken to determine the source of the infection.

Each clinic maintained a log of all children being tested for TB. A study worker reviewed the logbooks regularly to ensure completeness of enrollment. A case was defined as a child with a TST and no evidence clinically or radiographically of TB. Each case was matched to 2 controls in the same clinic by age (± 6 months for children 0-<3 years, ± 12 months for children 3-5 years). Controls were selected by reviewing the TST logbook and contacting in order the most recently tested age-eligible children seen in the clinic in the preceding month. If 2 controls could not be found during that time period, the study worker contacted eligible children from the preceding and subsequent months. These efforts were continued until either 2 controls were matched or 1 control had been found and 6 months had elapsed.

After obtaining informed consent, the study worker reviewed cases' and controls' medical records including information on patient demographics, birth history, immunizations, medical history, hospitalizations, medications, growth parameters, and previous TB testing. A questionnaire was administered in person to the child's parent or guardian in either English or Spanish. This questionnaire covered the same information as on the medical record review, as well as information on travel, dietary practices, contacts with TB cases, use of medical services out of the country, receipt of BCG, and immigration status. Each child was examined for the presence of a BCG scar.

This study was reviewed and approved by the institutional review boards at the University of California, San Diego, and the Centers for Disease Control and Prevention.

Statistical Analysis

Data abstraction and interview forms were checked for completeness by the study coordinator and then double-entered and validated in Epi Info (Centers for Disease Control and Prevention, Atlanta, GA).¹⁶ During analysis, all patients who had documentation of a positive skin test in the past were excluded. In addition, patients with a history of a positive skin test, for whom documentation of the skin test results were unavailable, were also excluded. We initially analyzed the data maintaining the match between individual cases and controls. This was compared with an analysis stratifying cases and controls into 2 age groups, <48 months and 48 months, and 2 combined groups based on clinic demographics. The age groups were selected to reflect the local requirement for skin testing between the ages of 4 and 6 years for school entry and closely resembled the initial age-matching as specified in the study design. The 2 clinic groups were created by combining clinics with similar patient demographics in terms of race, ethnicity, and types of insurance. Because the point estimates from these 2 analyses did not vary, we are presenting the more powerful results of the combined analysis.¹⁷

In the univariate analysis, we compared risk factors between cases and controls using the stratified Mantel-Haenszel odds ratio and computed 95% confidence intervals. Median length of time between current and previous skin tests was compared using the Mann-Whitney test. Factors associated with TB infection in the univariate analysis were considered for inclusion in the multivariable model. Continuous data on the length of time between previous skin tests were categorized into 4 groups (ie, never tested, <1 year, 1-2 years, >2 years) for this purpose. Never tested was included as a category on length of time between previous skin tests so as not to lose a third of our observations in the final model. Final inclusion in the model was based on biological plausibility, lack of collinearity with other variables, and robustness of the model with and without each variable. We performed conditional logistic regression maintaining the age and clinic strata as defined above, using SAS software (SAS, Inc, Cary, NC). For all statistical tests, differences were considered significant at P < .05.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Demographics

Between January 1997 and November 1998, we enrolled 62 cases and 97 controls. Eleven cases and 25 controls were excluded because documentation of a positive skin test in the past, or a history of a positive skin test with no documentation that it was in fact negative. Cases and controls did not differ with respect to age, sex, or type of insurance (Table 1). Compared with controls, cases were significantly more likely to be born outside of the United States, to have foreign-born fathers but not mothers, and to live in households in which no English is spoken. There was no difference in the presence of chronic or acute medical conditions; or the receipt of steroids or other medications.

The average family for cases was slightly larger for cases than for controls, 5 versus 4 members, respectively, although this difference did not reach statistical significance. The families consisted of 2 adults and 2 to 3 children, and lived in a private house or apartment. The majority of children shared a bed with another person.

Skin Test History

The median time between placement and reading of the TST was 2 days for cases and controls. The median induration for cases was 13 (range: 10-27); 2 controls had any induration: one 6 mm and one 8 mm (Table 2). There was no relationship between the length of time since receipt of BCG and the size of the induration (r² = 0.01, 95% confidence interval: 0.38, 0.40).

Cases and controls were equally likely to have received TST in the past and the number of previous tests did not differ between the 2 groups. However, the median time between the most recent previous test and the current test was significantly shorter among cases than controls (12 vs 25 months, P = .002). Among those having been skin tested in the past, there was no association between having had a skin test more recently (in the past 2 years) and travel to Mexico, Hispanic ethnicity, or having a relative with a positive skin test (data not shown).

