Every year worldwide about 1.5 million children develop tuberculosis and 450 000 of them die from the disease. In certain African countries where human immunodeficiency virus (HIV) infection has become pandemic, tuberculosis has again become a leading cause of disease and death among young adults and their children. The United States has not been spared from this resurgence of tuberculosis. Between 1986 and 1991, reported tuberculosis cases for all ages in the United States increased by 18%; for children, the increase has been almost 40%. J
Historically, tuberculosis case rates have increased when populations have been exposed to various stresses such as war, displacement, malnutrition, poverty, poor access to health care, and other diseases. Many experts cite several major factors contributing to the current increase of tuberculosis in the United States. First, coinfection with HIV is the strongest risk factor known for the development of tuberculous disease in an adult infected with Mycobacterium tuberculosis.1 Although few HIV-infected children with tuberculosis have been reported in the United States, many non-HlV-infected children have acquired tuberculous infection from HIV-infected adults with infectious tuberculosis.3 Second, the immigration of people from countries with a high prevalence of tuberculosis has enlarged the pool of infected individuals in the United States. School-based surveys from Los Angeles,4 Boston,5 and Houston1 have shown that up to 30% of foreign-born school-aged children in the United States may be tuberculinpositive. Third, the general decline in public health services and access to medical care in many commun i - ties has hindered the rapid identification and treatment of adults with infectious tuberculosis and the timely completion of contact investigations. Until these associated factors are addressed adequately, it is likely that the increase in tuberculosis in the United States will continue.
A positive Mantoux tuberculin skin test is the hallmark of tuberculous infection, meaning that the person harbors M tuberculosis within the body. If the individual with a positive tuberculin skin test has a negative chest radiograph, a normal physical examination, and no symptoms, he or she has tuberculous infection without disease and, usually, is a candidate for isoniazid therapy. Tuberculous disease exists when clinical manifestations of pulmonary or extrapulmonary tuberculosis become apparent, either by chest radiograph or by clinical symptoms and signs. The word tuberculosis usually refers to disease, not infection. A child suspected of having tuberculous disease will be treated initially with at least two and usually three antituberculosis medications.
The time interval between the initial tuberculous infection and the onset of disease may be several weeks or many decades. In adults, the distinction between tuberculous infection and disease is usually clear. For young children, in whom tuberculous disease develops as an immediate complication of the primary infection, the distinction is less clear. In general, a child with a positive Mantoux tuberculin skin test who has any radiographic or clinical manifestations consistent with tuberculosis is considered to have tuberculous disease, even if no symptoms are present.
In more than 98% of cases, infection due to M tuberculosis is caused by inhalation of airborne droplets produced by an adult with infectious pulmonary tuberculosis. Most children with pulmonary tuberculosis are not infectious because their bacterial burden is low, sputum production is rare, and cough, when present, is rarely forceful enough to promote transmission. Some adolescents with adult-type tuberculosis (especially if they have a pulmonary cavity or extensive upper lobe infiltrate) can transmit infection to close contacts.
In general, prolonged and close contact is required for transmission of M tuberculosis. Adults with extensive pulmonary disease, frequent cough, and heavy sputum production that is acid-fast smear positive are most likely to be infectious. A second factor determining transmission is the physical environment in which contact occurs. Buildings with poor ventilation, little sunlight and overcrowding promote the transmission of infection. An adult who would be mildly infectious in an open environment may be highly infectious in a poorly ventilated area. A third factor in transmission is the vulnerability of the contacts to tuberculosis. There may be genetic and other host-related factors that help determine the risk of both tuberculous infection and the risk of progression to tuberculous disease.6
EPIDEMIOLOGY IN THE UNITED STATES
It is estimated that 10 to 20 million persons in the United States have tuberculous infection, representing the pool from which most future cases of disease will come.7 It is unknown how many children have tuberculous infection since no nationwide surveys have been conducted since 1970, and there is mandatory reporting of tuberculous infection without disease in children from only three states - Indiana, Missouri, and Kentucky. The rate of infection for most children is probably quite low. However, recent surveys in several large US cities have detected tuberculous infection rates of 1% to 10% in some groups of school children.1·4·5
All states mandate the reporting of suspected cases of tuberculous disease for all ages. From 1953 through 1984, the incidence of tuberculosis declined about 5% per year. However, from 1985 to 1991 the yearly number of cases increased 18%; approximately 50 000 more cases were reported than would have been expected if the previous decline had continued. Of the almost 26 300 reported cases of tuberculosis in 1991, 1656 occurred in children less than 15 years of age, which represents a 40% increase over the 1177 cases in children in 1987.
