February 1997
NEW YORK — If pediatricians want to control tuberculosis, Jeffrey Starke, MD, advised them to put down their skin tests and get out in their communities.
"We in our own offices don't keep children from getting TB," Starke said at the Infectious Diseases in Children symposium here. "It is the health department, either by doing a good job or a bad job of TB control, that largely controls whether children will be getting sick with TB. One of our most important activities as pediatricians is to be advocates to make sure the health department has appropriate resources and uses them wisely. That is much more important than all the skin testing that we are doing in most of our offices." Starke is director, Children's Tuberculosis Clinic, Houston.
Pediatric TB is prevented by ensuring prompt evaluation of all children who have been in close contact with an adult who has active disease.
"Physicians in the community must work closely with public health officials to design and implement appropriate control strategies. ... The public health department is really the only entity that can prevent most cases of childhood TB," Starke said. "The contact investigation is the key, and we have data from around the country that contact investigations are not being done properly. Why is it so important? It has the highest yield: 30% to 50% of the household contacts will be infected and up to 1% will have already developed active disease. It identifies the highest risk: recently infected children are more likely to develop TB in the next year or two than children or adults who were infected years ago. So it is the most important group to find and it may also be the best chance to treat because it is within the context of a family investigation and adherence to therapy probably can be better."
Source cases are often reported late to health departments, and no contact investigations are performed for up to a third of all reported cases.
![[bar]](/art/gradred.gif){kind=small}
As a result, TB remains the leading infectious disease worldwide. Globally, in the 1990s, 90 million people — 30 million of them children — will develop TB and 30 million — 5 million children — will die. In the United States, Starke estimated that about 10 million people are infected with Mycobacterium tuberculosis. 250,000 cases are expected during the 1990s, with 20,000 deaths.
The United States experienced a resurgence of TB between 1985 and 1992. Children were disproportionately affected: although cases increased by 20% among all age groups, they increased by 40% among children.
Pediatricians are in a critical position to ensure that TB control efforts are not abandoned or weakened.
"It is a truism that as a public health program to eliminate a disease approaches its goal, it is more likely that the control program rather than the disease will be eliminated," Starke said. "That is precisely what happened with TB control in the 1980s. With managed care and block grants and other things going on, many of us are concerned that this is going to happen again in the 1990s. Pediatricians will lead the way to make sure this does not happen."
The way to control pediatric TB is not to increase childhood screening but to use the tests more wisely.
The American Academy of Pediatrics (AAP) and the Centers for Disease Control and Prevention (CDC), as well as other advisory organizations, have abandoned routine Mantoux skin tests for all children. The multiple puncture test is not appropriate in any situation, Starke emphasized.
"Every TB expert and everybody that issues recommendations says that multiple puncture skin testing should no longer be done in the United States. Period. End of story. Absolutely no discussion. Remember, this is in the context of us also saying that many fewer skin tests should be done in general practice than have been done traditionally," Starke said. "But when a skin test is truly needed, it should be done by the best available technique, which is the Mantoux skin test, and it should be done properly."
TB skin tests have about a 90% specificity and sensitivity rate, which would give a 99% positive predictive value in a population with a very high prevalence of TB, for example, in a TB sanitarium, Starke said. But outside of that high-prevalence setting, such as in a population where the risk of TB infection is 1%, the positive predictive value falls to less than 10%, which means more than 90% of positive skin tests would be false positives.
"What's wrong with false positives?" Starke asked. "For most tests you have a screening test and a definitive test, and the definitive test has a sensitivity and specificity that approaches 100%. With TB, the screening test is also the definitive test. The problem is you have no test that can discriminate the true positives from the false positives, so you must forever treat both of those groups the same. If you are dealing with a low-prevalence population, the majority of what you are dealing with are false positives."
Treating people with false-positive skin tests as though they were truly infected incurs tremendous cost. It is not just the child who must be evaluated and treated; it also the child's family. And with public health departments increasingly strapped for funds and manpower, evaluating the families of children with false positive skin test results is a low priority, Starke said.
"This means the TB skin test is very useful for high risk children and next to useless — and I may even submit to you harmful — for low-risk children," Starke said. "I believe it actually does more harm than good because we don't have kids in the suburbs dying from TB. That is not what is happening. I take care of between 40 and 80 children a year with active TB and virtually none of the cases would have been prevented through any kind of screening program that ever has been set up.
"Screening has been de-emphasized by the CDC and AAP over the last few years," he continued. "It is a terrible test for low-risk children. Mass screening is probably not cost beneficial unless the mass infection rate is greater than 30% and you have more than 65% adherence to therapy; in other words, the treatment actually does some good. The only group of individuals who may fit this bill in certain communities are foreign-born children. School-based testing as mass testing is a horrible idea. One may use the schools to reach certain groups of high-risk children, and I am not underestimating the moral, social, and ethical problems that this causes, but in terms of a strategy it is what makes the most sense. Mass testing of children is a bad idea."
So what is the answer? The CDC and AAP devised an alternative method for evaluating skin tests depending on the child's risk for infection. For example, a 5-mm reaction is considered a positive result in people who are at the highest risk for developing disease. This highest risk population includes people who are immunocompromised, those who are in close contact with someone who has TB, or those who are suspected of having TB. Although this recommendation may create more false positive results, it may help ensure that those most at risk for developing the disease will be identified.
At the other end of the spectrum, Starke suggested no testing for children at low risk for infection. Pediatricians uncomfortable with that suggestion or those living in areas with a prevalence of TB, or for children whose history of exposure is unknown, may test children before entry into school or at adolescence. Testing children at 1 year of age is no longer endorsed by the AAP.
Children at high risk for infection should be tested periodically every two to three years; annual testing is not recommended.
When a child does have a positive skin test, the pediatrician must take a thorough history to determine whether the child had in fact been exposed to TB. A good physical exam is imperative; signs and symptoms are often absent in children, and about one fourth of cases are extrapulmonary. In addition to assessing the child physically, pediatricians must also evaluate whether the child and family will adhere to treatment regimens.
"How do we treat TB infection? Still pretty much good old isoniazid, unless you know for a fact that the person [the child] got it from has isoniazid-resistant TB," Starke said. "This is where the health department should help you. For the majority of children, you will not have a source case and isoniazid is the treatment of choice."
You can express your
views on this article, or other relevant themes, in the
Infectious Diseases in Children
Specialty Forums.