Travel-Related Activities

Cases were significantly more likely than controls to have traveled to a foreign country and to have stayed overnight with friends or family (Table 3). Ninety-six percent of all foreign travel was to Mexico. Among travelers, there was no difference in frequency of travel or total number of trips taken.

Cases were significantly more likely to have received BCG in a foreign country and to have an upper arm scar suggestive of BCG receipt.

Cases were significantly more likely than controls to have consumed raw dairy products; however, among those consuming these products, there was no difference in the total amount of product consumed (Table 3).

TB Contact History

Compared with controls, cases were no more likely to have had contact with a known active TB case; however, they were significantly more likely to live in a house with a relative who had a previously positive TST (Table 4). The treatment rates for these TST-positive relatives were similar between cases and controls. Surprisingly, cases were less likely than controls to have had contact with foreign visitors.

Multivariable Model

A multivariable model was developed to assess the independent association of risk factors with having a positive TST. Variables that were significant in univariate analysis at a level of P < .05 were tested in the multivariable model. Receipt of BCG, having a relative with a positive TST, and having been tested <12 months before the current test were independently associated with having a positive TST (Table 5). Foreign birth was collinear with receipt of BCG and was excluded from the model. No other variables were significantly associated with having a positive TST when included in this model, nor were any other variables significantly associated with having a positive TST when included in any model containing the variable "receipt of BCG." Inclusion of additional variables did not strengthen the model.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

To prevent a disease, it is essential to understand disease transmission at the local level. Risk factors in one region or community may not predict risk factors in another community. This is clearly the case with tuberculosis in San Diego where population dynamics and behaviors are so closely integrated with the Mexican border.

This study of risk factors for having positive TST in children  5 years old found strong independent associations between positive skin tests and receipt of BCG vaccine, contact with a relative with a positive TST, and short duration between previous and current TST testing. Other associations that were present in univariate but not multivariable analysis included demographic factors such as foreign birth, foreign paternity, lack of contact with foreign visitors to the United States, and Spanish language spoken in the home; and travel-related behaviors such as foreign travel, consuming raw dairy products, and staying overnight in a home while in Mexico.

The association of positive TST with BCG receipt requires careful interpretation. One of the reasons that BCG is not recommended for routine use in the United States is because of concern that receipt will interfere with interpretation of Mantoux testing in a country with low rates of endemic TB transmission.¹⁸ Our results highlight this difficulty. Children with positive TSTs were significantly more likely than controls to have received BCG vaccination. The magnitude of this relationship made it difficult to assess the independent association of other factors that were significantly associated with TB infection in the univariate analysis.

There are several possible explanations for the association of positive skin tests with BCG receipt. Children who received BCG may come from settings where TB is highly endemic and may truly be more likely to be infected than children who did not receive BCG. They are more likely to be born in Mexico, travel to Mexico, stay in homes when they travel to Mexico, and consume raw dairy products when in Mexico. It may be that BCG receipt is a marker for children with these, and other unknown risk factors for infection that are unrelated to anything truly related to BCG. Although children being tested for TB because of exposure to known cases were ineligible for participation in this study, it is possible that recent prior tests were performed because of known exposures to TB, and that these exposures were not documented in the patient chart or obtained by parental history.

It is also possible that children who have received BCG are having false-positive reactions to the Mantoux testing. Numerous studies have attempted to address the issue of Mantoux interpretation in children who have received BCG.^19-24 Generalization from these studies is difficult given the varied populations studied, the nonuniform composition of BCG, the number of times that BCG was administered, and the varied strength of purified protein derivative used by the study investigators. A study of healthy school children in Chile demonstrated significant boosting in children who had received BCG 5 years before skin testing, if a second skin test was applied 2 weeks after the first. The authors concluded that recent preceding Mantoux testing in children who had received BCG, might explain positive reactions. The design of their study did not allow them to determine the duration of the booster effect.

A recent study of young children in Botswana found that 70% of children who were Mantoux skin tested were negative 3 to 60 months after receiving BCG vaccine.²³ However, this study, as with most studies addressing this issue, was conducted in a setting where routine Mantoux testing is not performed and therefore would be unlikely to detect boosting. It would be interesting to retest these children within the following year to see if boosting would occur.

The Centers for Disease Control and Prevention recommends that 2-step testing be considered for health care workers who will be tested frequently as part of TB control efforts.²⁵ This procedure involves a second skin test 1 to 3 weeks after the first negative test for workers who have not had a documented negative test within the past 12 months. The purpose of 2-step testing is to separate out skin test converters from persons who may have been infected years before and are experiencing boosting. Persons who are positive after the second test are considered to be "boosted" and not newly infected.