In the early 20th century, the risk of tuberculosis for children was high across the entire population. Currently, most children in the United States are at very low risk for tuberculous infection and disease. However, there are some groups of children that remain at relatively high risk because of the environments in which they have lived or the risk factors of the adults with whom they have had significant contact (Table 1).
About 60% of pediatric tuberculosis cases occur in children less than 5 years of age, the group traditionally at highest risk for progression of infection to disease. The gender ratio for childhood tuberculosis is 1:1. The disease is geographically focal with seven states - California, Florida, Georgia, Illinois, New York, South Carolina, and Texas - accounting for 63% of reported cases among children. Infection and disease rates are highest within cities with populations greater than 250000. Although die majority of children with tuberculosis were born in the United States, 16% of cases in children less than 5 years of age and 49% of cases in adolescents occur among the foreign-bom.1 Six countries - China, Haiti, Mexico, the Philippines, South Korea, and Vietnam - account for the majority of foreign-bom cases. Most reported children have acquired tuberculosis in their household settings, but occasional outbreaks traced to an adult with infectious tuberculosis continue to occur in schools, family day care homes, and nursery schools.1,8
TUBERCULIN SKIN TESTING
In most children, tuberculin reactivity - a marker of delayed hypersensitivity to M tuberculosis - first appears about 3 to 8 weeks, and occasionally up to 3 months, after initial infection. When tuberculin reactivity is due to infection by M tuberculosis, it usually remains for the individual's lifetime, even after chemotherapy is given.9
Two techniques are used for tuberculin skin testing - multiple puncture tests (MPTs) and the Mantoux. The MPTs have metal or plastic prongs coated with either purified protein derivative (PPD) or Old tuberculin. They have been used widely in pediatrics, but have several major problems and limitations that should have made them obsolete. First, the exact dose of antigen delivered cannot be precisely controlled, so interpretation of the reaction size cannot be reliably standardized. As a result, MPTs cannot be used as diagnostic tests for tuberculous infection (the manufacturers even say this), and any suspicious or "positive" result must be verified with a Mantoux test. The need for a subsequent Mantoux test can lead to the second problem, which is the booster phenomenon. Boosting is an increase in the reaction to a skin test in a person already sensitized to mycobacterial antigens caused by repetitive tests. Placing a tuberculin skin test does not cause sensitization, but repeated tests can make a small reaction become larger. The incidence of boosting increases with age, in geographic areas where exposure to nontuberculous mycobacteria is common, and in bacille Calmette-Guérin (BCG)vaccinated individuals.10 Use of an MTP and a subsequent Mantoux test in a person who has a boostable reaction can create a false-positive Mantoux test.
The third important problem is that MPTs have extremely variable and often high rates of falsepositive (10% to 15%) and false-negative (5% to 10%) results compared with the Mantoux test.1 A recent study of various MPTs in high-risk SoutheastAsian refugees yielded sensitivities of 68% to 96% and specificities of 40% to 90% compared with the Mantoux test.11 This kind of performance is unacceptable, especially when testing high-risk children.
Children at High Risk for Tuberculous Infection in the United States
The fourth problem is the bad habits of medical practice often associated with use of MPTs, especially the practice of allowing parents to interpret and report test results. This horrendous practice assumes parental adherence to a broad range of motivational and practical behaviors, a "trust" that is unfounded in all practice settings. There is no other widely used test in pediatrics that is routinely interpreted by nonprofessionals. Physicians and clinics must document the result of all tuberculin tests, or such testing is a waste of time and resources.
Multiple puncture tests always are contraindicated for children with signs or symptoms of disease, children exposed to a case of tuberculosis, high-risk children, and children who have received BCG vaccine. The only group of health professionals who continue to use MPTs routinely is pediatricians; other groups have rejected these inferior tests long ago. It is time for pediatricians to abandon them also.