In a similar manner, it is worth studying 2-step testing in children who have received BCG to see if children who have received BCG are subject to boosting, as our results might suggest. Children who received BCG and have a negative skin test would be retested 1 to 3 weeks later. In this manner we could determine how commonly boosting occurs. This would provide valuable information on which to make recommendations for the appropriate interval between routine Mantoux tests in children from high prevalence countries who have no other risk factors for TB infection.

As the true prevalence of TB in the community continues to fall, a greater proportion of positive skin tests will be falsely positive. If we are to reduce the overuse of antituberculous medications and the potential for rare, occasionally life-threatening complications, it will be necessary to take measures to sort out children who are truly infected with M tuberculosis from those having falsely-positive reactions.

The association of TB infection with consumption of raw dairy products was not significant in the final multivariable model. This may be attributable to a true lack of association, but may also have been attributable to the overwhelming association of infection with BCG receipt, which made it difficult to evaluate other risk factors. There is much to support the association with raw dairy consumption: M bovis, a zoonotic pathogen, causes a significant amount of pediatric TB in San Diego,^10,11 and investigations of children in San Diego with confirmed TB infection and TB disease are less likely than in other cities to identify an active adult case.^26,27 In 1997, of 866 dairy herds tested for tuberculosis in Baja California, the state bordering California, 3% were positive (Dr. Alejandro Perera, United States Department of Agriculture, personal communication). United States Department of Agriculture is currently working as part of a binational taskforce to try to develop solutions to this problem. This makes it plausible that this association is not occurring by chance and that the lack of significance in the multivariable model has more to do with the limited power in this study to show an association.

Historically, the control of tuberculosis attributable to M bovis has been achieved through a combination of screening of dairy herds and pasteurization of dairy products.¹¹ This has resulted in the near elimination of this pathogen from the US dairy supply. The persistence of this pathogen in Mexican dairy herds and the emergence of M bovis as a significant pathogen in San Diego clearly demonstrate that borders are of political rather than biological relevance. Efforts to eliminate tuberculosis in San Diego will require binational cooperation and collaboration.

Current TB control efforts in San Diego do not speak to the role of raw dairy consumption in the spread of tuberculosis and will need to be addressed if this cause is to be eliminated. Although it may be difficult to control whether dairy herds are cleared of M bovis or dairy products are pasteurized, efforts can be undertaken to inform the public of the dangers of raw dairy consumption. Media campaigns targeting this risk factor would need to be in Spanish given the findings of this study. Focus groups using Mexican-American community groups might be useful in developing the most culturally appropriate messages.

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

Positive TST in children in San Diego are associated with many known and some new risk factors. Prevention efforts must target these new risk factors if true TB infections are to be prevented. Efforts must also be taken to determine if many children in San Diego who are being treated for TB infection are actually not infected.

	ACKNOWLEDGMENT

This study was funded through a contract from the Centers for Disease Control and Prevention.

	FOOTNOTES

Reprints requests to Division of TB Elimination, CDC, Mailstop E-10, Atlanta GA 30333.

Received for publication Jun 12, 2000; accepted Dec 18, 2000.

Address correspondence to Richard E. Besser, MD, Respiratory Diseases Branch, Mailstop C-23, Centers for Disease Control and Prevention, 1600 Clifton Rd, NE, Atlanta, GA 30333. E-mail: rbesser{at}cdc.gov