The Mantoux tuberculin skin test using five tuberculin units of PPD is the "gold standard* test. Testing technique must be precise and consistent. Although placing an MPT may be somewhat simpler than using the Mantoux technique, the many pediatricians in my area who now use only Mantoux tests tell me that, with practice, they can be placed fairly quickly, accurately, and almost painlessly in the office setting. A useful technique for children is to rest the tester's hand with the syringe on the side of the child's hand, in a manner similar to that often used with a cerumen spoon on the side of the head (Figure 1 ). This steadies the arm and allows the fingers to act as a fulcrum to control the angle and depth of the needle. The PPD solution is injected mtradermally in the transverse direction.
Figure 1.A useful technique for the rapid and accurate placement of a Mantoux tuberculin skin test on a child. The tester steadies her hand on the patient's arm while injecting the purified protein derivative. Reprinted with permission from Starke FR. Tuberculin skin tests: why the Mantoux? Pediatr Infect Dis J. 1993;12:623-624. Q & A. Copyright © 1993.
False-negative results with the Mantoux test can be caused by faulty technique or a variety of host-related factors such as young age, immunosuppression by drugs or disease (especially HIV infection), viral infection, and overwhelming tuberculosis.12 Falsepositive reactions can be caused by very recent immunization with the BCG vaccine, exposure to environmental non tuberculous mycobacteria, or biological variability. Most false-positive reactions are «slO mm to 12 mm, although large reactions can occur.
The suggested interpretation of the Mantoux skin test has changed recently, partly in response to the changing epidemiology of tuberculosis. The interpretation of the reaction should be influenced by the purpose for which the test was given and the consequences of false classification (Table 2). For children at the highest risk of tuberculosis, especially contacts of infectious adult cases, an indurated area Ss 5 mm is considered positive. For children in high-risk groups and very young children (because of their diminished ability to make a large reaction), induration 2=10 mm is positive. For children with no risk factors, it may be acceptable to increase the cutoff point for a positive test to 15 mm since in them most reactions less than this will be false positives.
This scheme may be scientifically valid, but it presents some problems for clinicians who deal witli children. Classifying children by risk factors depends on the willingness and ability of the practitioner and family to determine a thorough history about the child and the adults in contact with the child. This is a notoriously difficult and time-consuming task. Many public health authorities have chosen to call any reaction 5*10 mm in a child positive, especially in communities with significant tuberculosis case rates. In communities where the tuberculosis prevalence is low and tuberculin screening of children is continued, increasing the cutoff point to Ss 15 mm for truly low-risk children is a scientifically sound practice. The geographically focal nature of tuberculosis in the United States means that each local health authority should determine appropriate cutoff points for the community.
What Constitutes a Positive Mantoux Skin Test
Previous BCG vaccination has a variable influence on subsequent Mantoux tuberculin skin tests. Many children who receive BCG vaccine never develop a reactive tuberculin test. When there is a reaction, the induration is usually ^ 10 mm and it wanes after 3 to 4 years.13 Prior BCG vaccination is never a contraindication for tuberculin testing. In general, a reaction 2*10 mm in a BCG-vaccinated child - especially one from a country with a high prevalence of tuberculosis - probably indicates infection with M tuberculosis, necessitating further diagnostic evaluation and, usually, preventive chemotherapy.
RNDlNG INFECTED CHILDREN
The most efficient method of finding children infected with M tuberculosis is through contact investigations of adults with suspected or proven pulmonary tuberculosis.14 On average, 30% to 50% of household contacts to infectious adults will be infected. An integral part of the evaluation of a child with a positive or questionable Mantoux skin test is the testing of the family members to find both the infectious source case and other infected individuals. Many health departments will provide family investigâtions for children younger than 6 years old with a positive test, but limited resources often preclude them from performing investigations for the families of older children. The clinician should try to test the entire household whenever possible when a child has a positive Mantoux tuberculin test.
Figure 2. Algorithm for evaluating a child for tuberculous infection and disease.
All children with tuberculous infection acquired the organism from an infectious person. If health departments could conduct perfect contact investigations and immigration screening, tuberculin testing of other children would be completely unnecessary. Mass screening of low-risk children is extremely inefficient, costly, and often poorly conducted. An additional problem of testing low-risk children is that the Mantoux tuberculin skin test is extremely inaccurate in this group of children. Under normal conditions with immunocompetent individuals, the Mantoux test has a sensitivity and specificity of about 90% at a reaction size of 10 mm.10 When this test is applied to a population whose prevalence of tuberculous infection is 90% - such as in a tuberculosis sanitorium - 99% of the positive results will be true positives. However, when this same test is applied to a group whose infection prevalence is only 1%, less than 10% of the positive results will be true positives caused by tuberculous infection. Thus, when low-risk children are tested, the vast majority of positive reactions will be false positives, which nevertheless lead to further evaluation, treatment, cost, and anxiety for the family and the clinician. This is a powerful argument against broad-based tuberculin screening of low-risk children.
A suggested scheme for tuberculin testing in die office setting is presented in Figure 2. Children at high risk for tuberculosis (Table 1) should be tested periodically with a Mantoux tuberculin skin test, perhaps every 2 to 3 years. Children at low risk do not need to be tested routinely. If low-risk children are not tested, two important changes in practice will occur: 1) the ^15 mm classification for reading the Mantoux reaction will be mostly unnecessary, and 2) the use of MPTs will be eliminated.
It should be emphasized that frequent tuberculin testing of low- or high-risk school-age children will have virtually no impact on the reported numbers of tuberculosis cases in children. The majority of cases (60% to 70%) occur in preschool-age children, and most school-age children infected with M tuberculosis who will develop tuberculous disease will not do so until adolescence or young adulthood. Testing and treating high-risk school-aged children will prevent future cases of tuberculosis in adults but will do little to lower case rates in children. The way to decrease childhood tuberculosis rates is to identify adults with tuberculosis, render them noninfectious with effective chemotherapy, and bolster our ability to do proper contact investigations and immigration screening.
Without chemotherapy, 5% to 10% of infected immunocompetent adults will develop tuberculous disease in their lifetimes; about one half of this risk is in the first 3 years after initial infection. The risk for children is greater; up to 40% of children younger than 1 year of age with untreated tuberculous infection will develop pulmonary disease, as will 24% of children aged 1 to 5 years and 15% of adolescents aged 11 to 15 years.15
The vast majority of children with intrathoracic tuberculosis have no physical signs or symptoms. Infants and young children are more likely to have symptoms such as cough, fever, weight loss, and wheezing or decreased breath sounds. The pulmonary and systemic signs and symptoms may be alleviated by antibiotics, suggesting that bacterial superinfection may play a role in the pathogenesis of some cases.
The most common finding of tuberculous disease in a child is hilar or mediastinal adenopathy (Figure 3). If the nodes continue to enlarge, they may impinge upon, then erode through the nearby bronchus, leading to atelectasis of the affected lung segment or lobe (Figure 4). The resulting lesion is often referred to as a segmental or collapse-consolidation lesion. The radiographic shadows are similar to those caused by foreign body aspiration (the enlarged lymph nodes act like a foreign body). Occasionally, lobar pneumonia, multiple lobe involvement, or pleural effusion will occur. Complete radiographic resolution usually occurs after proper treatment, although this process may require 2 to 3 years, continuing long after effective chemotherapy has been discontinued. A normal chest radiograph is not necessary to stop treatment.
Figure 3. Hilar adenopathy in a child with tuberculosis.
About 30% of children with tuberculosis have extrathoracic disease. The most common site is the cervical or supraclavicular lymph nodes, followed by the meninges, bones and joints, kidneys and brain. Disseminated or miliary tuberculosis is most common in young infants and immunocompromised hosts.
The treatment of all children with tuberculous infection (positive Mantoux skin test) without disease is an established and effective practice. For children, the effectiveness of isoniazid (10 mg/kg/day) therapy has approached 100%, and the effect has lasted for at least 30 years.16 The American Academy of Pediatrics currently recommends a 9-month course of daily isoniazid. The limiting factor for die efficacy of treatment is often the family's adherence to giving tJhe medication. The clinician must be intimately involved in the ongoing treatment, providing encouragement for the family to give medication to the well-appearing child. Isoniazid therapy is extremely safe; routine biochemical monitoring and supplementation with pyridoxine (vitamin B6) are unnecessary. Monitoring for adherence and toxicity is best accomplished through frequent interviews and brief physical examinations.
Any young child - including newborn infants - exposed to an adult with suspected or proven pulmonary tuberculosis should be started on isoniazid therapy, even if the skin test is negative. These children may already be infected with M tuberculosis, but sufficient time has not passed for them to develop and display delayed hypersensitivity. For children with initially negative skin tests, testing can be repeated 3 to 4 months after contact with the case is broken. If die repeated skin test is negative, treatment can be stopped; if it is positive, a full course of therapy should be given. Newborns whose mother or other adult contacts have a positive skin test without disease do not need to receive treatment.