	ABBREVIATIONS

TB, tuberculosis, TST, Mantoux skin test.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. American Academy of Pediatrics Update on Tuberculosis Skin Testing of Children. Pediatrics. 1996; 97:282-284 [Abstract/Free Full Text]
2. Comstock GW, Livesay VT, Woolpert SF The prognosis of a positive tuberculin reaction in childhood and adolescence. Am J Epidemiol 1974; 99:131-138 [Abstract/Free Full Text]
3. Brailey M Factors influencing the course of tuberculous infection in young children. Am Rev Tuberc 1937; 36:347-354
4. American Thoracic Society and the Centers for Disease Control and Prevention Treatment of tuberculosis and tuberculosis infection in adults and children. Am J Respir Crit Care Med 1994; 149:1359-1374 [Abstract]
5. Jereb J, Kelly G, Porterfield DS The epidemiology of tuberculosis in children. Semin Pediatr Infect Dis 1994; 4:220-231
6. Starke J, Taylor-Watts K Tuberculosis in the pediatric population of Houston, Texas. Pediatrics 1989; 84:28-35 [Abstract/Free Full Text]
7. Centers for Disease Control and Prevention. Reported Tuberculosis in the United States, 1994. 1995:32-33
8. Kenyon TA, Driver C, Haas E, Valway SE, Moser KS, Onorato IM. Immigration and tuberculosis among children on the United States-Mexico border, County of San Diego, California. Pediatrics. 1999;104(1). URL: http://www/pediatrics.org/cgi/content/full/104/1/e8
9. San Diego Dialogue. Who Crosses the Border: A View of the San Diego/Tijuana Metropolitan Region. Available at: http://www.sddialogue.org
10. Dankner W, Davis C. Mycobacterium bovis as a significant cause of tuberculosis in children residing along the U. S.- Mexico border in the Baja California region. Pediatrics. 2000;105(6). URL: http://www.pediatrics.org/cgi/content/full/105/6/e79
11. Dankner WM, Waecker NJ, Essey MA, Moser K, Thompson M, Davis CE Mycobacterium bovis infections in San Diego: a clinicoepidemiologic study of 73 patients and a historical review of a forgotten pathogen. Medicine 1993; 72:11-37 [Medline]
12. Ussery X, Valway S, McKenna M, Cauthen GM, McCray E, Onorato I. Epidemiology of tuberculosis among children in the United States: 1985 to 1994. Pediatr Infect Dis J. 1996:15:697-704
13. Lobato MN, Hopewell PC Mycobacterium tuberculosis infection after travel to or contact with visitors from countries with a high prevalence of tuberculosis. Am J Respir Crit Care Med 1998; 158:1871-1875 [Abstract/Free Full Text]
14. Gessner BD, Weiss N, Nolan C Risk factors for pediatric tuberculosis infection and disease after household exposure to adult index cases in Alaska. Pediatrics 1998; 132:509-513
15. American Academy of Pediatrics, Committee on Infectious Diseases Screening for Tuberculosis in Infants and Children. Pediatrics 1994; 93:131-134 [Abstract/Free Full Text]
16. Dean AG, Dean JA, Coulombier D, et al. Epi Info, Version 6: A Word Processing, Database, and Statistics Program for Epidemiology on Microcomputers. Atlanta, GA: Centers for Disease Control and Prevention; 1994
17. Rothman K, Greenland S. Modern Epidemiology. Philadelphia, PA: Lippincott-Raven; 1998:266
18. Centers for Disease Control and Prevention. The role of BCG vaccine in the prevention and control of tuberculosis in the United States: a joint statement by the advisory council for the elimination of tuberculosis and the advisory committee on immunization practices. MMWR Morb Mortal Wkly Rep. 1996;45(RR-4):1-18
19. Fox AS, Lepow ML Tuberculin skin testing in Vietnamese refugees with a history of BCG vaccination. Am J Dis Child 1983; 137:1093-1094 [Abstract/Free Full Text]
20. Lifschitz M The value of tuberculin skin test as a screening test for tuberculosis among BCG-vaccinated children. Pediatrics 1965; 36:624-627 [Abstract/Free Full Text]
21. Thompson NJ, Glassroth JL, Snider DE, Farer LS The booster phenomenon in serial tuberculin testing. Am Rev Respir Dis 1979; 119:587-597 [Medline]
22. Karalliedde S, Katugaha LP, Uragoda CG Tuberculin response of Sri Lankan children after BCG vaccination at birth. Tubercle 1987; 68:33-38 [Medline]
23. Lockman S, Tappero JW, Kenyon TA, Runisha D, Huebner RE, Binkin NJ Tuberculin reactivity in a pediatric population with high BCG vaccination coverage. Int J Tuberc Lung Dis 1999; 3:23-30 [Medline]
24. Sepulveda R, Burr C, Ferrer X, Sorensen R 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 [Medline]
25. Centers for Disease Control and Prevention. Guidelines for preventing the transmission of mycobacterium tuberculosis in health-care facilities, 1994. MMWR Morb Mortal Wkly Rep. 1994;43(RR-13):1-132
26. Besser RE, Pakiz B, Haverkamp D, Schulte J, Moser K, Onorato I. Source Case Investigations of Children With Tuberculosis Disease. Presented at: 38th Interscience Conference on Antimicrobial Agents and Chemotherapy, American Society for Microbiology; September 24-27, 1998; San Diego, CA
27. Besser RE, Pakiz B, Havercamp D, Schulte J, Moser K, Onorato I. Source Case Investigation of Children With Tuberculosis Infection. Presented at: 38th Interscience Conference on Antimicrobial Agents and Chemotherapy, American Society for Microbiology; September 24-27, 1998; San Diego, CA