Figure 4. Hilar adenopathy and partial atelectasis of the right upper lobe in a child with primary tuberculosis.
The incidence of drug-resistant tuberculosis is increasing in the United States. For any child with tuberculous infection with or without disease, it is crucial to try to identify the source case so their children and susceptibility results can be used to direct the child's treatment. For tuberculous infection due to an isoniazid-resistant organism, rifampin ( 10 to 20 mg/kg/day) should be given for 9 months.
Standard treatment for most forms of tuberculous disease in children consists of isoniazid and rifampin for 6 months, supplemented during the first 2 months with pyrazinamide (30 kg/day).17 Treatment can be given daily or twice weekly under the supervision of a public health care worker if adherence with daily treatment is questionable. In general, pediatricians caring for children with suspected tuberculous disease should seek the advice of an expert in tuberculosis.
Tuberculous infection and disease rates among children in the United States are rising, but they also are being confined mostly to well-defined groups of highrisk children. Most children are at very low risk for infection, and routine screening of these large numbers of low-risk children is not necessary. Clinicians who care for children need to stratify their patients by risk for tuberculosis, and provide periodic testing for those truly at risk. Public health efforts need to concentrate on identifying and completing treatment of adults with infectious tuberculosis who are the sources of infection for high-risk children. We need to use the best tools available, and use them wisely.
1. Starke JR. Jacobs RF, Jereh J. Resurgence oí tuberculosis in children. ] Pediatr. 1992;120:839-855.
2. Selwyn PA, Hanel D, Lewis VA, et al. A prospective study of the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N Engl ; Med. 1989;320:545-550.
3. Jones DS. Malecki JM, Bigler WJ, Witte JJ, Oxtoby MJ. Pediatric tuberculosis and human immunodeficiency virus infection in Palm Beach County, Florida. AmJ Dis Chad. l99Z;f46:lio6-f[70.
4. Davidson PT, Ashkar B, Salem N. Tuberculosis testing of children entering school in Los Angeles County. California. Am Rev Respir Dis. 1990;141(suppl):A336. Abstract.
5- Barry MA, Shirley L, Grady MT, et al. Tuberculosis infection in urban adolescents: results of a school-based testing program. Am 1 Putite Health. 1990;80:439-441.
6. Stead WW. Senner JW. Reddick WT, Lofgren JR Racial differences in susceptibility to infection by Mycobacterium tuberculosis. N Engl J Med. 1990;322:422-427.
7. Bloch AB. Rieder HL, Kelly GD, Cauthen GM, Hayden CH, Sniden DE, Jr. The epidemiology of tuberculosis in the United States. Clin Chest Med. 1989;10:297-313.
8. Leggiadro RJ, Callory B, Dowdy S, Larkin J. An outbreak of tuberculosis in a family day care home. Pediatr Infect Dis J. 1989;8:52-54.
9. Hsu KHK. Tuberculin reaction in children treated with isoniazid. Am J Dis CMU. 1983;137:1090-1092.
10. Seibert AF1 Bass JB Jr. Tuberculin skin testing: guidelines for the 1990s. / Respir Du. 1990;11:225-234.
11. Cantanzaro A. Multiple puncture skin tests and Mantoux test in Southwest Asian refugees. Chest. 1985;87:346-350.
12. American Thoracic Society. Diagnostic standards and classification of tuberculosis. AmRei-RespfrDis. 1990;142:725-735.
13. Nemir RL, Teichner A. Management of tuberculin reactions in children and adolescents previously vaccinated with BCG. Pediatr Infect Dis J. 1983;2:446-451.
14- Hsu KHK. Contact investigation: a practical approach to tuberculosis eradication. Am} Public HeM. 1963;53:1761-1769.
15. Miller FJW, Seale RME. Taylor MD. Tuberculosis m Children. Boston. Mass: Little, Brown &. Co; 1963.
16. Hsu KHK. Thirty years after isoniazid: its impact on tuberculosis in children and adolescents. ,/AMA. 1 984.-2 51 .- J 233-1 285.
17. Starke JR. Multi-drug therapy for tuberculosis in children. Pediatr Infect Dis J. 1990;9:785-793.
Children at High Risk for Tuberculous Infection in the United States
What Constitutes a Positive Mantoux Skin Test